{"paper_id":"c6bb48a2-870a-467f-a14d-fe8a1c8c7de3","body_text":"The number of oocytes retrieved is regarded as a predictor of live birth rates (LBRs) following fresh embryo transfer in ovarian stimulation for IVF. Studies have shown that LBRs either plateau or decline after a particular number of oocytes are retrieved, and optimum yields in the range between 6 and 15 oocytes have been suggested, taking the risk of ovarian hyperstimulation syndrome (OHSS) in high responders into consideration ( van der Gaast  et al. , 2006 ;  Sunkara  et al. , 2011 ;  Ji  et al. , 2013 ;  Steward  et al. , 2014 ;  Magnusson  et al. , 2018 ;  Polyzos  et al. , 2018 ;  Lobo  et al. , 2025 ).\nThe LBR per fresh cycle is the conventionally reported outcome of IVF. In recent years, there has been a marked increase in cryopreservation of embryos and subsequent frozen embryo transfers (FETs). The switch of cryopreservation method from slow freeze to vitrification, as well as embryo culture to the blastocyst stage, has led to a significant increase in embryo cryo-survival rates ( Saket  et al. , 2021 ). Another factor behind the rise in FET cycles is the increase in single embryo transfers (to avoid the potential complications of multiple pregnancies), resulting in more excess embryos available for cryopreservation. In addition, ‘freeze-all’ cycles have emerged as an alternative to fresh embryo transfer cycles, initially as a strategy in combination with GnRH agonist triggering to avoid OHSS in high responders ( Devroey  et al. , 2011 ), but also used with preimplantation genetic testing, or to overcome asynchrony of embryo and endometrial receptivity ( Shapiro  et al. , 2008 ). The increase in number and successful outcomes of FET cycles suggests that the cumulative live birth rate (CLBR) has arisen as a more clinically relevant outcome than fresh cycle LBR in IVF treatment.\nStudies imply that CLBR continues to increase with number of oocytes retrieved beyond the plateau or decline observed for fresh cycle LBRs ( Ji  et al. , 2013 ;  Drakopoulos  et al. , 2016 ;  Toftager  et al. , 2017 ;  Magnusson  et al. , 2018 ;  Polyzos  et al. , 2018 ;  Law  et al. , 2019 ;  Neves  et al. , 2023 ).\nThis study aimed to investigate the association between the number of oocytes retrieved and CLBR in a pooled analysis of four randomized controlled trials using the recombinant FSH follitropin delta for ovarian stimulation. The trials included one stimulation cycle and assessed live birth from the fresh cycle and subsequent cryopreserved cycles initiated within 1 year of the ovarian stimulation.\n\nThis study is a pooled analysis investigating the association between number of oocytes retrieved and CLBR in 1746 patients from four randomized controlled trials (trial registration numbers  NCT01956110  [ESTHER-1],  NCT03564509  [RAINBOW],  NCT03740737  [RITA-1], and  NCT03738618  [RITA-2]) conducted in 12 countries in Europe, North America, and South America. The trials were identified (without a registered protocol) from clinical trials available in the Ferring Pharmaceuticals database up to June 2023 and were selected based on the use of follitropin delta for ovarian stimulation and the collection of outcome data from both fresh and frozen cycles as part of the study design. Follitropin delta dose–response trials were excluded as well as trials investigating follitropin delta in repeated ovarian stimulation cycles ( Fig. 1 ).\nSelection of trials and patients.  AMH, anti-Müllerian hormone; CG, chorionic gonadotropin.  Olsson  et al ., 2014 ,  2015 ;  Shao  et al ., 2023 .\nThe original trial protocols and results have been described previously ( Nyboe Andersen  et al. , 2017 ;  ClinicalTrials.gov, Identifier NCT03740737, 2018 ;  ClinicalTrials.gov, Identifier NCT03738618, 2018 ;  Fernandez Sanchez  et al. , 2022 ;  Doody  et al. , 2023 ;  Scheiber  et al. , 2023 ;  Grover  et al. , 2024 ). The trials had obtained regulatory and ethical approval and were performed in accordance with the principles of the Declaration of Helsinki, the International Conference on Harmonisation Guidelines for Good Clinical Practice, and local regulatory requirements. Written informed consent was obtained from all participants.