Low anti-Mullerian hormone in reproductive age is associated with MASLD in midlife

Abstract

Background:

Anti-Mullerian hormone (AMH) is produced by ovarian follicles and is clinically reflective of reproductive aging. AMH has been associated with the severity of steatohepatitis in those with known metabolic dysfunction–associated steatotic liver disease (MASLD), although its association with prevalent MASLD is not known.

Methods:

Using the multicenter longitudinal Coronary Artery Risk Development in Young Adults cohort, we evaluated the association of low AMH levels in healthy young women with subsequent MASLD in midlife, assessed by CT scan 10 years later. Alternate causes of steatosis were excluded, and multivariable logistic regression was adjusted for confounding metabolic risk factors.

Results:

Among 585 eligible participants, 50 (8.5%) had MASLD. MASLD was more common among participants with low AMH (14% vs. 7%, p=0.03). Adjusted for baseline covariates, low AMH was associated with 2.50-fold higher odds of prevalent MASLD (95% CI: 1.18–5.28, p=0.02). The findings persisted after adjusting for both baseline and change in metabolic profiles (OR: 2.36, 95% CI: 1.05–5.27, p=0.04). The relationship between low AMH and MASLD was not mediated by body mass index (p=0.82) or visceral adipose tissue volume (p=0.14).

Conclusions:

Low AMH levels in reproductive-aged women conferred a more than 2-fold higher odds of prevalent MASLD in midlife, independent of metabolic comorbidities. AMH levels may therefore serve as a valuable marker of MASLD risk in young women.

INTRODUCTION

Metabolic dysfunction–associated steatotic liver disease (MASLD) is a leading cause of liver-related morbidity and mortality worldwide, and cirrhosis from metabolic dysfunction–associated steatohepatitis (MASH) is now the most common indication for liver transplantation among women.¹

Sex differences in the epidemiology of MASLD are well recognized, where steatosis is more common in men before age 50, but more common in women thereafter. Postmenopausal women, in particular, have a disproportionately higher risk of developing MASH and progressing to advanced fibrosis compared with age-matched men with MASLD.²,³ The heightened risk of MASH following menopause may relate to a decline in estradiol levels, as estradiol has been shown to lower hepatic stellate cell activity and subsequent fibrosis.⁴

The menopausal transition is also characterized by worsening metabolic profiles, including redistribution of adipose tissue to visceral stores, worsening insulin resistance, and an increased risk of dyslipidemia and type 2 diabetes mellitus (T2DM).⁵ Such changes are particularly relevant to incident MASLD and to progressive fibrosis in women with underlying liver disease. With the rising burden of MASLD, it is imperative to identify risk factors that contribute to disease development and biomarkers that may assist with early diagnosis in women.

Anti-Müllerian hormone (AMH) is a glycoprotein produced by ovarian follicles; it is considered the most clinically accurate marker of reproductive aging and ovarian reserve.⁶ AMH levels can be used to reliably identify the menopausal transition, which may be particularly important for younger patients who experience secondary amenorrhea from other causes, such as chronic liver disease. Lower AMH levels have been associated with a range of metabolic comorbidities, including cardiovascular disease, T2DM, and dyslipidemia.⁷,⁸ Lower AMH levels have also been associated with the presence and severity of MASH in reproductive-aged women with MASLD.⁹,¹⁰ The relationship between AMH level and subsequent MASLD has not been evaluated, which is clinically relevant for the larger population of young women without baseline liver disease. We hypothesized that low AMH levels measured during reproductive years would confer an increased risk of MASLD in midlife.

METHODS

Data source and study cohort

This study utilized data from the prospective CARDIA cohort and the ancillary CARDIA Women’s study (CWS).¹¹,¹² As described elsewhere¹¹, CARDIA is a multicenter community-based longitudinal study of healthy young adults recruited from 4 cities across the United States (Birmingham, AL; Chicago, IL; Minneapolis, MN; and Oakland, CA). All research was conducted in accordance with the Declarations of Helsinki and Istanbul. Institutional review board approval was obtained from all participating centers.

