Basal FSH values are positively associated with aneuploidy incidence in pre-advanced maternal age (AMA) but not in AMA patients

Abstract

Purpose

To assess the impact of maternal age on the association between maternal basal FSH and aneuploidy.

Methods

A retrospective study including data from 1749 blastocysts diagnosed as euploid or aneuploid by PGT-A (preimplantation genetic testing for aneuploidy). Aneuploidy incidence was compared between embryos from mothers with high vs. low basal FSH levels (above and below the group median, respectively) in total, pre-AMA (advanced maternal age; < 35 years, 198 embryos) and AMA (≥ 35 years, 1551 embryos) patient groups, separately. To control for the interference of potentially confounding variables, the association between aneuploidy and high basal FSH levels was assessed by multivariate logistic analysis in overall, pre-AMA and AMA patient groups.

Results

Overall, aneuploidy rate was 9% higher (p = 0.02) in embryos from patients with high basal FSH (63.7%) compared to those with low basal FSH (58.4%). In the pre-AMA subgroup, aneuploidy incidence was 35% higher (p = 0.04) in embryos from patients with high basal FSH (53.5%) compared to those with low basal FSH (39.4%). Differently, aneuploidy occurrence did not vary between embryos from AMA patients with low (61.0%) and high (64.8%) basal FSH (p = 0.12). The multivariate analysis revealed that, in pre-AMA embryos, the association between aneuploidy occurrence and high basal FSH is independent of potential confounding variables (p = 0.04).

Conclusion

Maternal basal FSH values are associated with embryo aneuploidy in pre-AMA but not in AMA patients. The present findings suggest that basal FSH is a useful parameter to assess aneuploidy risk in pre-AMA patients and reinforce the hypothesis that excessive FSH signalling can predispose to oocyte meiotic errors.

Introduction

Embryo aneuploidy is a prevalent cause of pregnancy loss and genetic defects/syndromes [1, 2]. In IVF practice, increased aneuploidy incidence has been recognised as a major obstacle associated with advanced maternal age (AMA), which has led to the development and use of PGT-A (preimplantation genetic testing for aneuploidy) [3–5]. Interestingly, however, aneuploidy incidence is also relatively high in young mothers (40–50% from 20 to 23 years), falling to remain at its lowest rates (20–30%) between 25 and 34 years of age, after which it increases exponentially until menopause [6, 7]. Even during the lowest incidence phase, aneuploidy rates are still relatively high, ranging from 20 to 30% [6, 8]. Despite the considerable occurrence of embryo aneuploidy throughout the women’s reproductive lifespan (≥ 20%), the efficacy of PGT-A in improving IVF results has been consistently questioned [9, 10], particularly for patients under 40 years of age; a negative association between PGT-A use and cumulative live birth has been recently reported in pre-AMA patients [11]. Therefore, a better characterisation beyond the age of the maternal profiles associated with higher and lower aneuploidy risk could provide more robust parameters for the decision on whether to perform PGT-A, as well as new valuable insights about the mechanisms underlying this genetic error.

Like aneuploidy incidence, basal FSH levels substantially increase during AMA as a consequence of preantral ovarian reserve depletion. This leads to a reduction in the population of small antral follicles and, consequently, of inhibin B secretion, thus attenuating the negative feedback that modulates FSH production and release from the pituitary [12–16]. In addition to this temporal relationship, the observation that increased maternal FSH levels are associated with the occurrence of Down’s syndrome and recurrent unexplained pregnancy loss reinforced the long-postulated hypothesis that FSH may be involved in the mechanisms linking maternal age and aneuploidy [17–23]. However, discrepant findings have been reported regarding the association between aneuploidy and maternal basal FSH levels, rendering the reliability of this potential hormonal predictor of aneuploidy risk questionable until the present day. Elevated FSH was associated with trisomic pregnancy losses in the second, but not in the first study published by Kline and co-workers [24, 25]. Of note, an apparent relationship was observed when the data was dichotomized according to maternal age in the earlier study [25]. In IVF/ICSI patients undergoing spontaneous miscarriages, aneuploidy incidence was initially reported to positively correlate with basal FSH serum concentrations [26], which, however, was not observed in later studies [27, 28]. Interestingly, more recently, polar body aneuploidy occurrence was significantly correlated with intrafollicular but not basal FSH concentration [29], implying that basal and intrafollicular FSH levels do not necessarily correspond and suggesting that FSH may locally activate or enhance paracrine mechanisms leading to aneuploidy.

