FSH dose is negatively correlated with number of oocytes retrieved: analysis of a data set with ~650,000 ART cycles that previously identified an inverse relationship between FSH dose and live birth rate

Abstract

Purpose

To evaluate whether total FSH dose was negatively correlated with number of oocytes retrieved in a large data set where previously, a negative correlation between FSH dose and live birth rate was identified.

Methods

Data from 650,637 fresh autologous in vitro fertilization (IVF) cycles reported to the Society for Assisted Reproductive Technology between 2004 and 2012 were included. Logistic regression analysis was performed to determine if the relationship between total FSH dose used during ART with number of oocytes retrieved was impacted by the patient’s health prognosis, age, BMI, ovarian stimulation protocol, or infertility diagnosis.

Results

The number of oocytes retrieved was negatively correlated with FSH dose (P < 0.0001). Regardless of patient prognosis, age, BMI, ovarian stimulation protocol, and infertility diagnosis, the highest number of oocytes retrieved was in the 1001–2000 IU FSH group, and was 36–51% lower in the > 5000 IU compared with the optimal, 1001–2000 IU, FSH groups. Overall, ~80% of patients received FSH doses outside of the optimal FSH dose. Moreover, 61% of good prognosis patients (excludes individuals likely prescribed higher FSH doses) received doses exceeding the optimal dose range.

Conclusion

The inverse relationship between FSH dose and the number of oocytes retrieved independent of patient age or health implies that excessive FSH doses during ART may be detrimental to oocyte retrieval.

Introduction

Ovarian stimulation during assisted reproductive technology (ART) cycles utilizes exogenous gonadotropins to override the mechanism by which a single ovarian follicle completes folliculogenesis. This allows for the retrieval of multiple oocytes which can be used for ART. The total FSH dose administered during ovarian stimulation is dependent on patient characteristics (for example, age and diagnosis of diminished ovarian reserve) and responsiveness to FSH. Hence, patients that respond poorly or slowly to ovarian stimulation tend to receive higher doses [1–4]. However, whether increasing FSH dose improves oocyte retrieval has yet to be conclusively demonstrated and was investigated in the current study using the same data set used by Baker et al. (2015) to identify the negative correlation between total FSH dose and live birth rate [5].

Presumably one measure of the efficacy of an FSH dose is the number of oocytes that are recovered during an ART cycle. In terms of the number of oocytes retrieved, very few studies have examined the dose–response relationship to FSH across a range of doses. Whilst there is evidence of a positive dose–response relationship, there is conflicting evidence as to whether the highest doses evaluated in these studies were, in fact, detrimental to oocyte retrieval [6–10]. Comparatively, more studies have assessed the effects of two doses, generally lower compared to a higher dose, on ART outcomes. Results of these studies indicate that high FSH doses are either detrimental to or have no effect on ART outcomes, including the number of oocytes retrieved, fertilization rate, number of embryos produced, embryo quality, blastocyst rate, number of transferrable or frozen embryos, and live birth rate [11–20].

In addition to the number of oocytes retrieved, some studies also report the total number of follicles that develop or the number of mature oocytes retrieved in response to gonadotropin stimulation [9, 14, 15, 19, 21]. This is particularly useful as follicular development and oocyte maturation (including the number of mature oocytes) are considered indices of ovarian function and responses to gonadotropin stimulation. Whilst not available for the current data set, such data could be used to investigate whether excessive FSH doses negatively affect ovarian function. In addition, the mechanisms by which high FSH doses during ovarian stimulation impact ovarian function and subsequently follicular maturation are poorly understood. Data from studies in mice have shown that in vivo, high FSH doses may uncouple the oocyte from the somatic cells of the follicle and in vitro, accelerate nuclear maturation and may induce chromosomal abnormalities [22, 23]. Similarly, bovine studies have shown that using high gonadotropin doses during superovulation was detrimental to ovulatory follicle function, including estradiol production and ovulation rate [24] and decreased oocyte developmental competence resulting in decreased rates of transferrable, high quality, embryos [25] and in a decrease in ovulation rate [26]. These data demonstrate that high gonadotropin doses have a negative impact on follicular maturation and may explain, at least in part, the negative impact of high FSH doses on oocyte retrieval and ART outcomes including fertilization rate, number of embryos produced, embryo quality, blastocyst rate, number of transferrable embryos, and live birth rate.

