Bisphenol A Alternatives in Follicular Fluid: Novel Risk Factors for Ovarian Reserve Dysfunction.

A total of 104 FF samples were collected, including 15 from patients with PCOS, 34 from patients with DOR, and 55 from control women with normal ovarian function. PCOS was diagnosed according to the Rotterdam criteria, which require the presence of oligomenorrhea, amenorrhea, or irregular uterine bleeding in combination with at least one of the following: (i) clinical or biochemical hyperandrogenism and/or (ii) polycystic ovarian morphology (PCOM) on ultrasound. DOR was defined by meeting all the following criteria: (i) AFC ≤ 7 on transvaginal ultrasound; (ii) serum anti-Anti-Müllerian hormone (AMH) < 1.1 ng/mL, irrespective of menstrual cycle phase; and (iii) basal FSH ≥ 10 IU/L measured on days 2–3 of the menstrual cycle. The control group serves as a well-established reference for normal ovarian physiology and FF microenvironment, a criterion widely adopted in high-quality studies in this field to ensure comparability. , It should be noted, however, that while this control group provides a standardized functional reference, it represents a specific subpopulation of infertility patients and may not fully reflect the general naturally fertile population. Therefore, the findings of this study are primarily applicable to infertile populations with similar clinical characteristics. Demographic characteristics, including age, smoking status, alcohol consumption, income, occupation, ethnicity, education level, physical activity, infertility status, and comorbidity, did not differ significantly among the three groups (all p > 0.05, Table ). A comparative analysis of ovarian reserve markers revealed distinct profiles across the groups ( Table ). Compared to the control group, the PCOS group exhibited significantly elevated levels of AMH, AFC, number of oocytes, and LH. The PCOS group also had a significantly higher BMI, a finding consistent with the well-established link between obesity and PCOS pathogenesis, where approximately 50% of patients are reported to be overweight or obese. This phenotype is often accompanied by a higher prevalence of metabolic comorbidities, such as hypertension and dyslipidemia, and parallels broader trends associating rising global obesity rates with increasing PCOS prevalence. , Note: AFC, antral follicle count; AMH, anti-Müllerian hormone; BMI, body mass index; DOR, diminished or decreased ovarian reserve; P, progesterone; PRL, Prolactin; T, Testosterone; E2, estradiol; FSH, follicle-stimulating hormone; IQR, interquartile range; LH, luteinizing hormone; SD, standard deviation. Variables that are not normally distributed are represented as M (IOR). Categorical variables are expressed as n (%). The p -value was calculated using the Kruskal–Wallis test for continuous variables and the chi-square (χ 2 ) test for categorical variables according to the data distribution. Difference and p < 0.05 when compared with the control group; Difference and p < 0.05 when compared with the DOR group. Fisher’s Exact Test correction As expected from the diagnostic criteria, the DOR group demonstrated a biomarker profile markedly distinct from both the control and PCOS groups, characterized by significantly higher basal FSH levels alongside significantly lower AFC and AMH levels. Furthermore, women in the DOR group were significantly older and yielded fewer oocytes. This pattern of elevated FSH, reduced AFC and AMH, advanced age, and lower oocyte yield is highly consistent with the clinical definition of DOR and aligns with findings from prior studies. , Table summarizes the detection rates and concentrations of 35 BPs analyzed in FF samples from all 104 participants. The total BPs concentrations ranged from 101.7 to 2944.5 ng/L. Among the analytes, BPS and BPA exhibited the highest detection rates (both 100%). The concentrations of BPA ranged from 2.2 to 435.0 ng/L, with a median concentration of 13.1 ng/L ( Table S6 ). Notably, BPS concentrations were consistently higher than BPA across all samples, with a median concentration of 64.7 ng/L (range: 33.3–2228.4 ng/L). The 100% detection rate of BPA in our Beijing cohort aligns with previous studies from China and South Korea ( Table S7 ), where detection rates often exceed 80%, but contrasts with reports from Russia and France ( Table S7 ) showing much lower rates (9.0%–28.6%). , , This geographic heterogeneity likely reflects regional differences in environmental exposure patterns, industrial activity, and regulatory policies. Although our detection rate is consistent with a prior Chinese study, both the median and mean BPA concentrations in our cohort were an order of magnitude lower. This