Similarly, basal FSH levels were found to correlate inconsistently with ART outcomes [30–33]. We have recently shed light on this inconsistency by demonstrating that the relationship between basal FSH and ART success is strongly influenced by maternal age; FSH basal levels were found to be significantly and negatively associated with post-ICSI/IVF live birth achievement in pre-AMA (< 35) but not in AMA patients [34]. Moreover, basal FSH was not associated with oocyte recovery in pre-AMA patients, suggesting that FSH affects ICSI/IVF success through oocyte quality [34]. The absence of a significant association between basal FSH and ICSI/IVF outcomes in AMA patients has been interpreted as a consequence of increased inter-cycle basal FSH variability with age [35–37]. This phenomenon very likely reflects the higher incidence of luteal phase dominant follicles in AMA women, capable to inhibit FSH secretion during the early follicular phase, when blood samples are taken for basal FSH measurement [22, 38, 39].

In the light of the inconsistences in the existing literature, aiming to clarify the robustness and clinical utility of the association between aneuploidy and maternal basal FSH values, we assessed the influence of maternal age on this relationship. More specifically, we analysed a database composed by ICSI/PGT-A cycles to test the hypothesis that basal FSH values are significantly associated with aneuploidy occurrence in pre-AMA but not in AMA patients.

Materials and methods

Study design

This is a retrospective cohort study including data collected from January 2018 to March 2021, from 1749 blastocysts tested for aneuploidy by PGT-A, produced by 480 sub-fertile couples in 589 autologous ICSI/PGT-A cycles performed at “Huntington Medicina Reprodutiva”—Eugin Group, Sao Paulo, Brazil. Ethical approval was obtained from the Ethics in Research Committee/National Committee of Ethics in Research-Brazil (CAAE 42762820.0.0000.8054).

To test the hypothesis that maternal age influences the relationship between maternal basal FSH values and embryo aneuploidy, the incidence of PGT-A positive (aneuploid) and negative (euploid) diagnoses in embryos from mothers with basal FSH higher and lower than the median was assessed in the total (n = 1749 embryos), pre-AMA (< 35; age range, 19.9 to 34.9; n = 198 embryos) and AMA (≥ 35; age range, 35 to 47.2; n = 1551 embryos) populations, separately; the age cut-off was defined as previously determined and predominantly adopted in ICSI/IVF practice and research [3, 40]. In addition, the association between aneuploidy occurrence and high basal FSH levels (i.e. above the median) was assessed by univariate and multivariate logistic analyses in total, pre-AMA and AMA patient groups, separately. The primary outcome of the study, aneuploidy occurrence, was the dependent variable in the univariate and multivariate analyses. Mosaic embryos were excluded from the analysis.

Basal FSH measurement

Basal FSH serum concentrations were measured in blood samples collected within the first 3 days after menstruation by an electrochemical assay. Blood samples were collected within 60 days preceding the IVF cycle.

Ovarian stimulation, fertilisation and embryo culture

Ovarian stimulation was performed in combination with agonist or antagonist downregulation and was defined in accordance with patient profile. Gonadotropin treatment started on the third day of the menstrual cycle, and doses were adjusted according to treatment responses monitored by ultrasound examination and estradiol measurements. Ovulation was triggered by administering hCG (250 mcg) or GnRH (0.3 mg) 34 h before ovum pickup.

