Correlation between AMH levels and embryonic aneuploidy rate in PGT-A patients: a retrospective study

Abstract

Objective

The association between anti-Müllerian hormone (AMH) levels and embryonic aneuploidy rates was investigated by analyzing clinical and embryo laboratory data from patients with preimplantation genetic testing for aneuploidy (PGT-A). However, the nonlinear relationship and threshold effect of AMH on aneuploidy risk remain poorly understood.

Methods

This retrospective study analyzed the clinical data of 819 PGT-A cycles performed between January 2018 and August 2024 at the General Hospital of Northern Theater Command. We used restricted cubic spline (RCS) to investigate the dose–response relationship between the AMH levels and aneuploidy rate, adjusting for potential confounders.

Results

Significant differences were observed in normal fertilization rates, day 3 high-quality embryo rates, blastocyst formation rates, euploidy embryo rates, aneuploid embryo rates, and mosaic embryo rates among the three AMH groups (P < 0.05). A statistically significant nonlinear relationship between AMH levels and aneuploidy rate was identified (P < 0.05). RCS and threshold effect analyses revealed that the risk of a positive (≥ 50%) aneuploidy rate increased by 40% for each 1-unit decrease in AMH when AMH ≤ 2.54 ng/mL.

Conclusions

In the PGT-A population, advanced maternal age (AMA), recurrent spontaneous abortion (RSA), or recurrent implantation failure (RIF) have been identified as contributing factors. After adjusting for potential confounders such as female age, AMH remains a significant risk factor for embryonic aneuploidy rates. The findings suggested that lower AMH levels are associated with a higher risk of embryonic aneuploidy, indicating that ovarian reserve function may be correlated with oocyte quality. These results provide new insights for individualized decision-making in assisted reproduction.

Trial registration ChiCTR2500099710 (03/27/2025).

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-025-03619-x.

Introduction

Embryonic aneuploidy is a key biological mechanism causing early pregnancy loss, with an incidence ranging from 45 to 83% among embryos generated through assisted reproductive technology [1–3]. Notably, embryonic aneuploidy arises mainly from maternal meiotic errors [4, 5], and its incidence increases exponentially with advancing maternal age [6, 7]. In women over 35 years old, the loss of adhesins and the decline of spindle monitoring function lead to a significant increase in meiotic chromosome segregation errors [8, 9], which is further associated with a reduction in ovarian reserve function (manifested as decreased anti-Müllerian hormone (AMH) levels) [10, 11]. This suggests a dual trend of decreasing oocyte quality and quantity with age.

In response to this clinical challenge, preimplantation genetic testing for aneuploidy (PGT-A) has significantly improved clinical pregnancy and live birth rates by enabling the screening for aneuploid embryos [12–14]. However, invasive biopsy procedures may impair embryonic developmental potential [15, 16] . Prior studies have indicated that male factors exert a negligible influence on embryonic haploidy. However, some studies have shown that DFI levels are significantly higher in patients with severe dyszoospermiathan in the general population. This leads to higher rates of blastocyst aneuploidy [17–21]. The present study was conceived with the objective of elucidating the independent role of AMH and excluding patients with severe teratozoospermia. Therefore, a critical need for non-invasive assessment methods for oocyte quality is necessary. AMH, as an ovarian reserve marker [22, 23], reflects the size of the primordial follicular pool [24–26], and its reduced level essentially reflects a decrease in the number of follicles [11, 27]. There is a synchronized decline in oocyte number and quality with age [28–30]. This overlap in biological characteristics makes the correlation between AMH and embryonic aneuploidy rate susceptible to interference by age confounding effects. If it is confirmed that AMH can independently predict aneuploidy risk, PGT-A can be recommended in advance for young patients with very low AMH to increase the chances of a successful pregnancy.

To address the above contradictions, the present study adopts a novel approach through dose–response relationship analysis. It seeks to determine whether the prediction of embryonic aneuploidy based on AMH is independent of age and whether there exists a threshold effect of AMH levels on embryo ploidy. By systematically analyzing clinical and embryonic data of patients undergoing PGT-A, this study aims to provide a theoretical basis for optimizing the oocyte quality assessment system and offers new insights into individualized assisted reproductive technology (ART) development.

Methods

Study population

This study involved the collection of clinical data from patients who underwent PGT-A at the Center for Reproductive Medicine of the General Hospital of the Northern Theater of Operations from January 2018 to August 2024. The patients were selected based on the indications of advanced maternal age (AMA), recurrent spontaneous abortion (RSA), or recurrent implantation failure (RIF). A retrospective analysis was conducted, and the flowchart of this study is shown in Fig. 1.

Fig. 1.