discrepancy may result from temporal reductions in BPA exposure driven by evolving regulations or from differences in cohort characteristics. Another significant factor contributing to the substantial variation in reported BPA concentrations across different studies is the impact of sampling timing. Dietary intake constitutes the primary route of exposure to BPs, and BPA’s relatively short biological half-life of approximately 6 h results in considerable fluctuations in its levels within individual urine and blood samples over the course of a day. , Consequently, the timing of sample collection substantially influences the concentrations measured in these matrices. Although FF is anatomically isolated from systemic circulation by the blood-follicle barrier and exhibits low cellular density, factors that may mitigate the impact of short-term dietary variations, , the effect of sampling time on FF BPA concentrations remains unexplored. This gap primarily stems from the challenges of collection and ethical constraints. This constitutes a potential confounding factor in exposure assessment. Future studies utilizing animal models should focus on elucidating the temporal dynamics of BPs in FF. This would facilitate a more accurate assessment of its utility as a biological matrix for investigating the association between BPs exposure and reproductive health. nd: Below the detection limit is indicated as nd (not detected). -: The mean concentrations are not calculated for BPs with a detection rate less than 40%. : valid data. Intriguingly, BPS levels in our cohort not only exceeded those of BPA but also revealed distinct geographic patterns compared to prior reports. To date, only three studies have reported BPS concentrations in FF. Two French studies reported low detection rates (11.0%–13.0%), , contrasting sharply with a Chinese study showing a 97.3% detection rate. Our detection rate of 100% aligns with that observed in China. However, the BPS concentrations measured in our study were significantly higher than those in the prior Chinese study, while being more consistent with the levels detected in the French studies despite their lower detection rates. Furthermore, both the French studies and our findings consistently demonstrate that BPS concentrations exceed those of BPA in FF. , This pattern highlights distinct regional exposure profiles of BPS compared to BPA, reinforcing the potential displacement of BPA by BPS in certain populations. Similar substitution trends, where BPS emerges as a dominant or highly prevalent alternative, sometimes exceeding BPA levels, are being reported globally in other biological matrices. For instance, studies analyzing hair samples in Poland and urine samples in India support this emerging shift. Expanding beyond BPA and BPS, our study provides a comprehensive quantification of 21 BPA alternatives and 12 BPS alternatives in FF. We detected 11 BPA alternatives and 11 BPS alternatives among these compounds. Four of the 21 BPA alternatives exhibited detection rates exceeding 50%: BPAF (95.2%), BPF (57.7%), bisphenol Z (BPZ, 50.0%), and benzyl 4-hydroxybenzoate (PHBB, 50.0%). The mean and median concentrations of these prevalent BPA alternatives were 18.4–32.5 ng/L and 6.8–21.7 ng/L, respectively. Among the 12 BPS alternatives, two showed detection rates exceeding 50%: TGSA (63.5%) and dapsone (DDS, 51.0%). Their mean concentrations ranged from 4.0 to 22.5 ng/L, and their median concentrations ranged from 2.1 to 7.1 ng/L. These findings are significant, given the current scarcity of data on specific BPA and BPS alternatives in human FF samples. A French Assisted Reproductive Technology (ART) study, BPF was detected in only 1% of FF samples and no BPAF was detected, contrasting sharply with a Chinese study that quantified BPF and BPAF, reporting detection rates >50% for both BPF and BPAF and median concentrations of 10–100 ng/L, values broadly consistent with those observed in our study. Crucially, no prior studies have reported exposure to BPS alternatives in human FF. The detection of 11 BPS alternatives, including TGSA and DDS, with a prevalence of >50%, represents the first identification of these structural alternatives within this biological matrix. Previous research detected BPSIP, BPS-MAE and other alternatives primarily in urine and serum. , The maximum concentration observed for the BPS alternatives was 1326.6 ng/L. The high detection frequency and considerable concentrations of BPS alternatives indicate that these compounds are now widespread environmental contaminants with potential relevance to the ovarian microenvironment. The total BP concentration (Σ35BPs) exhibited a significant exposure gradient across the study