Cumulus-oocyte complexes were collected and incubated for 3 h in pre-equilibrated medium (GIVF Plus, Vitrolife) before enzymatic (HYASE, Vitrolife) and mechanical denudation. Semen was collected by masturbation, and sperm samples were prepared with either density gradient or swim-up protocols. Metaphase II oocytes were fertilised by ICSI [41], and presumptive zygotes were cultured in EmbryoSlides dishes (Vitrolife A/S) containing CSCM-C medium (Irvine Scientific) overlayed with paraffin oil (Ovoil, Vitrolife). Embryo culture was uninterrupted until the blastocyst stage in a time-lapse incubator (EmbryoScope Plus, Vitrolife A/S) at 5% O₂, 7.2% CO₂.

PGT-A

Trophectoderm biopsies were performed on day 4, 5, 6 or 7, depending on the time required to achieve the expanded blastocyst stage, allowing good distinction between the inner cell mass and the trophectoderm. Each blastocyst was placed into a 15 µL drop of buffered medium (GMOPS Plus, Vitrolife) under paraffin oil (Ovoil, Vitrolife), and the biopsy procedure was performed as previously described [42]. In brief, the zona pellucida was opened with 2–3 laser shots (ZILOS-tk®, Hamilton Thorne), after which 5–8 trophectoderm cells were aspirated using a biopsy micromanipulation pipette and applying another 2–3 laser shots directed to the cell junctions. The pulling or flicking methods were used to obtain the biopsy sample as applicable [43]. The biopsy samples were washed in droplets of PBS with 0.1% polyvinyl alcohol (PVA) and placed in 0.2 mL PCR tubes preloaded with 2.5 µL of PBS under a stereomicroscope for visual confirmation, after which they were immediately stored at − 20 °C. Both solutions and the protocol were provided by a local commercial genetics laboratory (Igenomix, Brazil), to which the biopsy samples were transported for embryo ploidy status determination through WGA (whole genome amplification) followed by next-generation sequencing (NGS) [5].

Statistical analysis

In the characterisation of maternal age groups, categorical variables are described by percentages, and continuous variables are represented by mean values and standard deviation. Pre-AMA and AMA groups were compared withFisher’s exact test for parameters represented as percentages and with the Mann–Whitney test for endpoints expressed as continuous variables, since the data was found not to be normally distributed (Table 1).

Table 1.

Maternal and cycle characteristics of embryos grouped according to maternal age

	Overall	Pre-AMA (< 35 years)	AMA (≥ 35 years)	p§
N. of embryos	1749	198	1551	NA
N. of cycles	589	54	535	NA
N. of patients	480	48	432	NA
N. aneuploid embryos (%)	1067 (61%)	92 (46.4%)	975 (62.9%)	< 0.0001
Basal FSH (IU/L; mean ± SD)	7.88 ± 6.78	7.36 ± 4.91	7.95 ± 6.98	0.1
N. oocytes recovered (mean ± SD)	12.22 ± 7.21	16.71 ± 8.95	11.65 ± 6.75	< 0.0001
Maternal age (mean ± SD)	38.47 ± 3.37	31.83 ± 3.26	39.32 ± 2.26	NA
Maternal BMI (mean ± SD)	23.08 ± 3.05 (n = 1456)*	22.75 ± 3.34 (n = 166)*	23.12 ± 3.01 (n = 1290)*	0.03
FSH total dose (IU; mean ± SD)	2708.53 ± 1194.11 (n = 1028)*	2478.82 ± 743.12 (n = 136)*	2743.55 ± 1245.23 (n = 892)*	0.03