Flowchart of the study on the correlation between AMH levels and embryonic aneuploidy rate

Inclusion criteria: Patients fulfilling any of the above criteria were qualified for inclusion.  (1) AMA (female age ≥ 38 years old); (2) RSA (defined as ≥ 2 spontaneous abortions); (3) RIF (defined as ≥ 3 transfers or a cumulative total of ≥ 4–6 high-quality embryos or ≥ 3 high-scoring blastocysts still not implanted).

Exclusion criteria: (1) chromosomal abnormality in either spouse; (2) history of ovarian endometriosis or ovarian surgeries; (3) endocrine or autoimmune diseases (e.g., diabetes mellitus and systemic lupus erythematosus); (4) use of egg donation; (5) severe teratospermia in the male partner (defined as having <1% normal sperm morphology); (6) incomplete core data in the medical record.

Ovulation regimen

Depending on the woman’s age and ovarian reserve, an appropriate ovarian stimulation regimen is used for ovulation or natural cycle egg retrieval. Ovulation induction regimens include agonist regimens [gonadotropin-releasing hormone (GnRH) agonist long regimen, GnRH agonist extra-long regimen, GnRH agonist short regimen], GnRH antagonist regimens, and other regimens. Other regimens included the use of recombinant follicle-stimulating hormone (recombinant follicle-stimulating hormone, r-FSH, Jinseheng, Changchun Jinsai Pharmaceutical Co. Ltd; gonadotropin, Merck Serono, USA)/urine-derived human menopausal gonadotropin (HMG, Lebold, Lizhu Pharmaceutical Factory, Zhuhai Lizhu Group) direct stimulation, microstimulation, luteal phase ovulation, and natural cycle egg retrieval regimens. The Gn initiation dose was determined based on the woman’s age, AMH, basal antral follicle count (bAFC), body mass index (BMI), and ovarian response in previous cycles. When at least one follicle was ≥ 18 mm in diameter, follicle maturation was induced by injection of 250 µg of recombinant human chorionic gonadotropin (r-hCG, Azer, Merck Serono Ltd., Switzerland). Egg retrieval was performed 24–36 h after r-hCG injection.

Culture of embryos

Mature oocytes at the metaphase II (MII) phase were fertilized in vitro using intracytoplasmic single sperm microinjection technique. The fertilized oocytes were initially cultured in oviduct fluid culture (ART-1026) for 60–72 h, followed by transfer to blastocyst oviduct fluid culture (ART-1029) for an additional 48–72 h.

Blastocyst biopsy

A morphologic evaluation system for oocytes and early embryos was used for embryo culture. Embryos with 6–10 cleavage cells on day 3, classified as grade 1 or 2, were considered high-quality embryos. Blastocysts were assessed according to the Gardner scoring criteria, and trophoblast cell biopsies were performed on blastocysts with a score of ≥ 4 BB. A laser was used to create a hole of approximately 50 µm in diameter at the contralateral position of the inner cell mass of the blastocyst to be biopsied. Six to eight trophoblast cells were aspirated from the biopsy pins into the notch and detached from the blastocysts using both laser and mechanical methods. The biopsied cells were transferred to PCR tubes containing phosphate-buffered saline and stored in an ice box for further examination. PGT-A was performed by whole genome SurePlex amplification of the biopsied trophoblast cells. Amplification and library construction were conducted using the PGT-A kit (Beijing Zhongyi Kangwei Medical Equipment Co., Ltd.). Sequencing was performed using the Illumina sequencing platform MiSeqDx sequencer, and the sequencing data were analyzed for chromosome copy number using the Embryonic Chromosome Alloploidy Analysis System software (Beijing Zhongyi Kangwei Medical Equipment Co., Ltd.). The analysis process involved quality control of the sequencing data, followed by comparison to the human reference genome (GRCH37/hg19). The whole reference genome was divided into multiple windows, and the number of reads in each window was calculated. Copy number analysis was performed by bioinformatics algorithm through the data correction, combined with the reference range established by the reference dataset. The resolution of the detection was 4 Mb. When the proportion of 4 Mb and above copy number variation and chromosome abnormality was greater than 80%, the blastocyst was judged to be aneuploid. If the proportion of abnormality was between 30 and 80%, the blastocyst was recognized as a chimeric blastocyst.