Participants were aged 18–30 years at enrollment (Study Year 0, 1985–1986). Follow-up examinations have been conducted at regular intervals, occurring at 2, 5, 7, 10, 15, 20, 25, 30, and 35 years after the Year 0 exam. Cardiometabolic laboratory testing was performed at each study visit. Serum AMH was collected as part of the CWS at Study Year 15 (2000–2001), and CT quantification of hepatic steatosis was performed at Study Year 25 (2010–2011). These time points mark the baseline and the end of the study in this analysis, respectively. Female participants who completed the CT quantification of hepatic steatosis at Year 25 were considered for inclusion (n=1805 of 1930) (Figure 1).¹³ Participants with alternate causes of hepatic steatosis (n=124) were excluded, including those with consumption of >2 alcoholic drinks per day (n=79), HIV (n=4), medication use associated with steatosis, including amiodarone, methotrexate, valproic acid, tamoxifen, steroids, diltiazem (n=23), and self-report of other causes of chronic liver disease (n=18). We then excluded participants without the required hormonal testing (n=1023). Participants who reported removal of an ovary (the site of AMH production) and/or uterus during their Year 15 visit (n=73) were also excluded. Hysterectomy was grouped with oophorectomy as bilateral salpingo-oophorectomy was routinely performed with hysterectomy before 2010.¹⁴ Thus, most participants who reported hysterectomy in our study would likely have had an oophorectomy. Written consent was given in writing by all participants.

FIGURE 1.

Participant flow diagram. Abbreviations: AMH, anti-Mullerian hormone; CARDIA, Coronary Artery Risk Development in Young Adults.

Measurements

Anti-Mullerian hormone

AMH levels were measured as a part of the ancillary CWS (Study Year 15).¹⁵ AMH was measured from stored serum samples using the Ultra-Sensitive AMH ELISA assay from Ansh Laboratories. The lower limit of quantification was 0.09 ng/mL, and values <0.09 ng/mL were recoded as 0.089 ng/mL. Low AMH was defined as an AMH level within the lowest quintile of the study cohort (AMH <0.15 ng/mL). AMH was characterized by quintile due to non-normal distribution.¹⁶ This threshold aligns with previously established definitions of “ultra-low” AMH and that of the menopausal transition.¹⁷,¹⁸

Hepatic steatosis

Hepatic steatosis was identified by noncontrast multidetector CT scan performed using GE models 750HD (64) at the Birmingham site, GE LightSpeed VCT (64) in Oakland (GE Healthcare), and Siemens Sensation 64 at the Chicago and Minneapolis sites (Siemens Medical Solutions).¹⁹ Image analysis and quality control were performed at a central reading center (Wake Forest University Health Sciences). Hepatic steatosis was defined as a mean liver attenuation of ≤40 Hounsfield Units, determined from 9 measurements on 3 CT slices of the right hepatic lobe.²⁰ The threshold of 40 Hounsfield Units has been shown to correlate with biopsy-confirmed moderate to severe steatosis.²¹,²²,²³

Covariates

Participant demographics, medical history, and substance use were obtained through standardized surveys.¹¹ Participants were categorized as “younger” or “older” using the median age of the study cohort at the time of AMH collection (≤42 y, >42 y). Race was defined as “Black” or “White” by self-report. Medication use was determined by prescription review at each study visit. Menarche was defined by self-reported age at first menstrual cycle; early menarche was defined as menarche before 11 years of age.²⁴ Potential amenorrhea was defined as a lack of menses within 3 months of AMH measurement without reported menopause (>12 mo without a menstrual cycle). Polycystic ovary syndrome (PCOS) was defined by self-report; the PCOS group was stratified by the presence of hyperandrogenism at Year 2.²⁵,²⁶ Height, weight, and waist circumference were measured by certified technicians at each visit,²⁷ and body mass index (BMI) was calculated as weight (kg)/height (m)². Visceral adipose tissue volume, measured on Year 25 CT scan, was defined as the sum of fat voxels within 10 mm set of slices centered at the L4–5 disk within the intraabdominal cavity.²⁶,²⁸ Standard protocols were employed for fasting serum collection and assays of plasma triglycerides, HDL, LDL, total cholesterol, serum glucose, and insulin levels.²⁹,³⁰ Homeostatic model assessment of insulin resistance (HOMA-IR) was calculated using the equation: (fasting glucose [mg/dL] × fasting insulin [mIU/L])/405. Lipid measurements were missing in 1%–2% of the cohort at Year 15. Because measurements from Year 10 and Year 15 were highly correlated (correlation values of 0.59–0.76, p values <0.01), missing Year 15 values were replaced with Year 10 values during model development.

Analysis

Baseline and end-of-study characteristics were compared using the chi-square and Mann-Whitney U tests, as appropriate. Median AMH levels were compared between participants with and without MASLD using the Mann-Whitney U test, as non-normal distributions were observed on the Shapiro-Wilk tests. AMH levels were also compared between participants with and without T2DM, hypertension, dyslipidemia, and PCOS, and between those with and without potential amenorrhea. Correlations between AMH levels and age, lipids, and BMI were evaluated using the Spearman correlation coefficients.