^*Number of embryos with BMI and FSH total dose data available

^§p values refer to the comparison pre-AMA vs. AMA

The incidence of aneuploidy and euploidy in embryos from patients with basal FSH higher and lower than the group median was assessed in total, pre-AMA and AMA patient groups with the chi-square test (Table 2). To investigate whether the association between high basal FSH values (above the median of the cohort) and aneuploidy occurrence is independent of confounding variables, multivariate logistic regression analyses were performed in total, pre-AMA and AMA embryos, separately. In preparation for the multivariate analyses, the association between patient/cycle variables potentially impacting aneuploidy manifestation (maternal age, maternal BMI, number of oocytes, total FSH administered during ovarian stimulation) and aneuploidy occurrence was tested in linear univariate analyses. Subsequently, in the multivariate analysis, all variables significantly or tending to be significantly associated with aneuploidy occurrence in the univariate analysis (p < 0.1) were included as independent covariates (Table 3). Throughout the study, differences with p < 0.05 were considered to be statistically significant, while p values between 0.05 and 0.1 were considered tendencies of a significant difference. The statistical analysis was performed using the STATA software 9.0 (Stata Corporation, USA).

Table 2.

Aneuploidy incidence in embryos from patients with low and high serum basal FSH, stratified according to maternal age

	Overall, n = 1749		Pre-AMA (< 35 years), n = 198		AMA (≥ 35 years), n = 1551
	Low bFSH	High bFSH	Low bFSH	High bFSH	Low bFSH	High bFSH
Euploid	41.56%, 372/895	36.30%, 310/854	60.60%, 60/99	46.46%, 46/99	39.04%, 303/776	35.22%, 273/775
Aneuploid	58.43%, 523/895	63.70%, 544/854	39.39%, 39/99	53.53%, 53/99	60.95%, 473/776	64.77%, 502/775
p	0.02		0.04		0.12

Notes: High and low basal FSH (bFSH) values are those higher and lower than the group median, respectively. p values refer to differences in euploid/aneuploid incidence between embryos from patients with low and high bFSH in three different maternal age groups (overall, pre-AMA and AMA)

Table 3.

Univariate and multivariate logistic analyses depicting the association of aneuploidy occurrence with basal FSH levels, maternal BMI, number of oocytes recovered, maternal age and FSH total dose in embryos grouped according to maternal age

	Overall				Pre-AMA (< 35 years)				AMA (≥ 35 years)
	Univariate		Multivariate		Univariate		Multivariate		Univariate		Multivariate
	OR (95% IC)	p	OR (95% IC)	p	OR (95% IC)	p	OR (95% IC)	p	OR (95% IC)	p	OR (95% IC)	p
Basal FSH (high vs. low)	1.02 (1.01–1.04)	< 0.0	1.02 (1.00–1.04)	0.03	1.08 (1.01–1.15)	0.03	1.08 (1.00–1.16)	0.04	1.02 (1.00–1.04)	0.03	1.01 (0.99–1.03)	0.22
Maternal BMI	1.03 (1.00–1.07)	0.07	1.02 (0.98–1.06)	0.26	0.91 (0.83–1.01)	0.08	0.93 (0.84–1.03)	0.18	1.05 (1.01–1.09)	0.01	1.02 (0.98–1.06)	0.46
N. oocytes	0.97 (0.96–0.98)	< 0.0001	0.99 (0.98–1.01)	0.34	1.00 (0.97–1.03)	0.87	___	___	0.97 (0.96–0.99)	< 0.0001	0.99 (0.97–1.01)	0.23
Maternal age	1.19 (1.15–1.23)	< 0.0001	1.18 (1.14–1.22)	< 0.0001	1.04 (0.96–1.14)	0.33	___	___	1.38 (1.31–1.45)	< 0.0001	1.34 (1.26–1.43)	< 0.0001
Total FSH	1.00 (1.00–1.00)	0.12	___	___	1.00 (1.00–1.00)	0.78	___	___	1.00 (1.00–1.00)	0.18	___	___

Note: High and low basal FSH values are those higher and lower than the group median, respectively

Results

Characterisation of AMA and pre-AMA ICSI/PGT-A cycles

As expected, the occurrence of aneuploidy was higher in AMA (62.9%) as compared to pre-AMA embryos (46.4%; p < 0.0001; Table 1). In addition, oocyte yield was higher in pre-AMA (p < 0.0001), whilst BMI and FSH total dose were higher in AMA cycles (p < 0.03; Table 1). Basal FSH levels tended to be higher in AMA as compared to pre-AMA cycles (p = 0.1; Table 1); medians of basal FSH were 6.80, 6.05 and 6.86 IU/L for overall, pre-AMA and AMA cycles, respectively.