Frozen embryo transfer (FET) protocol

After obtaining a euploid embryo, the endometrial preparation protocol prior to FET is selected based on the patient’s age, endometrial thickness, and the presence or absence of comorbid adenomyosis. The endometrial preparation regimen mainly consists of natural cycle regimen, estrogen replacement cycle regimen, GnRH-a down regulation plus estrogen replacement cycle regimen, and ovulation induction cycle regimen. The endometrial thickness, estrogen, and progesterone levels of all transfer programs must meet the criteria for transfer before the frozen-thawed embryo transfer procedure is performed. All types of transfer programs require pharmacological luteal support, the dose of which is adjusted according to the levels of estrogen and progesterone, and the luteal support is continued after the transfer. If pregnancy is confirmed, the original dosage will be maintained until the 8th week of pregnancy, and then the dosage will be gradually reduced until the 12th week of pregnancy.

Clinical indicators

The following clinical indicators were recorded: female age, male age, BMI, years of infertility, number of cycles, basic follicle-stimulating hormone (bFSH), basic luteinizing hormone (bLH), basic estradiol (bE2), AMH, bAFC, and total gonadotropin (Gn) dose.

Measurement of indicators

AMH levels were measured using venous blood collected at any point during the menstrual cycle in the year prior to ovulation induction. bFSH, bLH, bE2 were measured from venous blood collected on days 2–5 of the natural menstrual cycle. bAFC was determined by transvaginal ultrasonography on days 2–5 of the menstrual cycle, with follicles ranging in size from 2 to 8 mm considered antral follicles, and the total count of these follicles was recorded. BMI was calculated as body weight (kg) divided by height (m)². The euploidy rate was calculated by dividing the number of euploid blastocysts by the number of biopsied blastocysts, multiplied by 100%. Similarly, the aneuploidy rate was calculated by dividing the number of aneuploid blastocysts by the number of biopsied blastocysts, multiplied by 100%. The mosaic rate was calculated by dividing the number of chimeric blastocysts by the number of biopsied blastocysts, multiplied by 100%. The fertilization rate was calculated by dividing the number of normally fertilized oocytes by the total number of eggs obtained, multiplied by 100%. The day 3 high-quality embryo rate was calculated by dividing the number of high-quality embryos on day 3 by the number of embryos with two pronuclei (2PN), multiplied by 100%. The blastocyst formation rate was calculated by dividing the number of blastocysts formed on day 5 or day 6 by the number of embryos with 2PN, multiplied by 100%. The diagnosis of a live birth is made when the gestational age of the pregnancy is 28 weeks or more, the fetal weight is greater than 1000 g, and the fetus exhibits vital signs following delivery.

Statistical analysis

Patients who met the criteria were divided equally into three groups based on their AMH values by percentile, using the 25th percentile and 75th percentile as cutoffs. Continuous variables were tested for normality using the Shapiro–Wilk test. Normally distributed variables were expressed as mean ± standard deviation (SD), while skewed continuous variables were expressed as median and interquartile range (IQR). Categorical variables were expressed as frequencies and percentages (%). To assess differences between groups, the chi-square test was used for categorical variables, ANOVA for normally distributed continuous variables, and the Kruskal–Wallis test for skewed continuous variables. Potential confounders were identified based on the following criteria: a P-value < 0.1 in univariate analysis or a change of more than 10% in the effect size in covariate screening. Variables with covariates related to AMH were excluded from multifactor regression analysis. Univariate regression analyses were performed on the remaining variables to identify potential factors influencing the embryonic aneuploidy rate. Multifactorial logistic regression analysis was performed with aneuploidy rate as the dependent variable (with a rate ≥ 0.5 defined as a positive event) and AMH as the independent variable. Variables with P-values < 0.1 and changes in effect size greater than 10% during covariate screening were included in the multifactorial model using forward selection. Subsequently, variables with P-values < 0.05 from the multifactor logistic regression were included in the regression model for RCS, with nodes at specific percentiles (5th, 35th, 65th, and 95th). The results were expressed using odds ratios (OR) and their 95% confidence intervals (CI). A two-stage logistic regression model was used to analyze the threshold between AMH levels and aneuploidy rates. Inflection points were identified using likelihood ratio tests and bootstrap resampling methods. To assess the stability of the relationship between AMH and embryonic aneuploidy across populations, stratification analyses were performed based on the following subgroup variables: female age (< 35 years, 35–37 years, > 37 years), BMI (15–24 kg/m², > 24 kg/m²), infertility type (primary infertility, secondary infertility), and number of years of infertility (< 2 years, 2–4 years, > 4 years), PGT-A factors (AMA, RSA, RIF); heterogeneity between subgroups was assessed by multivariate logistic regression, and interactions between subgroups and AMH were assessed using the likelihood ratio test. Univariate and multivariate logistic regression analyses of live birth outcomes were performed using the same methods previously described. A subject operating characteristic (ROC) curve was established to explore the predictive ability of AMH values for live birth in euploid embryo transfer. Missing data were addressed using multiple interpolation methods, following the approach outlined by Van Buuren and Groothuis-Oudshoorn (2011). All analyses were performed using R statistical software (http://www.R- project.org, the R Foundation) and the Free Statistics analysis platform, and differences were judged to be statistically significant at P < 0.05 (two-tailed).