\nThe trials included women undergoing their first or second IVF/ICSI cycle (no women had previously received follitropin delta). The women were diagnosed with unexplained infertility, tubal infertility, endometriosis stage I/II, or had partners with male factor infertility. They were 18–42 years of age, had a BMI of 17.5–38.0 kg/m 2 , and regular menstrual cycles of 24–35 days. The exclusion criteria included endometriosis stage III–IV and a history of recurrent miscarriage. Patients included in the current analysis were treated with follitropin delta and had at least one oocyte retrieved. Overall, 1827 patients received follitropin delta in the four clinical trials. One patient with a missing anti-Müllerian hormone (AMH) value and 80 patients with no oocytes retrieved were excluded from the analysis. The remaining 1746 patients had at least one oocyte retrieved and were included in the analysis. Because of differences in the choice of comparators between the trials, comparator data were not included in the analysis ( Fig. 1 ).\nOvarian stimulation was performed with follitropin delta (REKOVELLE ® , Ferring Pharmaceuticals). The patients either received an individualized dosing of follitropin delta at a fixed daily subcutaneous dose determined by their serum AMH concentration at screening and body weight at randomization/stimulation Day 1 ( NCT01956110  and  NCT03564509 ; the individualized dosing algorithm of follitropin delta is detailed in  Nyboe Andersen  et al. , 2017 ), or they followed a dosing regimen with a daily starting dose of 12 or 15 µg, which could be adjusted in increments of 3 µg after the first 4 stimulation days ( NCT03740737  and  NCT03738618 ).\nPatients either followed a GnRH antagonist protocol, with follitropin delta initiated on Day 2–3 of the menstrual cycle and GnRH antagonist (0.25 mg daily) initiated on stimulation Day 5 or 6 and continued throughout stimulation, or a GnRH agonist protocol, with GnRH agonist (0.1 mg daily) started in the mid-luteal phase of the menstrual cycle and follitropin delta initiated after 14 days.\nTriggering of final follicular maturation was performed when ≥3 follicles or ≥2 follicles with a diameter ≥17 mm were observed. Triggering was performed with hCG if <25 follicles or <20 follicles ≥12 mm were observed. If ≥25 follicles or ≥20 follicles ≥12 mm were observed, triggering was performed with a GnRH agonist (not applicable for the GnRH agonist protocol), or the cycle was cancelled (when >25 follicles or >35 follicles ≥12 mm were observed, depending on the trial). If the investigator judged that the triggering criterion could not be reached by Day 20, the cycle was cancelled (in one of the trials, triggering could still be performed if at least 1 follicle ≥17 mm was observed).\nOocyte retrieval took place 36 ± 2 h after triggering. Insemination was performed by IVF or ICSI, and single or double blastocyst transfer was performed on Day 5. Single blastocyst transfer was mandatory for patients ≤37 years or ≤34 years (depending on the trial), whereas in patients ≥38 years or ≥35 years (depending on the trial), single blastocyst transfer was performed if a good-quality blastocyst was available; otherwise, double blastocyst transfer was performed. All viable surplus blastocysts were cryopreserved. For patients who were triggered with a GnRH agonist, all blastocysts were cryopreserved, and no transfer was performed in the fresh cycle. All pregnancies from the fresh cycle and frozen cycles initiated within 1 year after the start of stimulation were followed until birth. In two of the trials, frozen cycles were performed in accordance with local clinical practice. In the other two trials, the protocol mandated either natural cycles or programmed cycles with oestradiol and progesterone. In the cryopreserved cycles, there was no defined transfer policy, and the choice between single and double blastocyst transfer was at the discretion of the investigator.\nThe primary outcome was cumulative live birth, defined as the birth of at least one live neonate (irrespective of gestational age) in the fresh or a subsequent frozen cycle using blastocysts derived from a single oocyte retrieval. Each patient underwent one stimulation cycle. CLBR was calculated as the number of patients with at least one live birth (only the first delivery was considered), divided by the number of all patients included in the analysis (i.e. patients treated with follitropin delta who had at least one oocyte retrieved).