The impact of maternal age on the association of basal FSH with aneuploidy

In overall embryos subjected to PGT-A, regardless of maternal age, the occurrence of aneuploidy varied significantly, although not drastically, between embryos from patients with high and low basal FSH (p = 0.02; Table 2); aneuploidy incidence was approximately 9% higher in embryos from mothers with high basal FSH (63.7%) as compared to embryos from patients with low basal FSH (58.4%). The same distribution pattern was observed in embryos produced by pre-AMA patients, although in this case, aneuploidy incidence was approximately 35% higher in embryos from patients with high basal FSH (53.5%) as compared to those from mothers with low basal FSH (39.4%; p = 0.04; Table 2). In contrast, the incidence of aneuploidy did not vary significantly between embryos from AMA patients with high or low basal FSH levels (p = 0.12; Table 2).

The univariate and multivariate analyses revealed that, in overall embryos, although basal FSH, maternal age and number of oocytes recovered were significantly associated with aneuploidy occurrence, only the first two relationships are independent of the influence of other variables (Table 3). Alternatively, in embryos from pre-AMA patients, among all variables included in the univariate analysis, only basal FSH was significantly associated with aneuploidy (p = 0.03; positive association), while a trend towards a negative association with maternal BMI (p = 0.08) was observed. The multivariate analysis indicated that the relationship between high basal FSH and aneuploidy occurrence is independent of maternal BMI in pre-AMA embryos (p = 0.04; Table 3). Differently, in embryos from AMA patients, aneuploidy incidence was positively associated with basal FSH (p = 0.03), number of oocytes recovered (p < 0.0001) and maternal age (p < 0.0001) in the univariate analysis, while, according to the multivariate analysis, maternal age was the only parameter significantly associated with aneuploidy independently of the interference of other variables (p < 0.0001; positive association; Table 3).

Discussion

In the face of previous contrasting studies, the association of maternal basal FSH levels with aneuploidy and, consequently, the utility of this hormonal parameter to assess aneuploidy risk have remained unclear [24–29]. Motivated by previous data indicating a strong influence of maternal age on the accuracy of a single basal FSH measurement and thus on its fertility prognostic value [22, 34–39], herein, we report novel and clarifying evidence that basal FSH values are associated with post-ICSI aneuploidy occurrence in pre-AMA but not in AMA patients.

As previously proposed in studies assessing the utility of FSH basal levels to predict ICSI/IVF outcomes [22, 34], the observation that basal FSH values are only significantly associated with aneuploidy in pre-AMA patients is very likely a consequence of increasing inter-cycle variability of FSH measurements as maternal age advances [35–37]. This phenomenon appears to result from altered follicle dynamics in AMA (i.e. increased incidence of long-lasting lutein phase dominant follicles), leading to a short-term decrease in FSH secretion at the beginning of the follicular phase [22, 38, 39]. Importantly, this may cause frequent observation of misleading basal FSH values in AMA patients, non-representative of the predominant circulatory levels. Indeed, repeated measurements did increase the prognostic performance of basal FSH in IVF practice [44], supporting the interpretation above. Alternatively, the absence of a significant correlation between basal FSH and aneuploidy in AMA patients may in part reflect the occurrence of predominantly high values of basal FSH accompanied by high incidence of aneuploidy in this patient group, thus providing insufficient variation of both variables for the detection of a statistically significant association.