Other key considerations relating to efficacy of ovarian stimulation include oocyte maturation and quality (reflected in embryo production and the live birth rate), minimizing the risk of ovarian hyperstimulation syndrome (OHSS) and cost of treatment. Generally, increasing oocyte retrieval rates have been positively associated with improved IVF outcomes (blastocyst formation, pregnancy rates) and live birth rate [27–30]. Based on these data, current ovarian stimulation procedures aim to induce the development of 5–15 oocytes, with low- (< 4 oocytes) and high-(> 15 oocytes) responders predicted to have poorer outcomes [31], in terms of live birth rate and OHSS, respectively [32]. A number of studies have recommended using 150 IU as the daily FSH dose to achieve optimal oocyte retrieval and maintain high live birth rates [13, 17, 18].

Previous analysis of the same data set of > 650,000 ART cycles used herein concluded that high FSH doses were negatively correlated to live birth rate and within FSH dose categories, the number of oocytes retrieved was positively correlated with live birth rate irrespective of age, BMI, and patient diagnoses (diminished ovarian reserve or ovulatory disorder) [5]. It was also noted in this previous publication that daily FSH dose ranged from 25 to 450 IU and total FSH dose administered per cycle ranged from 75 to 18,000 IU indicating that the daily and the total FSH dose varied substantially between cycles [5]. However, the relationship between FSH dose and the number of oocytes retrieved was not evaluated in the previous study [5]. Given that a key outcome of ovarian stimulation is to increase the number of oocytes retrieved in an ART cycle, the study reported herein was designed to investigate the hypothesis that total FSH dose is also negatively correlated with the number of oocytes retrieved using this same data set. A wide range of factors will influence a patient’s response to ovarian stimulation and likewise, a range of factors also influence a clinician’s choice of FSH dose. The purpose of this analysis was not to identify all factors that contribute to these, as this is not possible with the current data set. Instead, using this data set, arguably one of the largest currently available, consisting of an unselected population and patients with different characteristics (for example, age, BMI), diagnoses (for example, diminished ovarian reserve, PCOS) known to affect the number of oocytes retrieved and who were treated with a wide range of doses, we investigated the aforementioned hypothesis.

Methodology

Data were collected and verified by the Society for Assisted Reproductive Technology (SART) and reported to the Centers for Disease Control and Prevention in compliance with the Fertility Clinic Success Rate and Certification Act of 1992 (Public Law 102-493). The data in the SART CORS include data from > 90% of clinics providing IVF in the United States and are validated annually with some clinics having on-site visits for chart review based on an algorithm for clinic selection. During each clinic visit, data reported by the clinic were compared with information recorded in patients’ charts. Ten out of 11 data fields selected for validation were found to have discrepancy rates of ≤ 5% [33]. Only data from 650,637 fresh, autologous IVF cycles reported between 2004 and 2012 were included in this study. In the current study, the following data were excluded from statistical analyses: ART cycles where no oocytes were retrieved, ART cycles where ovulation was induced using oral medications, and ART cycles where the number of oocytes retrieved was > 1000. These discrepancies were likely due to data entry errors. Logistic regression models were used to examine the relationship between FSH dose (< 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU) with number of oocytes retrieved during each ART cycle for patients within the following categories: health prognosis (good or negative), age (< 35, 35–39, ≥ 40 years), BMI (< 18.5, 18.5–24.9, 25.0–29.9, and ≥ 30), ovarian stimulation protocol (agonist suppression, agonist flare, antagonist suppression), and infertility diagnosis (endometriosis, PCOS, diminished ovarian reserve, tubal disease, unexplained infertility, and other reasons for ART). As noted previously, FSH dose refers to the total gonadotropin dose as reported to SART CORS and includes the total dose of FSH from FSH-only and human menopausal gonadotropin (hMG) preparations [5] and does not include any luteinizing hormone activity, if administered. Serum anti-Müllerian hormone (AMH) concentrations were first included in the SART CORS database from 2012 (the final year’s data included in this analysis) and subsequently have been used to inform FSH dose selection in patients. Due to the limited serum AMH concentration available for the data set, it was not included in the current analysis.

