Use of gonadotropins in ovarian stimulation in Spain: Delphi consensus

Abstract Two-round Delphi study carried out in Spain. Three theme-based blocks were set out: 1) Patient profiles: therapeutic goal and parameters to be analysed according to POSEIDON patient profiles; 2) Ovarian stimulation protocols with antagonists: monotherapy (FSH) vs combined therapy (FSH + LH/HMG); 3) Safety and effectiveness of the devices. The antral follicle count and the anti-Müllerian hormone level were considered indicators that can be used to predict ovarian response. More than 80% of the participants agreed that FSH monotherapy is the recommended regimen in normal/hyper-responsive patients of < 35 years of age; that 150–300 IU is the dose to be used in ovarian stimulation in monotherapy depending on clinical parameters; and that FSH monotherapy improves patients’ comfort compared to two combined drugs. It was unanimously considered that the type of device used by the patient influences the comfort of the treatment. IMPACT STATEMENT What is already known on this subject? There is currently no consensus on the optimal treatment for controlled ovarian stimulation for patients undergoing IVF which leads to highly variable clinical practices. What the results of this study add? This study’s strong point is that, since it is a consensus, it has been possible to include more topics than would normally be dealt with in a systematic review or guidelines, which are generally based on a strict method that restricts the scope of the research. Experts have reached a consensus on most of the statements and based on these they have issued consensus statements that will enable the optimal use of gonadotropins in IVF. What the implications are of these findings for clinical practice and/or further research? This Delphi consensus provides a real-life clinical perspective on gonadotropin usage in IVF.


Introduction
Ovarian stimulation is crucial in each cycle of in vitro fertilisation (IVF) because the development of a sufficient number of follicles and attainment of a sufficient number of oocytes increases not only the live birth rate per cycle but also the cumulative live birth rate per cycle of treatment (Drakopoulos et al. 2016;Siristatidis, 2013).
Ovarian stimulation faces different challenges.On the one hand, the live birth rate per mature oocyte retrieved is lower than 5% (Goldman et al. 2013).On the other, the number of oocytes necessary to obtain at least one live birth increases exponentially with age (Goldman et al. 2017).Furthermore, a significant number of patients (from 9% to 24%) show a poor ovarian response (POR) to stimulation with exogenous gonadotropins (Roque et al. 2021).As a result, low rates of pregnancy and live births are obtained, both varying from 3% to 14% (Drakopoulos et al. 2016;Humaidan et al. 2017;La Marca et al. 2016;Tarlatzis et al. 2003;Ulug et al. 2003).
In order to predict the clinical response to stimulation and choose the most appropriate protocol, patients should be stratified according to ovarian reserve markers (La Marca et al. 2016).However, the choice of a marker is controversial and there is still debate as to what marker (or combination of them) is the most suitable (Bulletti et al. 2021).
The POSEIDON criteria stratify patients into four groups depending on a combination of quantitative and qualitative parameters (Alviggi et al. 2016).They are based on the woman's age, ovarian reserve biomarkers, ovarian sensitivity to exogenous gonadotropins, and the number of oocytes retrieved in an IVF cycle (Alviggi et al. 2016).
The 'ideal' approach to performing ovarian stimulation with POR patients is still not known in terms of medication type, dose and devices used (Bulletti et al. 2021).There is no single procedure that appears to clearly improve IVF outcomes for this subgroup of patients, and clinical trials are generally highly biased (Papathanasiou, et al. 2016).Various ovarian stimulation protocols have been put forward that use exogenous FSH combined with gonadotropin-releasing hormone (GnRH) analogues (Bulletti et al. 2021).Another point of discrepancy is whether to use a combined treatment of FSH with luteinising hormone (LH), hormones with an LH effect, or else monotherapy with FSH (Ferrando et al. 2020).
It can be concluded that there is currently no consensus on the optimal treatment for controlled ovarian stimulation for POR patients undergoing IVF (Olgan and Humaidan 2017;Pandian et al. 2010;Papathanasiou, et al. 2016), which leads to highly variable clinical practices.
Given this background, this study took up the following goals: 1) To identify the points of agreement or disagreement as regards the use of gonadotropins in IVF in Spain; and 2) To issue recommendations to optimise the use of gonadotropins in IVF, both conventionally and using IVF-ICSI.