Since aneuploidy incidence drastically increases in AMA [6, 7], age-related inter-cycle variability may constitute a serious limitation for the use of basal FSH in the assessment of aneuploidy risk. This is in line with a previous study suggesting that the association between basal FSH and aneuploidy becomes apparent only if the data are controlled for maternal age [25]. The present study, however, is the first designed to specifically assess the impact of maternal age on the relationship between basal FSH and aneuploidy, demonstrating that a robust association is restricted to pre-AMA patients. Considering that, although lower, the incidence of aneuploidy is far from negligible in pre-AMA patients (ranging from approximately 20 to 35% in women aged from 25 to 35 years; [6] and present study), basal FSH may still represent a valuable parameter for treatment decisions in pre-AMA IVF/ICSI cycles. Therefore, the present study provides important new references for the assessment of aneuploidy risk, indicating that the utility of single basal FSH measurements decreases as maternal age advances.

Previous evidence indicates that increased intrafollicular FSH signalling can predispose to aneuploidy. Firstly, elevated FSH levels in the follicular fluid are linked to decreased oocyte developmental competence and increased risk of aneuploidy in women [29, 45]. Interestingly, it has been recently demonstrated that sustained/increased FSH activity through eCG (equine chorionic gonadotrophin) treatment increases aneuploidy rates in midlife and young mice [46]. Although specific studies are still required to reveal the mechanisms linking FSH signalling and aneuploidy, we have recently integrated a large body of data strongly suggesting that excessive FSH activity can compromise meiotic accuracy and oocyte competence to support embryo development, at least in part, by reducing TZP (transzonal projections)-mediated cumulus-oocyte communication [22, 47]. This would reduce the delivery of cumulus-derived factors directly involved in the control of cytoskeleton dynamics and other molecules indirectly affecting meiosis through intracellular energetic metabolism and oxidative stress control [22, 48]. In line with this hypothesis, we have recently demonstrated in the bovine model that, through secreted paracrine factors, the oocyte suppresses FSHR (FSH receptor) transcription, while increasing that of AMH (anti-Müllerian hormone), a potent intra-ovarian paracrine inhibitor of FSH signalling, in cumulus cells [49]. It would thus appear that through paracrine regulation of cumulus cell gene expression, the oocyte might prevent meiotic errors induced by FSH, safeguarding its own developmental competence. Conversely, the present data corroborate all these previous findings suggesting that excessive FSH intrafollicular activity may compromise oocyte quality, which represents an important alert for medical decisions in current ICSI/IVF practice.

Our study is limited by its retrospective nature, and even if the interference of potential confounding variables was controlled for by a multivariate analysis, we cannot rule out the influence of other possibly non-identified variables unequally balanced across groups of patients with high and low basal FSH. Another potential limitation of our study is the predominance of AMA compared to pre-AMA embryos in the database, reflecting the prevalence of AMA in IVF/ICSI cycles applying PGT-A. We also recognise that the use of medians as basal FSH cut-offs may not constitute the most precise strategy to generate references to assist the decision of whether to perform PGT-A in pre-AMA patients, although it was sufficient to test/confirm the hypothesis of the study. In view of a potentially high variability of FSH basal levels and sensitivity to FSH activity across different patient populations, in-house determination of hormonal ranges associated with higher/lower aneuploidy incidence may maximise the accuracy of basal FSH as an assisting parameter for medical decisions (i.e. PGT-A performance) in ICSI/IVF practice. Lastly, we cannot rule out the possibility that the present data may have been influenced by PGT-A technical limitations leading to false positive or negative diagnoses of aneuploidy.

In conclusion, the present findings represent novel evidence that basal FSH values are significantly associated with aneuploidy occurrence in pre-AMA but not in AMA patients. Therefore, while shedding light on the utility of basal FSH as an indicator of aneuploidy risk in ICSI/IVF practice, the present data shall also contribute for a better understanding of the mechanisms underlying aneuploidy occurrence.

Data Availability

Anonymized data included in this study are available to the editors, reviewers and researchers upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.