groups: PCOS > DOR > Control group ( p = 0.048). Fourteen individual BPs showed statistically significant concentration differences among the groups ( p < 0.05; Table S6 and Figure ). The PCOS group demonstrated substantially higher detection rates, with 18 compounds detected in more than 50% of the PCOS samples, compared to 9 in the DOR group and 5 in the control group. Notably, the concentrations of BPA, BPS-MAE, DPS and BPAF were significantly elevated in the PCOS group compared to both the DOR and control groups. This aligns with prior reports of higher BPA and BPS levels in FF of PCOS patients. , Concentrations of 14 differential BPs in the PCOS, DOR, and control groups. Boxplots showing the distribution of 14 bisphenol compounds (BPs) across the three groups. The line within each box represents the median; the bottom and top edges represent the 25th and 75th percentiles, respectively; whiskers indicate the 5th and 95th percentiles. Statistical differences among groups were assessed using the Kruskal–Wallis test (* p < 0.05, ** p < 0.01, *** p < 0.001). In the DOR group, the median concentration of BPA was 21.0 ng/L, which was not significantly different from that in the control group. However, compounds such as BPS-MPE, DD-70, and TGSA showed significantly different concentrations than the control group. While epidemiological studies directly linking BPs exposure and DOR are scarce, the sole study measures BPA in 54 DOR patients and 67 non-DOR patients, and the results show that BPA levels in the FF of DOR patients were significantly higher than those of non-DOR patients (234.05 ng/L vs 193.30 ng/L, p < 0.01). Mechanistically, animal studies have indicated that exposure to BPB and BPAF led to changes in ovarian indices and a decrease in the number of follicles. BPs exhibit estrogenic activity, potentially disrupting hormonal regulation and impairing ovarian reserve function. , This provides biological plausibility for their potential role in DOR pathogenesis. To assess the relative contribution of individual BPs within each group, we calculated the proportional composition of detected BPs relative to the total BPs concentration per sample, expressed as percentage distributions ( Figure and Table S8 ). Across the control and DOR groups, the aggregate compositional profile consistently ranked as follows: BPA alternatives > BPS > BPS alternatives > BPA. The PCOS group exhibited a minor difference in ranking: BPA alternatives > BPS alternatives > BPS > BPA. Collectively, these profiles demonstrate that the combined contributions of BPA alternatives and BPS alternatives substantially exceed those of BPA itself. This shift strongly reflects the impact of BPA restrictions, which have driven the widespread production and use of alternative compounds. Further analysis identified BPS, DPS, and BPA as the compounds consistently contributing the highest individual proportions. However, the primary contributor differed by group: in the control groups, BPS had the highest proportion, followed by BPA and DPS; in the PCOS group, DPS predominated, followed by BPS and BPA; and in the DOR group, BPS was the primary contributor, followed by BPA and BPF. DPS, a derivative of BPS used in products such as thermal paper and currency, exemplifies this trend. Its prominence aligns with biomonitoring data from Chinese pregnant women, where DPS emerged as the second most prevalent BPs after BPA. Relative composition of individual BPs contributing to total BPs across the PCOS, DOR, and control groups. The Sankey diagram illustrates the proportional contribution of each bisphenol compound to the total bisphenol concentration (ΣBPs) in the three groups. Each color represents a distinct bisphenol, and node height corresponds to 100% of ΣBPs in each group. Flow width reflects the relative abundance of each compound, highlighting differences in exposure profiles among reproductive phenotypes. Spearman’s correlation analysis was performed to examine the relationships between BPs with detection rates exceeding 50% and the total BP concentration. The compounds analyzed included BPA, BPS, BPAF, BPF, PHBB, BPZ, TGSA, and DDS. As shown in Figure , except for DDS, most BPs exhibited positive correlations (r = 0.20–0.61, p < 0.05). Specifically, BPA was significantly positively correlated with BPAF (r = 0.25, p < 0.01) and TGSA (r = 0.50, p < 0.01). Similarly, BPS showed a strong positive correlation with TGSA (r = 0.39, p < 0.001). In addition, BPZ was positively correlated with PHBB (r = 0.35, p < 0.001) and TGSA (r = 0.39, p < 0.001), while the correlation between PHBB and TGSA was also highly significant (r = 0.43, p < 0.01). These interrelationships collectively suggest a co-occurrence