\nThe secondary outcome was live birth, defined as the birth of at least one live neonate, in the fresh cycle only. LBR was calculated as the number of patients with at least one live birth divided by the number of all patients included in the analysis.\nContinuous variables are presented as mean ± SD or median and interquartile range. Categorical variables are presented as number and percentages.\nThe association between the number of oocytes retrieved and cumulative live births was assessed using a logistic regression analysis with fractional polynomials to obtain predicted cumulative live births where the final model was selected using a backward selection method. The model was unadjusted. A sensitivity analysis was performed to evaluate the effect of age and AMH on the model. Furthermore, a logistic regression analysis was used to describe CLBR on one hand and the number of oocytes retrieved (grouped by 1–7, 8–14, 15–19, 20–24, and ≥25 oocytes), age (grouped by <35, 35–37, and 38–42 years), or AMH level (<15 or ≥15 pmol/l) on the other hand. Finally, subgroup analyses were performed based on follitropin delta dosing strategy (individualized dosing based on AMH level and body weight, fixed throughout stimulation [trials  NCT01956110  and  NCT03564509 ], or starting doses of 12 or 15 µg with potential adjustments during stimulation [trials  NCT03740737  and  NCT03738618 ]).\n\nOverall, 15 trials with follitropin delta were identified in the database. Of those, five were performed in healthy volunteers (with no pregnancy outcome data), two were dose–response trials, three did not collect outcome data from frozen cycles, and one evaluated follitropin delta in repeated cycles. As a result, four trials were selected for the analysis ( Fig. 1 ).\nIn total, 1746 patients were included in the analysis, selected based on ovarian stimulation with follitropin delta and retrieval of at least one oocyte ( Fig. 1 ). The mean age of the study population was 33.8 years, and the median AMH level was 17.0 pmol/l ( Table 1 ). The most common primary reason for infertility was unexplained infertility, followed by male factor and tubal infertility, and the vast majority of patients (94.2%) were treated with a GnRH antagonist protocol. For 1575 patients (90.2%), the trial stimulation cycle was their first IVF/ICSI cycle, while for 171 patients (9.8%), it was their second IVF/ICSI cycle. Overall, 1541 patients (88.3%) received hCG triggering, and 205 patients (11.7%) received GnRH agonist triggering.\nCharacteristics of the study population.\nAMH, anti-Müllerian hormone.\nOf the 1746 patients, 740 patients (42.4%) initiated frozen cycles. In total, the study population underwent 2948 cycles: 1746 fresh cycles (referring to ovarian stimulation cycles, with or without transfer) and 1202 frozen cycles (initiated cycles, with or without transfer). The maximum number of frozen cycles for a patient was six (the numbers of patients initiating cryopreserved cycle 1–6 are found in  Table 2 ). At the cut-off for data collection (1 year after start of stimulation), 279 patients (16.0% of the study population) had not achieved a live birth but had remaining cryopreserved blastocysts.\nTreatment outcomes of the study population.\nThe study outcomes of the overall study population are presented in  Table 2 . The mean number of oocytes retrieved was 12.4; the distribution is illustrated in  Fig. 2 . Transfer in the fresh cycle was performed for 1303 patients (74.6%). Overall, 1440 patients (82.5%) had single blastocyst transfer only, 60 patients (3.4%) had double blastocyst transfer only, 85 patients (4.9%) had both single and double blastocyst transfer (in separate cycles), and 161 patients (9.2%) had no transfer. The LBR in the fresh cycle was 29.1%. The LBR decreased by the number of repeated frozen cycles, and no live birth was achieved in the fifth or sixth frozen cycle. The overall CLBR was 51.4%.\nDistribution in the number of oocytes retrieved in the overall study population (N = 1746).