Scientific committee and panel of experts
This project was carried out using the Delphi method from March to December 2021 in Spain.Initially, an advisory committee was set up tasked with analysing the main controversies in the sphere of ovarian stimulation in IVF, drawing up a questionnaire, interpreting the results from the Delphi rounds and critically reviewing the final report.
The committee members were chosen based on the following: they should be specialists and spend most of their working days in IVF sites (public or private); belong to a scientific society in this field; and be authors of publications or communications or have participated in clinical trials in the sphere of IVF.
In order to answer the questionnaire, a panel of experts was created with 40 fertility specialists from public and private sites with extensive experience in IVF and from all regions in the country.
As this study was based on a Delphi survey ethics approval was not required.

Creating statements. Questionnaire
The advisory committee first identified the areas of uncertainty that should form the basis for the questionnaire's structure.Three theme-based blocks were set out: 1) Patient profiles: therapeutic goal and parameters to be analysed based on POSEIDON patient profiles; 2) Ovarian stimulation protocols: monotherapy (FSH) vs combined therapy (FSH þ LH/HMG); 3) Safety and effectiveness of the devices.
Each member of the committee proposed different statements for the questionnaire with their corresponding bibliographic basis.To do so, the guidelines, protocols and other available evidence were reviewed, putting them into context with each committee member's clinical experience.A work meeting was held to pool all of the statements and validate the choices.The final questionnaire contained 33 statements divided into five blocks or sub-blocks.It was posted on a microsite that the participants accessed via a web link with a user password.

Statistical analysis and interpretation of results
Two Delphi rounds were carried out.In each of them, the panel of participants scored each statement on a 9-point Likert scale.The level of agreement was classified as 1-3 (disagree), 4-6 (neither agree nor disagree), or 7-9 (agree).
The mean values (standard deviation) were calculated, as well as the median and interquartile range (p25-p75) for each of the questionnaire's items.The level of significance was measured with the Kolmogorov-Smirnov goodness-of-fit test for distribution.
The criteria for consensus included 'unanimity' when 100% of participants agreed on the same Likert scale category, 'consensus' when there was agreement among � 80% of participants, 'majority' when there was agreement among � 66% of participants, and 'disagreement' when there was agreement among < 66% of the participants.For the purposes of this analysis, the 'unanimity', 'majority' and 'consensus' groups were considered all together as consensus.
After the first round, the questions that did not reach a consensus in replies went on to the second, reformulating the ones whose wording could be improved.When the statistical analysis of the results was available, the advisory committee met to discuss and define the conclusions from the study.
The initial questionnaire contained 33 statements divided among the three aforementioned blocks (with 16, 14 and 3 statements respectively).
� Patient profiles a. Parameters to stratify patients with a low ovarian response: 9 statements b.Goals of the ovarian stimulation treatment: 7 statements � Ovarian stimulation protocols: monotherapy (rFSH) vs.

Results
The questionnaire was answered by 40 fertility specialists.The participating entities are listed in Table 1.
Figure 1 shows the process of Delphi rounds.In the first round, a consensus was reached for 27 of the 33 matters raised.Of the six remaining questions, one was eliminated after being reviewed by the committee and considering that consensus would not be reached, while the rest went on to the second round, having been reformulated to make them more understandable, so that the final results included 32 statements.
Table 2 shows the results from the Delphi study.