pattern among these compounds. Spearman correlation analysis between individual and total BPs. Correlation coefficients are represented by the size of the circle and the intensity of the color, with red indicating positive associations and blue indicating negative associations. Asterisks denote statistical significance (*: p value <0.05. **: p value <0.01. ***: p value <0.001). Interestingly, although BPS, a common alternative for BPA, did not show a significant correlation with BPA, it demonstrated the strongest positive correlation with Σ35BPs (r = 0.61, p < 0.001). This finding indicates that BPS, as a replacement for BPA, has emerged as a predominant BP congener in FF, consistent with reported occurrence patterns in food [Zhang, China CDC Weekly]. , Moreover, significant positive correlations were observed between Σ35BPs and several individual BPs (BPAF, BPS, BPZ, BPF, PHBB, and TGSA), reinforcing that overall bisphenol exposure is not driven by a single compound, but rather results from the combined contribution of both conventional and emerging bisphenols. As a comprehensive metric, Σ35BPs effectively captures the aggregate burden of mixed bisphenol exposure. The results of this study provide foundational insights into the mixed bisphenol exposure profile in human FF. Given that FF directly bathes developing oocytes, the observed coexposure patterns may hold toxicological relevance, potentially influencing reproductive health via endocrine-disrupting mechanisms. , To further elucidate the impact of BPs on ovarian function, we examined their associations with key ovarian reserve markers, including AMH, AFC, T and oocytes. Existing literature suggests that elevated BPA levels are associated with both PCOS and DOR, despite the distinct clinical profiles of these conditions. PCOS is typically characterized by increased oocyte counts, elevated AMH, normal or low FSH, increased AFC and hyperandrogenism. In contrast, DOR is marked by reduced oocyte counts, decreased AMH, elevated FSH and diminished AFC. Therefore, we categorized each ovarian reserve marker into three ordinal levels (0, 1, and 2), representing below-normal, normal, and above-normal ranges, respectively. Ordinal logistic regression models were employed with these categorized markers as dependent variables and the concentrations of BPs as independent variables. Odds ratio (OR), 95% confidence intervals (CI), and p-values are summarized in Table S9 , with statistically significant associations ( p < 0.05) visualized in Figure . Logistic regression analysis of BPs and ovarian reserve markers. The forest plot presents odds ratio (ORs) and 95% confidence intervals (CIs) for the associations between bisphenol compound (BPs) levels and ovarian reserve markers. Only statistically significant associations ( p < 0.05) are displayed. The plot illustrates the relationships between BPs and ovarian reserve markers, along with their corresponding confidence intervals. For AMH, DDS and PHBB with increasing concentrations, raising the probability of AMH levels falling outside the normal range (i.e., either below [Category 0] or above [Category 2]). In contrast, BPA demonstrated a consistent and significant positive association with AMH, with higher exposure increasing the probability of levels exceeding the normal range in both univariable and multivariable models (OR = 10.01, p < 0.05). Regarding AFC, elevated concentrations of BPAF, PHBB, and Σ35BPs were significantly associated with an increased probability of AFC exceeding the normal range, with BPAF showing the strongest effect (OR = 11.91, p = 0.015). For oocyte count, BPA was the only compound significantly associated in both models (OR = 5.08–7.74, p < 0.05), with higher exposure increasing the likelihood of above-normal oocyte yield. However, after FDR correction, the association between BPA and oocyte number was no longer statistically significant ( p = 0.184). In contrast, the associations of BPA, DDS, BPAF, PHBB, and Σ35BPs with abnormal AMH and AFC levels remained statistically significant following FDR adjustment. These findings indicate that the potential biological relevance of these compounds warrants further investigation. Overall, these findings indicate that exposure to BPA and its alternatives correlates with alterations in ovarian reserve markers. Previous studies have suggested that BPA exposure may disrupt ovarian homeostasis and contribute to heterogeneous reproductive outcomes such as PCOS. , However, due to the limited sample size in the present study ( Text S2 and Table S4 ), we can only demonstrate an association between exposure to certain BPA alternatives, such as PHBB, BPAF and markers of ovarian reserve function. These