Results

Participant selection

A total of 1222 cycles from January 2018 to August 2024 were initially considered for this study. Of these, 403 cases were excluded from this study, of which 156 were excluded because no embryos were available for delivery. Thus, 819 cycles were included in the analysis, at least one euploid embryo was available in 542 cases and 401 FET cycles.

Baseline characteristics of participants

The baseline characteristics of all participants, divided into three groups based on AMH levels, are presented in Table 1. A total of 819 cycles were divided into three groups: Group 1 (AMH < 1.08) included 205 cycles; Group 2 (1.08 ≤ AMH < 3.34) included 410 cycles; Group 3 (AMH ≥ 3.34) included 204 cycles. The baseline characteristics, including number of eggs retrieved, number of normal fertilizations, day 3 high-quality embryos, number of blastocysts formed, euploid embryos, mosaic embryos, and euploid embryo rate, showed a gradual increase across the groups (P < 0.05). Conversely, female age, male age, fertilization rate, blastocyst formation rate, and aneuploidy embryo rate gradually decreased (P < 0.05). There were no significant differences in infertility type, years of infertility, or BMI in each group (P > 0.05).

Table 1.

Baseline patient characteristics, embryo culture, and testing results

	Total (n = 819)	AMH < 1.08(n = 205)	1.08 ≤ AMH < 3.34 (n = 410)	AMH ≥ 3.34(n = 204)	P	Statistic
Female age (years)	36.7 ± 4.2	38.6 ± 4.1	36.6 ± 4.0	34.7 ± 3.9	< 0.001	48.427
Male age (years)	37.8 ± 5.7	40.0 ± 5.5	37.8 ± 5.7	35.8 ± 5.2	< 0.001	30.080
Type of infertility					0.709	0.687
Primary infertility	132 (16.1)	30 (14.6)	66 (16.1)	36 (17.6)
Secondary infertility	687 (83.9)	175 (85.4)	344 (83.9)	168 (82.4)
Years of infertility (years)	2.0 (1.0, 4.0)	2.0 (1.0, 4.0)	2.0 (1.0, 4.0)	2.0 (1.0, 3.0)	0.647	0.872
BMI (Kg/m2)	23.6 ± 3.5	23.2 ± 3.6	23.7 ± 3.3	23.8 ± 3.8	0.172	1.766
PGT-A factors					< 0.001	30.779
AMA	264 (32.2)	94 (45.9)	128 (31.2)	42 (20.6)
RSA	421 (51.4)	83 (40.5)	212 (51.7)	126 (61.8)
RIF	134 (16.4)	28 (13.7)	70 (17.1)	36 (17.6)
bFSH (mIu/mL)	6.2 (5.0, 7.6)	7.3 (5.5, 9.5)	6.2 (5.1, 7.3)	5.6 (4.6, 6.8)	< 0.001	58.034
bLH (mIu/mL)	4.6 (3.2, 6.2)	4.2 (2.9, 5.4)	4.3 (3.1, 6.1)	5.7 (4.0, 7.9)	< 0.001	46.884
bE2 (Pg/mL)	40.6 (29.7, 55.3)	41.2 (30.5, 60.1)	41.0 (29.3, 56.4)	39.0 (29.0, 49.2)	0.186	3.360
bAFC	11.0 (7.0, 16.0)	6.0 (4.0, 8.0)	10.0 (8.0, 14.0)	19.0 (13.8, 25.0)	< 0.001	371.377
Total dose of Gn (U)	3176.1 ± 991.4	2985.2 ± 1224.5	3248.3 ± 946.2	3222.7 ± 775.0	0.006	5.164
Oocytes retrieved (n)	11.3 ± 7.9	4.6 ± 3.1	10.7 ± 5.5	19.4 ± 8.4	< 0.001	325.436
Number of normal fertilization (n)	7.0 (4.0, 11.0)	3.0 (2.0, 5.0)	7.0 (4.0, 10.0)	13.0 (8.8, 17.0)	< 0.001	335.513
Fertilization rate (%)	71.9 ± 19.8	77.8 ± 22.6	70.6 ± 19.1	68.5 ± 17.0	< 0.001	13.377
Day 3 high-quality embryos (n)	3.0 (2.0, 5.0)	2.0 (1.0, 3.0)	3.0 (2.0, 5.0)	6.0 (4.0, 10.0)	< 0.001	207.568
Day 3 high-quality embryo rate (%)	57.1 (37.5, 73.5)	66.7 (50.0, 100.0)	50.0 (33.3, 70.0)	57.1 (39.0, 69.2)	< 0.001	22.205
Number of blastocyst cultures (n)	7.0 (3.0, 11.0)	3.0 (2.0, 5.0)	7.0 (4.0, 10.0)	13.0 (9.0, 17.0)	< 0.001	340.812
Number of blastocysts formed (n)	3.0 (2.0, 6.0)	2.0 (1.0, 2.0)	3.0 (2.0, 5.0)	7.0 (4.0, 10.0)	< 0.001	255.878
Blastocyst formation rate (%)	57.2 ± 25.2	66.2 ± 27.6	54.2 ± 24.6	54.1 ± 21.7	< 0.001	18.396
Biopsied embryos (n)	3.0 (1.0, 4.0)	1.0 (1.0, 2.0)	3.0 (1.0, 4.0)	4.0 (3.0, 6.0)	< 0.001	187.897
Euploid embryos (n)	1.0 (0.0, 2.0)	0.0 (0.0, 1.0)	1.0 (0.0, 2.0)	2.0 (1.0, 3.0)	< 0.001	105.68
Aneuploid embryos (n)	1.0 (0.0, 2.0)	1.0 (1.0, 1.0)	1.0 (0.0, 2.0)	1.0 (0.0, 2.0)	0.142	3.907
Mosaic embryos (n)	1.0 (0.0, 1.0)	0.0 (0.0, 1.0)	1.0 (0.0, 1.0)	1.0 (0.0, 2.0)	< 0.001	73.516
Euploid embryo rate (%)	33.3 (0.0, 60.0)	0.0 (0.0, 50.0)	33.3 (0.0, 60.0)	50.0 (25.0, 60.6)	< 0.001	15.295
Aneuploid embryo rate (%)	50.0 (0.0, 75.0)	66.7 (33.3, 100.0)	50.0 (0.0, 66.7)	25.0 (0.0, 50.0)	< 0.001	52.728
Mosaic embryo rate (%)	20.0 (0.0, 50.0)	0.0 (0.0, 33.3)	25.0 (0.0, 50.0)	25.0 (0.0, 40.7)	< 0.001	22.345