Good prognosis patients were defined as patients < 35 years of age with a body mass index (BMI) < 30 kg/m² and no diagnosis of either diminished ovarian reserve or endometriosis. This category excluded patients with the most significant confounding characteristics or diagnoses that would account for being administered higher FSH doses based on the information available in the SART CORS database (mean total FSH dose: 2551 ± 3 IU). In contrast, the negative prognosis category included patients with one or more of the following characteristics and diagnoses: patient age of ≥ 35 years, BMI ≥ 30 kg/m², diminished ovarian reserve, or endometriosis. By comparison, this category included patients who were more likely to be prescribed high FSH doses (mean total FSH dose: 3778 ± 3 IU). The rationale for examining the relationship between FSH dose and oocyte retrieval within different categories was that patient characteristics (for example, patient age and BMI), type of ovarian stimulation protocol used, and infertility diagnoses (for example, diminished ovarian reserve) can potentially affect response to ovarian stimulation and choice of dose.

Differences between the proportions of patients in each FSH dose and patient prognosis category were analyzed using chi-square tests. The FSH dose range with the highest number of oocytes retrieved was compared to the highest FSH dose range, to determine the proportional decrease in oocyte retrieval. All analyses were conducted in SAS (9.4). The study was previously approved by the Institutional Review Board at Michigan State University [5].

Results

The proportion of patients treated with each FSH dose range is included in Table 1 and patient characteristics and diagnoses for the overall, good and negative prognosis analyses is included Table 2. As expected, the proportion of patients receiving the three highest doses was higher in the negative, compared to the good prognosis patients (P < 0.0001; Table 1), and increased with age (P < 0.0001; Table 1). Likewise, the proportion of patients with each diagnosis varied in ways that were expected. For example, ~50% of patients in the good prognosis data set had a diagnosis of male factor infertility and/or unexplained infertility (Table 2).

Table 1.

Summary of the proportion of ART cycles categorized by FSH dose (IU) range for each analysis conducted

FSH dose (IU)	Proportion of patients (%)
	All patients	Good prognosis	Negative prognosis	< 35 years	35–39 years	≥ 40 years	BMI < 18.5	BMI 18.5–24.9	BMI 25.0–29.9	BMI > 30	Agonist suppression	Agonist flare	Antagonist suppression	Endometriosis	PCOS	Diminshed ovarian reserve	Tubal disease	Unexplained infertility	Other reason for ART
< 1000	3.0	5.1	2.0	4.4	2.0	1.7	4.6	3.4	2.4	1.6	3.4	2.0	2.6	2.4	5.5	1.7	2.4	2.9	3.1
1001–2000	19.8	34.4	12.9	30.8	13.7	5.9	23.4	21.6	20.1	16.8	26.9	6.2	14.8	17.4	34.3	5.7	18.8	20.9	16.6
2001–3000	27.4	33.6	21.9	32.9	26.5	16.5	25.3	26.8	27.5	28.6	32.5	15.7	24.4	28.3	29.6	15.0	29.2	29.9	25.5
3001–4000	17.3	13.5	21.7	14.6	19.6	19.2	16.4	16.7	17.1	18.3	16.1	19.8	18.1	18.1	13.7	18.0	18.0	17.7	18.5
4001–5000	15.6	7.9	19.3	9.6	18.6	24.0	14.7	15.0	15.5	16.2	11.4	25.0	18.3	16.4	8.9	24.3	15.7	14.3	17.5
> 5000	16.9	5.4	22.3	7.8	19.5	32.8	15.6	16.5	17.4	18.5	9.7	31.4	21.9	17.4	8.0	35.4	15.9	14.3	18.9
No. of cycles	650,637	206,248	444,389	288,583	236,743	125,311	9761	191,120	79,965	61,614	316,242	80,837	241,901	78,007	95,800	128,927	112,738	88,619	85,060

The total number of ART cycles for each analysis is included in the last row

Table 2.