results should not be extrapolated to infer a link with specific reproductive disorders; further validation through larger epidemiological studies and mechanistic investigations is warranted. The observed associations may be explained by the documented disruptive effects of bisphenols on ovarian cellular function. BPA and its alternatives (e.g., BPAF) are known to impair granulosa cell activity. Specifically, BPA disrupts steroidogenesis and inhibits follicular development. Similarly, BPAF and other alternatives at environmentally relevant concentrations may bind to the follicle-stimulating hormone receptor (FSHR), leading to aberrant estrogen synthesis. − In theca cells, BPA downregulates key receptors, including FSHR and the luteinizing hormone receptor (LHR), , thereby promoting follicular atresia and reduced ovarian reserve. Furthermore, BPA and its alternatives induce reactive oxygen species (ROS) generation, activating apoptotic pathways in theca cells. , Combined exposure to multiple bisphenols may lead to complex cumulative or antagonistic effects. Notably, low-dose bisphenols may transiently stimulate AMH secretion via estrogen-mimicking effects, whereas higher doses typically exert toxic impacts. , In summary, BPs likely impact ovarian reserve through multiple interconnected mechanisms, including disruption of granulosa cell receptor signaling, induction of oxidative stress, and epigenetic alterations. , However, current evidence is largely derived from in vitro studies and cross-sectional human data, underscoring the critical need for longitudinal studies to establish causality and clarify dose–response relationships.

Standards of the 35 target BPs (purity >98%) and their isotopically labeled internal standards (ISs) were purchased from Sigma-Aldrich (St. Louis, MO, USA), Alta Scientific Co., Ltd. (Tianjin, China), J&K Scientific (Beijing, China) and Dr.Ehrenstorfer GmbH (Augsburg, Germany). The detailed information is presented in Table S1 . Liquid chromatograph–mass spectrometer (LC-MS) grade methanol, acetonitrile, acetic acid and ammonium fluoride with purity above 99% were obtained from Dikma Technologies Inc. (Beijing, China). Ammonium acetate was obtained from J&K Scientific (Beijing, China). β-Glucuronidase/arylsulfatase ( Helix pomatia ) was obtained from Sigma-Adlrich (St. Louis, MO, USA). Ultrapure water was generated by a Milli-Q Gradient system (Millipore, Bedford, MA, USA). This study collected FF samples from patients at the Department of Human Reproductive Medicine, Beijing Maternity Hospital, Beijing, China, between July and September 2024. The study protocol received approval from the Ethics Committee of Beijing Maternity Hospital, Capital Medical University (Approval No. 2023-KY-097-01). Written informed consent was obtained from all participants prior to enrollment. All diagnoses were confirmed by board-certified reproductive medicine specialists in accordance with internationally recognized guidelines. The inclusion criteria were as follows: (1) age between 20 and 50 years; and (2) availability of complete data on FF analysis, serum hormone levels and oocyte count. Exclusion criteria included: (1) comorbidities that could influence androgen levels or ovulatory function, such as congenital adrenal hyperplasia, thyroid dysfunction, or hyperprolactinemia; (2) known genetic or chromosomal abnormalities affecting ovarian function; (3) history of ovarian surgery or use of hormone medication within the past six months; (4) incomplete data on FF, serum hormone levels, or oocyte; or (5) occupational exposure to BPs. The control group comprised women with confirmed normal ovarian reserve undergoing in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) due to tubal or male factor infertility. Eligible participants completed standardized questionnaires that included information on education level, income, and occupation. Serum samples were collected for hormone analysis, including follicle-stimulating hormone (FSH), luteinizing hormone (LH), estradiol (E2), progesterone (P), testosterone (T), and prolactin (PRL) levels. FF samples were collected for the analysis of BPs, obtained from each participant’s largest follicle before ovulation using a follicular aspiration needle guided by ultrasound. Samples were immediately centrifuged, and the supernatant was aliquoted and stored at −80 °C. Antral follicle count (AFC) (follicles 2–8 mm in diameter) was assessed by transvaginal ultrasound on menstrual cycle days 2–3. Detailed protocols for hormone assays, FF collections and AFC assessment have been described in our previous study. To further minimize potential batch effects during sample processing, all collected FF samples will