BMI body mass index, PGT-A preimplantation genetic testing for aneuploidy, AMA advanced maternal age, RSA recurrent spontaneous abortion, RIF recurrent implantation failure, bFSH basic follicle stimulating hormone, bLH basic luteinizing hormone, bE2 basic estradiol, AMH anti-Müllerian tubular hormone, Gn gonadotropin

Logistic regression analysis

Covariate screening identified variables with an effect size change greater than 10%. Variables included female age, male age, PGT-A factors, bFSH, bLH, bE2, bAFC, oocytes retrieved, number of normal fertilizations, day 3 high-quality embryos, number of blastocysts cultured, and number of blastocysts formed. Covariate screening suggested covariance between the number of oocytes retrieved, the number of normal fertilizations, day 3 high-quality embryos, the number of blastocysts cultured, and the number of blastocysts formed, which were excluded from the multivariate analysis. Univariate logistic regression analysis revealed potential co-factors associated with aneuploidy rates (P < 0.10), including female age, male age, years of infertility, PGT-A factors, bLH, and AMH (Table 2). Following this, multivariate logistic regression analysis was performed (Table 2). The unadjusted OR for AMH to aneuploidy rate was 0.78 (95% CI 0.73–0.84, P < 0.05). After adjusting for female age, male age, years of infertility, PGT-A factors, bFSH, bLH, and bE2, the adjusted OR was 0.87 (95% CI 0.84–0.90, P < 0.05). Multifactorial logistic regression analysis revealed OR was 1.02 (95% CI 0.95–1.09, P = 0.602) for AMH versus live birth, after adjusting for female age, BMI, type of infertility, and endometrial thickness (EMT) on the day of embryo transfer (Supplementary Table 1). The ROC curve demonstrated a predictive model AUC of 0.596 (0.539–0.652) for combined AMH, female age, BMI, infertility type, and EMT on the day of embryo transfer (Supplementary Fig. 1).

Table 2.