Discussion
This Delphi consensus provides a real-life clinical perspective on gonadotropin usage in IVF.The experts reached a consensus on most of the statements and based on these they have issued recommendations (Table 3) that will enable optimal use of gonadotropins in IVF, whether conventionally or as IVF-ICSI.
As regards the parameters for stratifying patients with low ovarian response, initially there are different parameters available to stratify the patients, such as age, baseline folliclestimulating hormone (FSH) concentration, anti-M€ ullerian hormone (AMH) concentration, and the antral follicle count (AFC), used alone or in combination via algorithms or clinical nomograms.Despite this, the choice of marker is    controversial and there is still debate as to what marker (or combination of them) is the most suitable (Bulletti et al. 2021).Although other authors have described baseline FSH, body mass index and older age as predictors of ovarian response, the experts considered antral follicle count and the AMH value to be predictors of response, which are also indicators that are preferentially recommended by the guidelines of the European Society of Human Reproduction and Embryology (Bosch et al. 2020) compared to other markers.The participants agreed on recommending antral follicle counts for all patients before stimulation and in each cycle.The experts also consider that when a previous cycle is available, an antral follicle count or a determination of AMH should be carried out, but that it is unnecessary to measure the basal FSH.
Other authors have described factors such as baseline FSH, BMI, age or the number of follicles smaller than 11 mm as the most conclusive variables for patients under 35 years of age treated with recombinant FSH (rFHS) monotherapy (Howles et al. 2006).
As for the goals of treatment for ovarian stimulation, there was agreement of greater than 80% that the goal of ovarian stimulation for normal-responding patients could be set at obtaining 10-15 oocytes.There was also consensus that the adequate number of oocytes to be obtained in a cycle may vary depending on the patient's age and clinical profile, with a higher risk of complications if the result is greater than 20 oocytes; and that the increase in this risk determines the number of oocytes to be obtained in an ovarian stimulation cycle.In this vein, a recent Delphi study by Bulletti et al has revealed a broad consensus on assertions such as the goal of 8-14 oocytes retrieved after ovarian stimulation in fresh cycles (Bulletti et al. 2021).According to data from the National Registry of the Spanish Fertility Society (SEF) for 2019 (Sociedad Española de Fertilidad, 2019), the estimated number of oocytes necessary to achieve pregnancy was 14.3, and the number of inseminated or injected oocytes necessary to achieve pregnancy was 11.3 (in both fresh and cryopreserved transfers) (Sociedad Española de Fertilidad, 2019).
In terms of the ovarian stimulation protocols, a consensus was achieved in all of the items in this block.The experts agreed that the most usual indications for receiving ovarian stimulation treatment combined with LH are an age of over 35 years and poor response in previous cycles, although the scientific evidence is controversial, to say the least.It was unanimously accepted that other indications for adding LHacting drugs are excessive pituitary LH suppression or a lack of response to FSH alone in previous cycles.
As regards ovarian stimulation, there is no single intervention that clearly improves IVF outcomes for patients with poor ovarian response.In general, clinical trials give controversial results (Papathanasiou et al. 2016).In a systematic review of 75 clinical trials carried out with POR patients over 15 years, there was a notable heterogeneity detected in defining this type of patient and in the interventions studied, as well as serious methodological problems (Papathanasiou et al. 2016).Only 10% of the trials reported a significant improvement in reproductive outcomes after applying different protocols, many of which were based on a single clinical trial (Papathanasiou et al. 2016).
Although there are guidelines on ovarian stimulation in IVF-ICSI published in 2019 by the ESHRE (Bosch et al. 2020), they are limited by the fact that only a small proportion of patients are included in randomised controlled trials; it is estimated that only 35% of POR patients would meet the inclusion criteria used in large clinical trials (Orvieto et al. 2021).Thus, prospective studies with an adequate sample size and Table 3. Recommendations to optimise the use of gonadotropins in IVF, both conventionally and in IVF-ICSI.
Stratification of patients with a low ovarian response � The ovarian stimulation protocol must be designed taking into account the patient's age.� The result from the previous cycles carried out by the patient is fundamental in designing a new ovarian stimulation protocol.� For patients with previous stimulation cycles, the most important indicators to take into account on designing the ovarian stimulation protocol are the doses of gonadotropins used and the number of oocytes retrieved.� For patients with no previous cycles, the indicators to be considered on designing the ovarian stimulation protocol are the antral follicle count, anti-M€ ullerian hormone, age and body mass index.� The indicators that can be used to predict ovarian response are the antral follicle count and the anti-M€ ullerian hormone value.� The antral follicle count evaluation is a clinical practice recommended for all patients prior to stimulation, in each cycle.Ovarian stimulation protocols: monotherapy (rFSH) � FSH monotherapy is the regimen commonly recommended for normal-or hyper-responding patients under 35 years of age.