be transported to the analytical laboratory in a single shipment. Upon arrival, quantification of BPs was completed within three-days. An aliquot of 0.5 mL of FF was spiked with 1 ng of mixed IS and equilibrated at room temperature for 30 min. Sodium chloride (0.25 g) was added, followed by liquid–liquid extraction with 0.5 mL acetonitrile. The samples were vortex-mixed for 30 s, sonicated in a water bath for 15 min, and centrifuged at 14,000 rpm for 15 min at 4 °C. The acetonitrile layer was then collected. Enzymatic hydrolysis was performed by adding β-glucuronidase/arylsulfatase to an acetic acid/ammonium acetate buffer (1:100, v/v). After hydrolysis for at least 4 h ( Figure S1 and Text S1 ), the supernatant was evaporated to near dryness under a gentle nitrogen stream at 35 °C. The residue was reconstituted in 200 μL of methanol/water (2:3, v/v) and centrifuged at 14,000 rpm for 5 min at 4 °C. The supernatant was analyzed by ultraperformance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Chromatographic separation was performed using a Waters ACQUITY UPLC system (Waters, Milford, MA, USA) equipped with a CORTECS UPLC C18 column (100 mm × 2.1 mm, 1.6 μm; Waters, Milford, MA, USA). Mobile phases consisted of 1 mmol/L ammonium fluoride aqueous solution (A) and methanol (B) at a flow rate of 0.25 mL/min, with the column temperature maintained at 40 °C. A 5 μL injection volume was employed with the following gradient program: 0–0.5 min, 35% B; 0.5–1.5 min, 35% B–50% B; 1.5–5.0 min, 50%–100% B; 5.0–7.0 min, 100% B; 7.0–7.1 min, 100% B–35% B; 7.1–9.6 min, 35% B. Quantitative analysis was performed on a QTRAP-6500 mass spectrometer (SCIEX, California, USA) with electrospray ionization (ESI) in positive and negative ion modes. Phenyl sulfone (DPS) and 1-Prop-2-enoxy-4-(4-prop-2-enoxyphenyl)­sulfonyl-benzene (BPS-DAE) were monitored in positive mode, while all other compounds were observed in negative mode. Multiple reaction monitoring (MRM) was employed. The ion source parameters were optimized as follows: desolvation temperature, 500 °C; auxiliary gases (GS 1 and GS 2), 55 psi; curtain gas, 35 psi; and collision gas, medium. The electrospray voltages were set at +5500 V for the positive mode and −4500 V for the negative mode. The other mass spectrometry parameters are listed in Table S2 . Rigorous quality control measures were implemented throughout the sampling and analyses. Before sample collection, five unused oocyte retrieval needles from the same batch were randomly selected. Following the standard operating procedures, these needles were used to aspirate ultrapure water samples. The aspirated ultrapure water was subsequently analyzed. During sample processing, procedural blanks (ultrapure water alternatived for FF), solvent blanks (reconstitution solution directly injected), and matrix-spiked samples (blank FF spiked at 100 ng/L) were concurrently analyzed. The method detection limit (LOD) and quantification limit (LOQ) were estimated at signal-to-noise (S/N) ratios of 3 and 10, respectively, based on the matrix-spiked sample, yielding LODs of 0.1–18.2 ng/L and LOQs of 0.3–60.0 ng/L for the 35 analytes. The mean recoveries across the three spiking levels ranged from 77.1% to 107.7%, with relative standard deviations (RSD) <11.6% ( Table S3 ). The concentrations of BPs detected in the ultrapure water aspirated by the needles, as well as in the procedural blanks and solvent blanks, were consistently below the LOD. Demographic, clinical, and lifestyle covariates were obtained from electronic health records and structured questionnaires. These variables included: (1) Demographic characteristics, comprising age (categorized as <35 years and ≥35 years), body mass index (BMI, kg/m 2 ), classified using World Health Organization criteria (24.0), and current tobacco use; and (2) Reproductive indices, including duration of infertility, recorded in specific years. All covariates were systematically collected and verified to ensure data completeness and accuracy. The normality of the continuous variables was assessed using the Shapiro-Wilk test. Normally distributed variables were presented as mean ± standard deviation (SD), and intergroup differences were analyzed using one-way ANOVA. Non-normally distributed variables were presented as medians and interquartile ranges (IQRs), with intergroup differences assessed using the Kruskal–Wallis test. Categorical variables are expressed as frequencies (n, %) and compared using the chi-square test. For BPs concentrations, values below the LOQ but above the LOD were assigned the numerical value reported by the analytical method. Values below the LOD were imputed using the regression on order statistics (ROS) method