Univariate and multivariate analysis of embryonic aneuploidy rates according to AMH values

	Univariate analysis		Multivariate analysis
	OR (95% CI)	P	OR (95% CI)	P
Female age (years)	1.27 (1.22 ~ 1.32)	< 0.001	1.22 (1.19 ~ 1.25)	< 0.001
Male age (years)	1.14 (1.10 ~ 1.17)	< 0.001	1.00 (0.98 ~ 1.02)	0.948
Type of infertility	1.04 (0.72 ~ 1.51)	0.833
Years of infertility (years)	1.05 (1.00 ~ 1.10)	0.049	0.97 (0.95 ~ 0.99)	0.009
BMI (kg/m2)	0.99 (0.95 ~ 1.03)	0.700
PGT-A factor (RSA)	0.57 (0.35 ~ 0.91)	0.019	0.58 (0.47 ~ 0.72)	< 0.001
PGT-A factor (RIF)	0.20 (0.14 ~ 0.29)	< 0.001	1.11 (0.87 ~ 1.41)	0.412
bFSH (mIu/mL)	0.37 (0.23 ~ 0.59)	< 0.001	0.98 (0.96 ~ 1.00)	0.103
bLH (mIu/mL)	0.96 (0.92 ~ 1.00)	0.078	1.00 (0.98 ~ 1.03)	0.773
bE2 (Pg/mL)	1.00 (1.00 ~ 1.01)	0.142	1.00 (1.00 ~ 1.01)	0.075
AMH (ng/mL)	0.78 (0.73 ~ 0.84)	< 0.001	0.87 (0.84 ~ 0.90)	< 0.001
Total Gn (U)	1.00 (1.00 ~ 1.00)	0.498
Fertilization rate	1.00 (1.00 ~ 1.01)	0.390
Day 3 high-quality embryo rate (%)	1.00 (1.00 ~ 1.01)	0.961
Blastocyst formation rate	1.00 (0.99 ~ 1.00)	0.437

Embryonic aneuploidy rate was dichotomized (≥ 50% defined as positive)

AMH anti-Müllerian hormone, BMI body mass index, PGT-A preimplantation genetic testing for aneuploidy, RSA recurrent spontaneous abortion, RIF recurrent implantation failure, bFSH basic follicle stimulating hormone, bLH basic luteinizing hormone, bE2 basic estradiol, Gn gonadotropin, OR odds ratio, CI confidence interval

Threshold effect analysis

As shown in Fig. 2, RCS suggested a nonlinear association between AMH level and aneuploidy rate. After adjusting for female age, PGT-A factors, and years of infertility, the risk of aneuploidy rate positivity increased by 40% for each unit decrease in AMH when AMH ≤ 2.54, with an adjusted OR of 0.60 (95% CI 0.45–0.81, nonlinear P < 0.05). There was no significant change in aneuploidy rate when AMH > 2.54 (Table 3).

Fig. 2.

Dose–response relationship between the AMH levels and aneuploidy rate. Adjustments: female age, PGT-A factors, years of infertility, only 95% of the data are shown; solid and dashed lines indicate the estimated values and their corresponding 95% confidence intervals. AMH: anti-Müllerian hormone; embryonic aneuploidy rate is dichotomized (≥ 50% defined as positive)

Table 3.

Threshold effect analysis of AMH values and embryonic aneuploidy rate

	Adjusted OR (95% CI)	P
AMH ≤ 2.54	0.60 (0.45 ~ 0.81)	< 0.001
AMH > 2.54	1.00 (0.81 ~ 1.23)	0.978
Likelihood ratio test		0.005

Embryonic aneuploidy rate was dichotomized (≥ 50% defined as positive)

Adjustments: female age, PGT-A factors, years of infertility, only 95% of the data were analyzed

OR odds ratio, CI confidence interval, AMH anti-Müllerian hormone

Subgroup analysis

Stratified analyses were conducted to evaluate potential factors influencing the relationship between AMH and embryonic aneuploidy rates. Subgroup analyses were performed based on female age, BMI, type of infertility, years of infertility, PGT-A factors, and adjusted for age, PGT-A factors, and years of infertility. No significant interactions were observed in any of the subgroups (Fig. 3).

Fig. 3.

Subgroup analysis of the correlation between AMH and embryonic aneuploidy rate. Adjustments: female age, PGT-A factors, years of infertility. OR, odds ratio; CI, confidence interval; AMH, anti-Müllerian hormone; BMI, body mass index; PGT-A, preimplantation genetic testing for aneuploidy; AMA, advanced maternal age; RSA, recurrent spontaneous abortion; RIF, recurrent implantation failure, embryonic aneuploidy rate was dichotomized (≥ 50% defined as positive); subgrouping factors: female age, BMI, infertility type, and PGT-A factors