� The dose used in ovarian stimulation in monotherapy is from 150 to 300 IU depending on the clinical parameters defined at the start of stimulation.� In the event of a low previous response or low ovarian reserve, a dose of 300 IU is the most common.� For hyper-responsive patients, the starting dose for FSH is usually 75 to 150 IU. � For patients with polycystic ovaries, the starting dose of FSH is usually 75 to 150 IU. � FSH in monotherapy increases the patient's comfort compared to using two different drugs in combination.� FSH in monotherapy reduces mistakes in handling the medication.� Stimulation can begin between day 2 and day 4, giving greater flexibility for the first control.� There is no evidence that changes or variations in the dose used after ovarian stimulation has begun gives advantages in the final outcome.Ovarian stimulation protocols: monotherapy (rFSH) vs. combined therapy (rFSH 1 LH/HMG) with antagonists � Combined ovarian stimulation treatment is only indicated for certain groups of patients.� The most common indications for receiving ovarian stimulation treatment combined with LH are an age of over 35 years and poor response in previous cycles.� Other indications for adding LH-acting drugs are excessive pituitary LH suppression or a lack of response to FSH alone in previous cycles.� Generally, normal-responding patients under 35 years of age do not benefit from ovarian stimulation treatment combined with drugs with a LH effect.Safety/efficiency of the device � The type of device used by the patient influences the safety of ovarian stimulation treatment.� The type of device used by the patient influences the comfort of ovarian stimulation treatment.
statistical power are still needed to respond to the clinical questions raised about ovarian stimulation for these patients.
Some authors have put forward different protocols using exogenous FSH combined with gonadotropin-releasing hormone (GnRH) analogues, whether agonist or antagonist (Bulletti et al. 2021).On the one hand, proper individualisation of the FSH dose has been put forward as one of the most sensitive steps to give a successful outcome and reduce complications from IVF (Bulletti et al. 2021).On the other, fixed-dose gonadotropin stimulation regimens (as opposed to variable doses during stimulation) have been considered a suitable option for some patients due to their lower cost and greater comfort, with fewer follow-up visits compared to the protocol with variable doses of gonadotropins (Bulletti et al. 2021).Another point of discrepancy is whether to use monotherapy wit FSH or else combine FSH with luteinising hormone (LH) or hormones with an LH effect (Ferrando et al. 2020).The scarcity of scientific evidence and the variability in clinical practice have prompted several Delphi studies to seek consensus (Bulletti et al. 2021;Orvieto et al. 2021).
The clinical effectiveness of rFSH monotherapy has been demonstrated in both clinical trials and real-life studies (Ferrando et al. 2020;Selman and Rinaldi 2016;Vlaisavljevi� c et al. 2010;Strowitzki et al. 2016).In one of them, carried out in Spain with 1,222 patients treated in 26 sites, the pregnancy rates varied from 23.2% in poor responders to 37% in normal-responders (Ferrando et al. 2020).
The range of rFSH doses chosen by experts for ovarian stimulation protocols generally coincides with the range published in the ESHRE guide, which recommends doses ranging from 150 IU for good responders to 300 IU for poor responders (Bosch et al. 2020).
The results obtained in terms of safety and efficiency of the device agree with various studies comparing devices used in ovarian stimulation.However, these show differences in the reduction of handling errors (Imthurn et al. 2014;Saunders et al. 2020), comfort of use (Quintero et al. 2016) and the patient's preference (Quintero et al. 2016).This study's strong point is that, since it is a consensus, it has been possible to include more topics than would normally be dealt with in a systematic review or guidelines, which are generally based on a strict method that restricts the scope of the research.Nevertheless, the consensus also has its limitations.For example, not all of the statements reached 100% agreement.Furthermore, although these recommendations represent experts' points of view, they are not universal.The patient's individual characteristics should always be taken into account before choosing the type of treatment.
We trust that this analysis will provide a contribution in improving care and therapeutic outcomes for POR patients undergoing IVF.

Conclusions
This Delphi study gives relevant data regarding the clinical reality of ovarian stimulation for POR patients undergoing IVF in Spain: 1) The different gonadotropins cannot alter the quality of the oocytes obtained; 2) The goal of ovarian stimulation is to obtain an adequate number of oocytes (in poor responders this is the maximum number possible, and in normal-or hyper-responders it is between 10 and 15); 3) The starting dose is determined well by response prediction factors; 4) The most valid prediction factor is the response in a previous cycle; 5) If there is no previous cycle, the antral follicle count and AMH value are the most reliable indicators of response.
This practical perspective is of great value in a field with numerous clinical questions raised and a high variability in clinical practice due to the scarcity of conclusive scientific evidence.

Table 2 .
Final results from the Delphi study.Statements that did not reach consensus by the end of the study are marked in italics.
FSH: follicle-stimulating hormone; LH: luteinising hormone.a: Statements that did not reach consensus in first round; b: statements revised for 2 � round � statement eliminated.