The association between AMH level and prevalent MASLD was evaluated using logistic regression. A simplified bivariate model assessed the association of low AMH with MASLD after adjusting for baseline age and BMI. Sensitivity analyses were also performed after excluding those with baseline T2DM, hypertension, dyslipidemia, and PCOS, as these conditions independently increase MASLD risk.³¹,³²,³³

A multivariable model was developed using covariates selected a priori for clinical relevance, including age, race, BMI, triglyceride levels, PCOS, and HOMA-IR.³⁴,³⁵ To capture longitudinal change in metabolic conditions, the association of low AMH with prevalent MASLD was also evaluated after adjusting for change in covariates from Years 15 to 25, including BMI, HOMA-IR, triglyceride levels, in addition to new PCOS diagnosis and interval surgical menopause. The final model was developed using forward selection, including those with 2-sided p values <0.05.

Lastly, we evaluated the role of BMI, triglyceride levels, visceral adipose tissue volume, and HOMA-IR as potential mediators of the association between low AMH and MASLD, given prior data supporting the link between low AMH levels and these cardiometabolic indices.³⁶ Mediation analysis was performed using the SPSS PROCESS 4.2 macro.³⁷ Significance was evaluated at p<0.05. All analyses were performed in SPSS Version 29.

RESULTS

Cohort characteristics

A total of 585 participants were included in this analysis (Table 1). The median age at baseline was 42 years (IQR: 39–45). Cardiometabolic comorbidities included hypertension (15.9%, n=93), dyslipidemia (15.9%, n=93), T2DM (7.5%, n=44), and PCOS (6.5%, n=38). The median age at menarche was 13 years, and 6.0% of participants (n=35) reported early menarche, defined as the onset at age <11 years. Most participants reported prior hormonal contraception use (84.3%, n=493), but few had used hormone replacement therapy (1.7%, n=10). Just under 40% (n=227) experienced menopause by the end of the study, including 152 (26.0%) with natural menopause and 75 (12.8%) with surgical menopause.

TABLE 1.

Cohort characteristics at baseline and study end (n=585)

	Baseline (Year 15)	Study End (Year 25)	p
Age, median (IQR), y	42 (39–45)	50 (47–53)	<0.01
Age at menarche, median (IQR), y	13 (11–12)	—	—
Race, n (%)
White	297 (50.8)	—	—
Black	288 (49.2)	—	—
PCOS, n (%)	39 (6.7)	—	—
T2DM, n (%)	44 (7.5)	72 (12.4)	<0.01
Dyslipidemia, n (%)	93 (15.9)	156 (26.8)	<0.01
Hypertension, n (%)	93 (15.9)	200 (34.4)	<0.01
HRT exposure, n (%)	10 (1.7)	37 (6.4)	<0.01
Waist circumference, median (IQR), cm	82 (74–96)	89 (79–101)	<0.01
BMI, median (IQR), kg/m2	27.6 (24–33)	29.3 (25–36)	<0.01
Total cholesterol, median (IQR), mg/dL	178 (157–200)	195 (171–217)	<0.01
LDL, median (IQR), mg/dL	106 (86–126)	110 (90–133)	0.02
HDL, median (IQR), mg/dL	53 (45–63)	61 (51–72)	<0.01
Triglycerides, median (IQR), mg/dL	74 (54–102)	84 (64–120)	<0.01
Glucose, median (IQR), mg/dL	81 (76–87)	91 (85–98)	<0.01
Insulin, median (IQR), mcU/mL	11 (8–17)	13 (9–20)	<0.01
HOMA-IR, median (IQR)	2.3 (1.6–3.6)	2.9 (2.1–4.2)	<0.01

Baseline refers to Year 15 when AMH levels are measured. Study end refers to Year 25 when steatosis was evaluated. Missing Year 15 triglycerides (n=7) and HOMA-IR (n=1).

Abbreviations: BMI, body mass index; HOMA-IR, homeostatic model assessment of insulin resistance; HRT, hormone replacement therapy; PCOS, polycystic ovary syndrome; T2DM, type 2 diabetes mellitus.

AMH

The median AMH level in this cohort was 0.77 ng/mL (IQR: 0.23–2.13). As expected, younger participants were more likely to have higher AMH quintiles, while those who had transitioned through menopause by Year 25 had lower AMH quintiles (Supplemental Table S1, http://links.lww.com/HC9/C228). Regarding cardiometabolic comorbidities, baseline hypertension was more common in the highest AMH quintile (25.6% vs. 8.0%–18.2%, p<0.01), while dyslipidemia was more common in the lowest AMH quintile (20.9% vs. 11.6%–16.5%, p=0.11) (Supplemental Table S1, http://links.lww.com/HC9/C228). However, median AMH levels did not differ by the presence of baseline cardiometabolic comorbidities (p=0.35) (Supplemental Table S2, http://links.lww.com/HC9/C228).