Patient characteristics including age, BMI, and percentage of cycles with each diagnosis for the overall, good and negative prognosis analyses

	All cycles	Good prognosis	Negative prognosis
Mean patient age in years (range)	35.6 (18.2, 57.6)	31.2 (18.2, 34.9)	37.6 (18.7, 57.6)
Mean BMI (range)	25.4 (14.1, 68.4)	23.2 (14.1, 29.9)	26.3 (14.1, 68.4)
Diagnoses	Proportion of patients (%)
Male factor	38.1	45.9	34.5
Endometriosis	12.0	NA	17.6
Polycystic ovarian syndrome	14.7	22.2	11.2
Diminished ovarian reserve	19.8	NA	29.0
All tubal disease	17.3	17.0	17.5
Uterine abnormality	4.8	3.0	5.6
Other	13.1	12.5	13.4
Unexplained	13.6	17.5	11.8
No. of cycles	650,637	206,248	444,389

The mean patient age and BMI are included with the range (minimum, maximum) for the appropriate analysis. Likewise, the proportion of cycles including patients with a positive diagnosis for endometriosis, PCOS, diminished ovarian reserve, tubal disease, uterine abnormality, and other reasons for ART and unexplained infertility were calculated for each study population. The total number of ART cycles included for each analysis is included in the last row. Note that column totals are not equivalent to 100% because a patient could have more than one diagnosis recorded. NA not applicable

A negative relationship between the number of oocytes retrieved and FSH dose category (P < 0.0001; Fig. 1 top panel) was identified. The number of oocytes retrieved decreased by 48% from the 1001–2000 to > 5000 IU FSH dose categories and > 80% of patients received doses outside the 1001–2000 dose range (Table 1).

Fig. 1.

Relationship between total FSH dose (IU) and number of oocytes retrieved in patients undergoing fresh autologous IVF cycles for the overall data set (all patients; top panel) and parsed by patient prognosis (good and negative prognosis; bottom panel). Good prognosis patients were selected based on the following criteria: < 35 years, body mass index < 30 kg/m², no diagnosis of diminished ovarian reserve, or endometriosis. Negative prognosis patients were selected based on the following criteria: either ≥ 35 years and/or body mass index ≥ 30 kg/m² and/or diagnosis of diminished ovarian reserve, or endometriosis. FSH dose categories included < 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU. FSH dose category was significantly negatively correlated with the number of oocytes retrieved in all analyses (P < 0.0001). The figure also depicts (red line and text) the decrease in the number of oocytes retrieved in the highest (> 5000 IU) FSH dose category compared to the category with the highest number of oocytes retrieved (1001–2000 IU). The number of patients included in each analysis (N =) is depicted on the appropriate graph

A negative relationship (P < 0.0001) existed between the number of oocytes retrieved and FSH dose in patients without (good prognosis, Fig. 1) and with (negative prognosis, Fig. 1) characteristics or diagnoses increasing the likelihood of receiving higher FSH doses. In both analyses, the highest oocyte retrieval was observed in the 1001–2000 IU FSH dose category. The number of oocytes retrieved decreased by 38 and 46% in the > 5000 IU FSH dose category in the good and negative prognosis analyses, respectively (Fig. 1). In spite of the fact that the good prognosis subgroup excluded all patients likely receiving high FSH doses due to patient characteristics or diagnoses, 60% of these patients received doses > 2000 IU FSH (Table 1).

Regardless of patient age category, FSH dose category was negatively correlated (P < 0.0001; Fig. 2) with the number of oocytes retrieved and the highest oocyte retrieval was observed in the 1001–2000 IU FSH dose category. The decrease in the number of oocytes retrieved in the > 5000 compared to the 1001–2000 IU FSH dose groups was 42, 45, and 39% in the < 35, 35–39, and ≥ 40 year categories, respectively (Fig. 2). The proportion of patients receiving higher doses of FSH also increased with age (P < 0.0001; Table 1).