\nFigure 3  shows the observed and predicted CLBR by number of oocytes retrieved. Due to the high distribution in the number of oocytes retrieved and the low number of patients at each fraction of the highest ovarian responses observed, the figure only displays data up to 35 oocytes. The predicted CLBR increased with the number of oocytes retrieved, reaching above 60% at >15 oocytes retrieved and above 70% at >20 oocytes retrieved. At 21–25 oocytes, the CLBR started to plateau. Sensitivity analysis of the model, including adjustments for age and AMH, did not change the conclusion of the analysis. Subgroup analyses based on dosing strategy ( Supplementary Fig. S1 ) displayed similar trends; however, the CLBR was higher in patients treated with individualized follitropin delta compared to patients who received starting doses of 12 or 15 µg (peaks were observed at 80.7% and 72.3%, respectively), and the plateau was shifted towards higher oocyte numbers, starting at around 25 oocytes.\nObserved and predicted cumulative live birth rate by number of oocytes retrieved.  Blue bars represent observed rates, and red bars represent predicted rates. Predicted cumulative live birth rates were obtained using a logistic regression analysis with fractional polynomials.\nThe number of patients by subgroup (based on age, AMH, and number of oocytes retrieved) is presented in  Table 3 . The subgroup analyses demonstrated that the CLBR decreased with increasing age, from 57.1% in patients <35 years to 51.6% in patients 35–37 years and to 35.8% in patients 38–42 years. A continued increase in CLBR from 15 oocytes retrieved was observed in older patients (≥38 years); from 41.3% at 15–19 oocytes retrieved to 53.4% at 20–24 oocytes retrieved to 58.7% at ≥25 oocytes retrieved. No equivalent benefit from retrieving more oocytes was observed in younger patients (<38 years), where the corresponding rates were 72.5%, 68.0%, and 78.8% in patients <35 years and 70.3%, 73.1%, and 71.5% in patients 35–37 years ( Fig. 4A ). The CLBR was similar for patients with AMH <15 pmol/l and ≥15 pmol/l (52.0% vs 50.9%), with a similar tendency towards a higher CLBR with higher number of oocytes retrieved in both subgroups ( Fig. 4B ). Subgroup analyses by dosing strategy ( Supplementary Fig. S2 ) displayed similar patterns with some variations.\nSubgroup analyses of the cumulative live birth rate.  ( A ) Cumulative live birth rate by number of oocytes retrieved (grouped) and age group. ( B ) Cumulative live birth rate by number of oocytes retrieved (grouped) and anti-Müllerian hormone (AMH) level. Number of oocytes retrieved are grouped as ≤7 oocytes (blue bars), 8–14 oocytes (red bars), 15–19 oocytes (green bars), 20–24 oocytes (brown bars), and ≥25 oocytes (purple bars). Data are presented as percentage (95% CI). Predicted cumulative live birth rates were obtained using logistic regression analyses.\nDistribution of patients by age group, AMH group, and number of oocytes retrieved.\nAMH = anti-Müllerian hormone.\nA comparison of the observed fresh cycle LBR and the CLBR by number of oocytes retrieved demonstrated that the LBR decreased beyond 14 oocytes retrieved, while the CLBR continued to increase by number of oocytes retrieved ( Fig. 5 ).\nObserved cumulative live birth rate versus fresh cycle live birth rate by number of oocytes retrieved (grouped).  The solid blue line represents observed cumulative live birth rates, and the dotted red line represents observed live birth rates.\n\nThis pooled analysis of four randomized clinical trials evaluated the association between cumulative live birth and the number of oocytes retrieved in patients undergoing ovarian stimulation with follitropin delta. It demonstrated an increase in CLBR by the number of oocytes retrieved that started to plateau at 21–25 oocytes. Subgroup analyses by dosing strategy displayed similar trends with some variations. These variations may be attributable to differences in baseline characteristics of the trial populations, where patients treated with individualized follitropin delta on average were slightly younger and had lower BMI. Previous studies implied that CLBR continues to increase with the number of oocytes retrieved ( Ji  et al. , 2013 ;  Drakopoulos  et al. , 2016 ;  Toftager  et al. , 2017 ;  Polyzos  et al. , 2018 ;  Law  et al. , 2019 ;  Neves  et al. , 2023 ), although a plateau or decline beyond a certain number of oocytes retrieved, as observed for fresh cycle LBR, was also reported ( Nelson  et al. , 2016 ;  Magnusson  et al. , 2018 ). Some of the previous studies reporting a continuous increase in CLBR by number of oocytes retrieved used grouped categories of ovarian response with a highest cut-off at >15 oocytes retrieved ( Ji  et al. , 2013 ;  Drakopoulos  et al. , 2016 ;  Toftager  et al. , 2017 ). The same cut-off in this study would have led to the same conclusion. However, in the current study, the additional subgrouping of patients with responses at >15 oocytes and the analysis using fractional polynomials enabled further evaluation of CLBR at higher responses. Variations from the results of studies using similar approaches ( Polyzos  et al. , 2018 ;  Law  et al. , 2019 ) may be due to the lower number of patients in this study.