Discussion

This study revealed a significant nonlinear relationship between AMH levels and the embryonic aneuploidy rate and highlighted the threshold effect of AMH. Among infertile women undergoing PGT-A, the embryonic aneuploidy rate was notably higher when AMH was ≤ 2.54 ng/mL. The risk of aneuploidy increased by 40% for every 1-unit decrease in AMH. Studies have shown that reduced AMH levels may indicate granulosa cell hypoplasia, which affects oocyte maturation and chromosome stability. Normal follicular fluid AMH is greater than 15 pmol/L (approximately 2.1 ng/mL) [31]. This is close to the cutoff value of the present study. The aneuploidy rate leveled off when AMH was > 2.54 ng/mL. In contrast, the incidence of chromosomal aneuploidy slowly increases with AMH values above 4 ng/mL. Previous studies have shown that patients with AMH levels greater than 4 ng/mL have a significantly higher risk of ovarian hyperstimulation syndrome (OHSS) during in vitro fertilization (IVF) (40% vs. 4% in the low AMH group) [32]. Elevated levels of reactive oxygen species (ROS) in the follicular fluid of patients with OHSS attack the microtubular proteins of the oocyte. This interferes with the formation of the spindle and increases the risk of chromosomal aneuploidy [33, 34]. After adjusting for female age, PGT-A factors, and years of infertility, AMH still remained an independent predictor of aneuploidy risk. These findings are partially consistent with previous studies but also offer new perspectives. For example, Li et al. (2023) noted that AMH is a better predictor of aneuploidy than age in women under 35 and that both AMH and age impact embryo quality in older women [35]. However, the present study found that a threshold effect of AMH may be independent of age, suggesting a direct regulatory role of AMH on oocyte quality. In addition, Jiang et al. (2020) found a higher aneuploidy rate in women with low AMH based on tertile groupings (< 1.5, 1.5–5.6, ≥ 5.6 ng/mL) (66.7% vs. 42.9%) [36]. The present study further refined the threshold to 2.54 ng/mL and revealed a nonlinear risk gradient, providing a more precise clinical assessment. Our findings align with those reported in previous publications, including those of Carnesi et al. (2025) and Arnanz et al. (2023) [37, 38]. Some studies concluded that AMH is not directly associated with aneuploidy rate [39, 40], and the variability in conclusions may be due to differences in categorization methods that do not account for nonlinear effects or variations in the classification methodology for aneuploidy rates.

After adjusting for the confounders of female age, BMI, infertility type, and endometrial thickness on the day of transplantation, infertility type and endometrial thickness (cm) on the day of transplantation were found to be significantly associated with live birth (P < 0.05), whereas AMH was not found to be independently predictive of live birth outcome. The low efficacy of the prediction model combining AMH, female age, BMI, infertility type, and endometrial thickness on the day of transplantation in predicting live birth was confirmed by the ROC curve. AMH predicts aneuploidy rates but does not extend to live births. AMH has been shown to independently predict the risk of embryonic aneuploidy (e.g., aneuploidy is elevated in those with a low AMH). However, the implantation and development of aneuploid embryos is more dependent on endometrial receptivity, maternal metabolism, and other factors. A multicenter study by Li et al. (2023) of PGT-A cycles similarly noted that AMH was associated with aneuploidy rates but was not predictive of live birth outcomes in a single transfer [35].

At the molecular level, AMH may affect chromosome stability through two primary pathways. First, low AMH levels suggest granulosa cell hypoplasia, potentially leading to insufficient secretion of pro-survival factors (e.g., GDF9, BMP15) during oocyte maturation, which could trigger meiotic errors [41–45]. Second, reduced ovarian reserve is accompanied by reduced mitochondrial DNA copy number and accumulation of oxidative stress, which interferes with spindle assembly and chromosome segregation [46, 47]. However, polycystic ovary syndrome (PCOS) patients frequently exhibit elevated AMH levels, which, by impeding the FSH signaling pathway, result in diminished sensitivity of granulosa cells to FSH, suppression of aromatase activity, and impediment of the conversion of androgens to estrogens [48, 49]. This process gives rise to a hyperandrogenic and hypoestradiolic follicular microenvironment, which has the potential to impede oocyte maturation [50–55]. Furthermore, elevated levels of ROS in the follicular fluid of PCOS may interfere with the meiotic maturation of oocytes and the fertilization process, thereby affecting the chromosomal stability of the embryo [56]. The findings of the RCS curves in this investigation align with the aforementioned conclusions. However, this study is limited by the biological depth of the clinical data. Future research combining oocyte single-cell sequencing and metabolomics will be essential to further investigate these mechanisms.