Fig. 2.

Relationship between total FSH dose (IU) and number of oocytes retrieved in patients undergoing fresh autologous IVF cycles, parsed by patient age (< 35 years, 35–40 years, and > 40 years). FSH dose categories included < 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU. FSH dose category was significantly negatively correlated with the number of oocytes retrieved in all patient age categories (P < 0.0001). The figure also depicts (red line and text) the decrease in the number of oocytes retrieved in the highest (> 5000 IU) FSH dose category compared to the category with the highest number of oocytes retrieved (1001–2000 IU). The number of patients included in each analysis (N =) is depicted on the appropriate graph

When ART cycles were stratified into BMI categories (< 18.5, 18.5–24.9, 25.0–29.9, > 30), FSH dose was negatively correlated with the number of oocytes retrieved, irrespective of BMI (P < 0.0001; Fig. 3). The number of oocytes retrieved decreased by 51, 50, 47, and 40% when patient BMI was < 18.5, 18.5–24.9, 25.0–29.9, and > 30, respectively, in the > 5000 compared to the 1001–2000 IU FSH dose categories (Fig. 3).

Fig. 3.

Relationship between total FSH dose (IU) and number of oocytes retrieved in patients undergoing fresh autologous IVF cycles, parsed by patient BMI (< 18.5, 18.5–24.9, 25.0–29.9, and ≥ 30). FSH dose categories included < 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU. FSH dose category was significantly negatively correlated with the number of oocytes retrieved in all patient BMI categories (P < 0.0001). The figure also depicts (red line and text) the decrease in the number of oocytes retrieved in the highest (> 5000 IU) FSH dose category compared to the category with the highest number of oocytes retrieved (1001–2000 IU). The number of patients included in each analysis (N =) is depicted on the appropriate graph

A negative relationship between FSH dose and the number of oocytes retrieved was also identified when subgroup analyses were conducted on the three most common ovarian stimulation protocols (agonist suppression, agonist flare, and antagonist suppression; all P < 0.0001). Compared to the 1001–2000 IU group, the number of oocytes retrieved decreased by 47, 41, and 46% in the agonist suppression, agonist flare, and antagonist suppression categories, respectively (Fig. 4).

Fig. 4.

Relationship between total FSH dose (IU) and number of oocytes retrieved in patients undergoing fresh autologous IVF cycles, parsed by the three ovarian stimulation protocols (agonist suppression, agonist flare, and antagonist suppression) most common in the data set. FSH dose categories included < 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU. FSH dose category was significantly negatively correlated with the number of oocytes retrieved in all ovarian stimulation protocol categories (P < 0.0001). The figure also depicts (red line and text) the decrease in the number of oocytes retrieved in the highest (> 5000 IU) FSH dose category compared to the category with the highest number of oocytes retrieved (1001–2000 IU). The number of patients included in each analysis (N =) is depicted on the appropriate graph

Likewise when evaluating the relationship between FSH dose and the number of oocytes retrieved in cycles where patients were diagnosed with endometriosis, PCOS, diminished ovarian reserve or tubal disease, unexplained infertility, or other reasons for seeking ART, a negative relationship was identified (P < 0.0001). For all of the above, the highest number of oocytes retrieved was noted in the 1001–2000 IU FSH dose category (Fig. 5). By comparison, the proportional decrease in the number of oocytes retrieved in the > 5000 IU FSH dose group was 46, 44, 36, and 46% in patients diagnosed with endometriosis, PCOS, diminished ovarian reserve, and tubal disease, respectively. For patients seeking ART due to unexplained infertility or for other reasons, the decrease was 42 and 48%.

Fig. 5.