implemented in ProUCL (version 5.1; U.S. Environmental Protection Agency). Concentrations below the LOD for BPs with a detection rate lower than 40% were replaced with zero. BPs with a detection rate below 50% in the overall sample were excluded from subsequent analyses. Spearman rank correlation analysis was used to test the correlation between various BP concentrations, and the results are presented as Spearman correlation coefficients (r) and their corresponding p-values. Due to the limited sample size in the polycystic ovary syndrome group, the statistical power was insufficient. We performed a sample size calculation ( Table S4 and Text S2 ) to provide a reference for future studies. Subsequently, ordinal logistic regression was used to explore the association between BP concentration and key ovarian reserve markers. Ovarian reserve markers were categorized according to medical reference values into ″below normal range, within normal range, and above normal range″ and assigned values of 0, 1, and 2 as dependent variables. Independent variables included various BP concentrations and potential confounding factors ( Table S5 ). False discovery rate-adjusted p -values were obtained from multivariable regression models using the Benjamini–Hochberg procedure. The association between BPs concentration and ovarian reserve markers was quantified by calculating the odds ratio (OR) and its 95% confidence interval (CI). The experimental flowchart is shown in Figure . All statistical analyses were performed using two-tailed tests, and p < 0.05 was considered statistically significant. Data processing and visualization were primarily performed in R software (version 4.5.1), with supplementary analyses conducted in SPSS (version 31.0, USA). Experimental flowchart. This figure illustrates the recruitment process and participant classification for patients from Beijing Maternity Hospital between July and September 2024. Inclusion and exclusion criteria were applied to identify eligible participants, who were subsequently categorized into three groups: polycystic ovary syndrome (PCOS), control group, and diminished ovarian reserve (DOR). The lower portion of the diagram outlines the procedures for sample and data collection, measurement of bisphenol compounds (BPs) in follicular fluid (FF), and the statistical analyses conducted.

This study uses FF as a biomonitoring matrix for assessing BPs exposure, overcoming limitations of traditional blood/urine analyses. Compared to urine and serum, where BPs concentrations are lower and exhibit substantial interindividual variability due to metabolic clearance differences, FF directly reflects the ovarian microenvironment. Its proximity to developing oocytes provides more biologically relevant exposure assessment, offering superior evidence for BP-reproductive disease associations. Our comprehensive analysis covered 35 BPs, including the first reported detection of several novel alternatives in human FF. The limitations of this study primarily include: (1) The relatively small sample size and single-center recruitment may limit the generalizability of the findings and preclude in-depth exploration of underlying mechanisms, mixed exposure effects, or potential temporal variability; (2) Owing to ethical constraints, the control group consisted of women with infertility due to tubal or male factors rather than completely healthy individuals; therefore, the results should be interpreted primarily within the context of infertile populations with similar clinical characteristics and not generalized to the general healthy population; (3) Despite implementing stringent exclusion criteria, the possibility cannot be entirely excluded that unidentified subclinical conditionssuch as mild autoimmune abnormalities or undetected metabolic disordersmight subtly influence the FF microenvironment, representing another potential source of bias. These constraints should be addressed through the expansion of multicenter cohorts in future research.

This study revealed a high prevalence and diversity of BPs in human FF, including the first identification of several BPA and BPS alternatives in this biological medium. Total BPs were significantly higher in patients with PCOS than in the control group. Furthermore, BPs levels in FF were significantly and positively correlated ( p < 0.05) with abnormally elevated ovarian reserve markers, including AFC, and oocyte. These findings collectively highlight the significant impact of BPs exposure on ovarian reserve dysfunction. Given the widespread environmental presence of BPs, further research is urgently needed to elucidate the long-term reproductive consequences of both chronic and mixed exposures.