The median AMH level was higher in PCOS participants (1.61 vs. 0.69 ng/mL, p<0.01), which is expected in the setting of higher ovarian follicle number and size. The highest AMH levels were evident in those with hyperandrogenic PCOS (2.81 ng/mL, IQR: 0.94–6.28) (Supplemental Table S3, http://links.lww.com/HC9/C228). When stratified by AMH quintile, the prevalence of PCOS was significantly higher among the upper quintiles (11%–12%) compared to the lower quintiles (1%–3%).

MASLD

MASLD was present in 8.5% (n=50) of the cohort at Year 25. Baseline hypertension (16.0% vs. 15.9%, p=0.98) and dyslipidemia (18.0% vs. 15.7%, p=0.67) did not differ by MASLD status, although there was a trend toward increased T2DM among those with prevalent MASLD (14.0% vs. 6.9%, p=0.07) (Supplemental Table S4, http://links.lww.com/HC9/C228). In contrast, metabolic comorbidities at Year 25 were more common in those with prevalent MASLD, including hypertension (52.0% vs. 32.5%, p<0.01), dyslipidemia (38.0% vs. 25.8%, p=0.05), and T2DM (32.0% vs. 10.5%, p<0.01), with a trend toward increased PCOS prevalence (12.0% vs. 6.2%, p=0.11). Early menarche occurred in 6% of participants with and without MASLD. More participants with MASLD reported menopause in the interval between AMH measurement and Year 25 CT (52.0% vs. 37.6%, p=0.05), driven by differences in surgical menopause (26.0% vs. 11.6%, p<0.01).

Association between AMH and MASLD

The median AMH level was similar in the presence and absence of MASLD (0.61 vs. 0.78 ng/mL, respectively, p=0.24). However, MASLD was most common in the lowest quintile of AMH compared to the rest of the quintiles (13.6% vs. 7.4%, p=0.03) (Figure 2). We noted the highest MASLD prevalence among younger participants with low AMH (n=5 of 20, 25.0%) (Supplemental Figure S1, http://links.lww.com/HC9/C228). Interestingly, among those with low AMH, prevalent MASLD was higher in participants who did not transition to menopause during the study than in those who did, though this was not statistically significant (14.8% vs. 9.7%, p=0.30).

FIGURE 2.

Prevalence of MASLD by AMH quintile. Bars represent the proportion of women with prevalent MASLD across quintiles of AMH, with Q1 representing the lowest and Q5 the highest AMH levels. Percentages shown above bars indicate the prevalence of MASLD within each quintile; sample sizes for each quintile are shown on the x-axis. Abbreviations: AMH, anti-Mullerian hormone; MASLD, metabolic dysfunction–associated steatotic liver disease.

In unadjusted analysis, low AMH was associated with 2-fold higher odds of prevalent MASLD (OR: 1.99, 95% CI: 1.04–3.78, p=0.04), and there was increased strength of association after adjustment for baseline age (aOR: 2.67, 95% CI: 1.30–5.47, p=0.01). In fully adjusted analysis, low AMH was associated with 2.5-fold higher odds of prevalent MASLD in midlife (aOR: 2.50, 95% CI: 1.18–5.28, p=0.02), independent of baseline age, race, PCOS, BMI, triglycerides, and HOMA-IR (Table 2). This relationship persisted in sensitivity analyses excluding patients with baseline T2DM (p=0.04), dyslipidemia (p=0.01), or PCOS (p=0.01) (Supplemental Table S5, http://links.lww.com/HC9/C228). Importantly, low AMH remained significantly associated with MASLD with adjustment for both baseline and change in cardiometabolic parameters (aOR: 2.36, 95% CI: 1.05–5.27, p=0.04) (Table 3).

TABLE 2.