Relationship between total FSH dose (IU) and number of oocytes retrieved in patients undergoing fresh autologous IVF cycles, parsed by cycles where patients were recorded as having a positive diagnosis for either endometriosis, PCOS, diminished ovarian reserve, tubal disease, unexplained infertility, and those seeking ART for other reasons. FSH dose categories included < 1000, 1001–2000, 2001–3000, 3001–4000, 4001–5000, and > 5000 IU. FSH dose category was significantly correlated with the number of oocytes retrieved in each of the categories (P < 0.0001). The figure also depicts (red line and text) the decrease in the number of oocytes retrieved in the highest (> 5000 IU) FSH dose category compared to the category with the highest number of oocytes retrieved (1001–2000 IU). The number of patients included in each analysis (N =) is depicted on the appropriate graph

Discussion

The most significant findings of this study show as follows: firstly, that a highly significant inverse relationship, independent of age and patient prognosis, exists between total FSH dose and oocyte retrieval in fresh, autologous ART cycles. The highest number of oocytes retrieved was observed in the 1001–2000 IU FSH dose category and comparatively, the highest dose category (> 5000 IU) was associated with 45 ± 1% decline in the number of oocytes retrieved. This supports increasing evidence that high FSH doses, even in low or poor responders, are detrimental to oocyte retrieval and ART outcomes. Secondly, ~80% of patients received a non-optimal dose, that is, a dose outside of the 1001–2000 IU FSH dose category. Consequently, there is limited evidence for use of high FSH doses during ovarian stimulation.

In previous studies, where more than two gonadotropin (including hMG) doses were investigated, the relationship between dose and oocyte retrieval was curvilinear. Oocyte retrieval increased with dose but the highest doses either did not increase or were detrimental to oocyte retrieval [6, 8, 9, 19, 34–37]. Likewise, when two doses were investigated or doses were termed high or low, the higher dose did not increase the number of oocytes retrieved [1, 12–15, 20, 38, 39]. However, a small number of studies reported contradictory findings [2, 7, 10], contributing to the idea that increasing FSH dose is a suitable strategy for increasing oocyte retrieval rates, particularly, in low or poorly responding patients.

One possible criticism of retrospective correlative analyses such as the one reported herein is that high FSH doses are administered to offset the possibility of cycle cancelation due to a poor response, thereby confounding the analysis. Increased patient age and diagnosis of a diminished ovarian reserve are associated with a poor response to ovarian stimulation and as a result, tend to be associated with higher total FSH doses [18, 40]. However, data presented herein, which were limited to fresh ART cycles, and by others show that using higher FSH doses in patients with either decreased antral follicle counts [14, 20], decreased serum AMH concentrations [16], or increased maternal age [10, 13, 19, 41] did not improve oocyte retrieval rates. We also conducted analyses where ART cycles were categorized by BMI, stimulation protocol, and infertility diagnoses. In all cases, FSH dose was negatively correlated with the number of oocytes retrieved and the highest number of oocytes retrieved was in the 1001–2000 IU dose range. Furthermore, we conducted a subgroup analysis excluding ART cycles where patients were predicted to respond poorly to FSH stimulation, namely, the good prognosis analysis. Despite excluding patients more likely to be prescribed high FSH doses, > 60% of these good prognosis patients still received doses higher than the optimal dose range.

A related criticism could be that patients receiving the lowest doses (< 1000 IU) may be undergoing minimal stimulation IVF cycles, particularly, since data for canceled cycles were not included in the analysis. Whilst the precise reasons for the choice for such a low total FSH dose are not available for this data set, previous studies have shown improved outcomes using minimal stimulation protocols for patients expected to respond poorly, for example, due to increased age or diagnosis of a diminished ovarian reserve [42, 43]. However, < 2% of patients with a diagnosis of a diminished ovarian reserve were included in the < 1000 IU FSH dose group and when patients were categorized by age, the oldest age category (≥ 40 years) had the lowest proportion of patients in the < 1000 IU FSH dose category. Whilst not supporting the use of minimal stimulation IVF protocols either, the analysis indicates that patient characteristics are unlikely to be the sole factors influencing choice of FSH dose during ART. Given that a major objective of ovarian stimulation is to increase the number of high quality oocytes retrieved for IVF whilst minimizing treatment cost and risk of OHSS, the data herein strongly support the use of total FSH doses within the 1001–2000 IU dose range regardless of the predicted responsiveness to ovarian stimulation.