\nThe difference between the optimum number of oocytes to achieve a live birth in the fresh cycle and the optimum number of oocytes to maximize the cumulative live birth was evident in the current study. The fresh cycle LBR declined substantially beyond 14 oocytes retrieved, while the CLBR continued to increase beyond 14 oocytes retrieved. This evidence is relevant when personalizing treatment for patients. Similar observations of diverging curves for LBR and CLBR after ∼15 oocytes were made in previous studies evaluating both outcomes ( Ji  et al. , 2013 ;  Magnusson  et al. , 2018 ,  Polyzos  et al. , 2018 ).\nThe CLBR decreased with increasing age, which is in agreement with previous studies ( Polyzos  et al. , 2018 ;  Law  et al. , 2019 ;  Neves  et al. , 2023 ). While Polyzos  et al.  observed a continuous increase in CLBR by the number of oocytes retrieved in all age groups, Neves  et al.  and Law  et al. , noted a plateau around 25 oocytes for patients <35 years, whereas patients 35–44 years displayed a continued increase in CLBR beyond 25–30 oocytes. The current analysis suggests that patients <38 years do not benefit from oocyte numbers ≥20, since no evident increase in CLBR was observed, while an increase in CLBR from 20 oocytes was observed in patients ≥38 years. There was no difference in CLBR based on AMH level. Based on these findings, the optimum number of oocytes retrieved from a stimulation cycle is individual and dependent on patient characteristics. In younger patients with a higher risk of OHSS, who do not benefit from increased oocyte yields, the target number of oocytes could be lower than in older patients, who may increase their chances of achieving a live birth with a higher number of oocytes retrieved. However, it should be noted that in patients undergoing repeated frozen cycles, the success rates decreased as the number of repeated cycles increased.\nAs far as we are aware, this is the first pooled analysis based on patient-level data from randomized controlled trials to investigate the association between the number of oocytes retrieved and CLBR. It has the advantages of including a well-defined patient population, representing multiple centres and countries, and employing study procedures in accordance with current clinical practice.\nFrom a clinical trial perspective, the inclusion of frozen cycles initiated up to 1 year after the start of stimulation is a satisfactory time period. On the other hand, the time limit prevents frozen blastocysts not used at the cut-off from inclusion in the analysis, which constitutes a limitation of the study. However, the majority of the patients (84.0%) in the study had either achieved a live birth or had no remaining blastocysts at the 1-year cut-off. Another limitation is the relatively small number of patients with more than 20 oocytes retrieved. In total, 249 patients (14.3%) had ≥20 oocytes retrieved, and 38 patients (2.2%) had ≥35 oocytes retrieved. In two of the included trials, no cycle cancellations were performed in case of excessive ovarian response; instead, triggering was performed with GnRH agonist, transfer was cancelled, and all blastocysts were cryopreserved. As a consequence, some patients had a very high number of oocytes retrieved. The low number of patients with a high ovarian response makes predictions unreliable in this range.\nIn conclusion, this analysis suggests an increase in CLBR with the number of oocytes retrieved up to a plateau starting at 21–25 oocytes retrieved following ovarian stimulation cycles with follitropin delta and subsequent frozen cycles. A benefit of an increase from 20 oocytes retrieved was age-dependent and evident in older but not in younger patients.","source_license":"CC-BY-4.0","license_restricted":false}