This study introduces several methodological innovations. A statistical model based on dose–response relationship was used to reveal the nonlinear association between AMH and aneuploidy rate, addressing the limitations of the traditional linear assumptions. The determination of biomarker threshold was made more objective through the use of automated inflection point detection, eliminating the need for subjective threshold settings. Additionally, the construction of an AMH threshold model provides a new biomarker reference system for PGT-A indication screening. Compared with the current PGT-A indication based on age ≥ 38 years (e.g., ASRM 2024), the present study confirms that AMH still maintains an independent predictive value even after adjusting for age. Incorporating AMH into PGT-A screening could optimize clinical decision-making for patients under 38 years of age, potentially reducing both the psychological and economic burdens of ineffective embryo transfers. However, this study has some limitations. Some of the AMH data were obtained from an external hospital testing system, although this bias was minimized by standardizing measurement units. Metabolic indicators such as blood glucose and thyroid stimulating hormone (TSH) were not included in the study, although available data suggest that these parameters lack clinical decision-making value in ART. Confounding was reduced by excluding diabetic patients and strictly regulating TSH levels prior to ovulation induction. The study’s retrospective design resulted in < 5% missing data for bFSH, bE2, and bLH, which were addressed through multiple interpolation. Fourthly, the study was conducted among infertility patients undergoing PGT-A treatment, a population that exhibited a notable tendency toward advanced age and a specific pathological background (e.g., repeated implantation failures, recurrent miscarriages). It is noteworthy that the study strictly excluded non-chromosomal factors contributing to pregnancy failure, such as patients with a history of immune and endocrine disorders. Furthermore, relevant covariates, including patient age and indications for PGT-A, were adjusted for in the analysis. Despite the study’s rigorous exclusion of non-chromosomal factors, such as pregnancy failure in patients with a history of immune and endocrine system disorders, and its adjustment for covariates, including patient age and PGT-A indication, the predictive value of AMH on embryo quality remains to be validated in future studies within natural pregnancy cohorts. Furthermore, the study examined the correlation between aneuploidy rate and AMH levels in blastocyst-stage embryos. This design, while adhering to the clinical practice criteria for PGT-A (in which only blastocysts can be biopsied), does restrict the scope of the study findings. Future studies should incorporate follicular developmental dynamics, such as chromosome analysis of GV-stage oocytes, and explore the association of AMH with aneuploidy in embryos at different developmental stages. Notwithstanding, the conclusions of the present study maintain clinical relevance. Chromosomal status at the blastocyst stage is a pivotal factor in determining the live birth rate. The clinical value of PGT-A is contingent upon this stage of screening, and the present study directly addresses this scenario.

Conclusions

This study revealed the nonlinear relationship between AMH levels and embryonic aneuploidy rates, highlighting its threshold effect. These findings provide a crucial theoretical foundation and practical guidance for optimizing assisted reproduction strategies in clinical settings. Future research should focus on further exploring the underlying biological mechanisms and validating the generalizability of these results in larger and more diverse patient populations.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

PGT-A

Preimplantation genetic testing for aneuploidy

AMH

Anti-Müllerian hormone 

AMA

Advanced maternal age

RCS

Restricted cubic spline

RSA

Recurrent spontaneous abortion

RIF

Recurrent implantation failure

GnRH

Gonadotropin-releasing hormone

r-FSH

Recombinant follicle stimulating hormone

HMG

Human menopausal gonadotropin

bAFC

Basal antral follicle count

BMI

Body mass index

r-hCG

Recombinant human choriogonadotropin alfa solution for injection

bFSH

Basic follicle-stimulating hormone

bLH

Basic luteinizing hormone

bE2

Basic estradiol

Gn

Gonadotropin

MII

Metaphase II

2PN

Two pronuclei

SD

Standard deviation

IQR

Interquartile range

RCS

Subject work characteristics

OHSS

Ovarian hyperstimulation syndrome

ROS

Reactive oxygen species

PCOS

Polycystic ovary syndrome

OR

Odds ratio

CI

Confidence interval

TSH

Thyroid stimulating hormone

ART

Assisted reproductive technology

Author contribution

ST designed the study and was responsible for collecting and organizing the data, analyzing the data, and writing the manuscript. PZ participated in the data collection and organization. KS and QZ assisted in analyzing the data. YY participated in the design, directed the research, and critically reviewed and revised the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This research and researchers were funded as follows: Liaoning Provincial Science and Technology Program—Key R&D Projects (2024JH2/102600271 to Yuexin Yu).

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethical approval

The study was approved by the Medical Ethics Committee of the General Hospital of the Northern Theater of Operations, and due to the retrospective nature of the study, no personal information of the patients was used, and the Institutional Review Board of the hospital approved the study with a waiver of the requirement to obtain informed consent. The Ethical Approval Number: Y (2025) 122. All reports follow the guidelines of Enhancing Reporting of Observational Studies in Epidemiology.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.