In combination with the previously published analysis of our data set [5], we have demonstrated that total FSH doses greater than 2000 IU are likely to be excessive and detrimental to number of oocytes retrieved and live birth rate [5]. This is consistent with an increasing number of dose–response trials and retrospective, correlative studies also showing that high FSH doses had either no effect or were detrimental to IVF success [7, 10–12, 14–16] and live birth rate [4, 10, 17, 19, 20, 38]. Studies in mice have also shown that high FSH doses may perturb bidirectional communication between the oocyte and surrounding granulosa and cumulus cells and alter the kinetics of nuclear maturation, both of which are known to be detrimental to oocyte quality [22, 23]. Given the potential detrimental effects of high FSH doses on oocyte quality and the number of oocytes retrieved, there appears to be inadequate justification for the use of high FSH doses during ovarian stimulation.

With the available data, it is not possible to determine the precise reasons for the decrease in number of oocytes retrieved observed in the highest FSH dose category. Two potential reasons include a poor ovarian response (number of follicles that developed) or a reduction in oocyte pickup efficiency, that is, the number of oocytes retrieved in relation to the total number of follicles aspirated. Both support our hypothesis and demonstrate that high FSH doses during ovarian stimulation are likely to impair follicular maturation and subsequently decrease oocyte retrieval. With increasing data demonstrating the complexity of FSH and LH signaling pathways [44], it has become apparent that current understanding of the effects of exogenous gonadotropins during ovarian stimulation is incomplete. For instance, luteinization is a normal physiological process occurring in response to LH, following ovulation. However, premature luteinization has been documented in women following administration of high FSH doses during ovarian stimulation and results in significant perturbations in the follicular microenvironment [45–48]. Bovine studies also demonstrate that in cattle, superovulation with high FSH doses did not increase the number of ovulatory sized follicles produced and in fact decreased ovulation rate relative to lower doses [24]. Thus high FSH doses during ovarian stimulation and the resultant changes in the follicular microenvironment may be detrimental to oocyte and embryo quality, and potentially, oocyte recovery rate [22–24, 45, 49–51]. There is also evidence that delayed frozen embryo transfers mitigated the negative effect of high FSH doses on live birth rate and indicate that these high FSH doses may also have detrimental effects on the endometrium [52]. Although, the molecular mechanisms by which high FSH doses alter follicular development and endometrial receptivity are unclear; they warrant further investigation.

A possible limitation of this study is the absence of data regarding the maturity and quality of the oocytes that were retrieved at the different FSH doses. However, the strong inverse relationship between FSH dose and the number of oocytes retrieved (shown in the present study) and FSH dose with live birth rate (shown in the previous study) using the same data set [5] imply that relatively high FSH doses may also be detrimental to oocyte quality.

In conclusion, irrespective of age or other confounding factors, 1001–2000 IU FSH yielded the highest number of oocytes and non-optimal doses, which were administered in the majority of ART cycles, were detrimental to number of oocytes retrieved.

Author contribution

Z Clark performed the analyses and initial interpretation of the data, prepared the majority of the first draft of the manuscript, and contributed to all subsequent revisions. M Thakur contributed to interpretation of data, all revisions of the manuscript, and liaised with SART to secure approval for the use of the data set. R Leach provided advice and assisted in the interpretation of data, securing approval from SART for the use of the data set, and contributed to revising the manuscript. J Ireland oversaw this project, including the analysis of the data and the planning and preparation of all drafts of this manuscript.

Data Availability

Not applicable

Code availability

Not applicable

Declarations

Ethics approval

The study was previously approved by the Institutional Review Board at Michigan State University [5].

Consent to participate

Not applicable

Consent for publication

Not applicable

Conflict of interest

The authors declare no competing interests.