Bisphenol A (BPA) is a well-known environmental endocrine disruptor (EDC) that has raised significant concern due to its ubiquitous use and estrogen-mimicking properties. In recent years, growing recognition of the toxicity of BPA has led to increasingly stringent regulatory restrictions worldwide. For instance, the tolerable daily intake (TDI) for BPA has been progressively lowered from an initial 50 μg/kg body weight (bw)/day to the current level of 0.2 ng/kg bw/day as of 2023. Many countries and regions have also phased out BPA in infant bottles and other food contact materials. These regulatory measures have accelerated the production and application of BPA alternatives. Among these alternatives, bisphenol S (BPS) is extensively used in plastics, dyes, pharmaceuticals, and pesticide intermediates. − The global market demand for BPS continues to grow, with its market size reaching approximately USD 750 million in 2022, representing a 13% year-on-year increase. These alternatives are structurally analogous to BPA and exhibit similar endocrine-disrupting effects. In vitro and in vivo studies have revealed that bisphenol AF (BPAF), bisphenol B (BPB) and bisphenol FL (BPFL) demonstrate stronger estrogenic effects than BPA, − whereas BPS exhibits more potent antiandrogenic activity. , Notably, BPS is not regarded as a safe alternative and has been classified by the U.S. Environmental Protection Agency as a substance of very high concern (SVHC). This has prompted the adoption of other BPS alternatives, such as 4-((4-(Allyloxy)­phenyl)­sulfonyl)­phenol (BPS-MAE) and 4-hydroxy-4’-isopropoxydiphenylsulfone (BPSIP), which are now widely used in the manufacture of thermal paper and related products. In vitro evidence suggests that 4-[(4-benzyloxyphenyl)­sulfonyl]­phenol (BPS-MPE) and 2-[(4-hydroxyphenyl)­sulfonyl]­phenol (2,4-BPS) act as specific antagonists of estrogen receptor alpha in HepG2 cells, while BPSIP and 4,4′-Sulfonylbis­(2-allylphenol) (TGSA) modulate estrogen-responsive genes in avian models. BPSIP has also shown estrogenic activity comparable to that of BPS in zebrafish. Therefore, BPA, its alternatives and BPS alternatives can collectively disrupt female endocrine homeostasis via estrogenic activity or antiandrogenic effects, potentially leading to disorders of the female reproductive system. Follicular fluid (FF) constitutes the microenvironment surrounding oocytes and granulosa cells, providing a more direct reflection of chemical exposure at the ovarian level than blood or urine. It serves as an accurate ‘snapshot’ of the reproductive microenvironment. , Studies have shown that BPA concentrations in FF are 2–5 times higher than those in serum, underscoring the value of FF as a biomonitoring matrix of ovarian-specific exposure. Epidemiological studies have established BPA exposure as a significant environmental risk factor for both polycystic ovary syndrome (PCOS) and diminished ovarian reserve (DOR). Although PCOS and DOR represent distinct pathological states of ovarian dysfunction, with PCOS characterized by follicular developmental arrest and endocrine-metabolic disturbances, and DOR primarily defined by an accelerated depletion of the ovarian follicular pool. Both conditions ultimately lead to impaired female fertility. Critically, BPA concentrations in FF are significantly elevated in women with PCOS and DOR compared to controls, , suggesting a potential role for follicular exposure in ovarian dysfunction. However, despite the reported occurrence of BPA in FF across multiple countries, data on BPA alternatives remain scarce. To date, only three studies have reported BPA alternatives in FF: Lebachelier de la Riviere et al. (2023b) detected conjugated forms of BPS, BPF, and BPA; Li et al. (2025) reported the presence of Bisphenol BP (BPBP) together with BPA, BPAP, BPS, and BPAF; and more recently, Lebachelier de la Riviere et al. (2025) identified BPS, BPF, BPAF and BPA. Moreover, no study to date has investigated the levels of BPS alternatives in this biologically critical matrix and evidence linking BPA alternatives to ovarian disorders remains quite limited. Therefore, this study aims to 1) determine the concentrations of BPA, BPS, 21 BPA alternatives, and 12 BPS alternatives (collectively referred to as bisphenols, BPs) in human FF samples collected from women in Beijing, China, 2) characterize female exposure profiles to BPs, and 3) explore their potential associations with ovarian dysfunction.