Association of low AMH with prevalent MASLD, adjusted for baseline covariates (n=585)

	Unadjusted		Adjusted
	OR (95% CI)	p	aOR (95% CI)	p
Low AMH (reference: non-low AMH)	1.99 (1.04–3.78)	0.04	2.50 (1.18–5.28)	0.02
Black race (reference: White)	1.06 (0.59–1.89)	0.86	1.63 (0.83–3.18)	0.15
Age > 42 y (reference: ≤42)	0.72 (0.40–1.29)	0.27	0.58 (0.29–1.16)	0.13
PCOS (reference: no PCOS)	2.07 (0.82–5.22)	0.12	2.17 (0.78–6.08)	0.14
BMI (per kg/m2)	1.09 (1.05–1.13)	<0.01	1.07 (1.03–1.12)	<0.01
Triglycerides (per mg/dL)	1.01 (1.00–1.01)	<0.01	1.00 (1.00–1.01)	0.18
HOMA-IR (per 1 unit)	1.19 (1.09–1.30)	<0.01	1.09 (0.97–1.23)	0.17

Abbreviations: AMH, anti-Mullerian hormone; BMI, body mass index; HOMA-IR, homeostatic model assessment of insulin resistance; MASLD, metabolic dysfunction– associated steatotic liver disease; PCOS, polycystic ovary syndrome.

TABLE 3.

Association of low AMH with prevalent MASLD, adjusted for baseline and change in covariates from baseline to end of study (n=585)

	aOR (95% CI)	p
Low AMH (reference: non-low AMH)	2.36 (1.05–5.27)	0.04
Age >42 y (reference: ≤ 42)	0.72 (0.35–1.50)	0.39
Hysterectomy (reference: no hysterectomy)	2.53 (1.16–5.49)	0.02
BMI (per kg/m2)	1.10 (1.04–1.15)	<0.01
ΔBMI (per Δ5 kg/m2)	2.03 (1.31–3.15)	<0.01
Triglycerides (per mg/dL)	1.01 (1.00–1.01)	<0.01
ΔTriglycerides (per Δ10 mg/dL)	1.09 (1.02–1.15)	<0.01
HOMA-IR (per 1 unit)	1.04 (0.91–1.19)	0.58
ΔHOMA-IR (per 1 unit)	1.07 (1.00–1.15)	0.04

Abbreviations: AMH, anti-Mullerian hormone; BMI, body mass index; HOMA-IR, homeostatic model assessment of insulin resistance; Δ, delta; MASLD, metabolic dysfunction–associated steatotic liver disease; PCOS, polycystic ovary syndrome.

On mediation analyses, no covariates were found to be statistically significant mediators of the relationship between low AMH and prevalent MASLD including baseline BMI (Effect =–0.02, Bootstrap SE = 0.07, 95% CI: −0.15 to 0.11, p=0.82), triglyceride levels (Effect = 0.08, Bootstrap SE = 0.05, 95% CI: −0.03 to 0.18, p=0.12), or HOMA-IR (Effect = 0.00, Bootstrap SE = 0.05, 95% CI: −0.09 to 0.11, p=0.98). Similarly, visceral adipose tissue volume as measured on Year 25 CT was not a significant mediator (Effect = 0.21, Bootstrap SE = 0.15, 95% CI: −0.05 to 0.50, p=0.14).

DISCUSSION

Leveraging the large, prospective CARDIA cohort, we found that low AMH measured in reproductive-aged women was associated with more than 2-fold higher odds of prevalent MASLD in midlife. These findings remained independent of potential confounders, including age, race, BMI, insulin resistance, triglycerides, and PCOS.

There is growing recognition that metabolic disease encompasses a broad spectrum of pathologies related to biological aging or cellular senescence.³⁸,³⁹ Unlike the fixed, linear trajectory of chronological aging, biological aging is influenced by a variety of genetic and environmental factors. Recent studies have demonstrated accelerated biological aging in patients with MASLD compared to chronologically age-matched controls without liver disease.³⁸,⁴⁰ Although chronological milestones associated with risk of MASLD have been identified (eg, age >50), shifting focus to biological aging represents an opportunity to refine risk stratification in women.

Reproductive aging is a subset of biological aging that represents a decline in ovarian function, changes in hormonal regulation, and culminates in menopause.⁴¹ Indeed, menopause has been identified as a risk factor for MASLD, independent of chronological age, in part due to a decline in estradiol and the withdrawal of its antifibrotic effects.⁴² Though characteristic of reproductive aging, studying estradiol and other endogenous estrogens as biomarkers for MASLD in younger women is complex due to hormone fluctuations during the menstrual cycle. In contrast, AMH is a reproductive hormone that is more stable within the menstrual cycle and declines predictably in the years preceding menopause.¹⁵ For these reasons, AMH is feasible to study and has the potential to bridge differences in biological and chronological aging in investigations of chronic disease in premenopausal women.⁶

Low AMH levels have been linked with several metabolic conditions, including insulin resistance and cardiovascular disease, but few studies have explored AMH in the context of liver disease.⁷ A study conducted within the NASH Clinical Research Network (NASH CRN) demonstrated an inverse relationship between AMH levels and the presence and severity of steatohepatitis and fibrosis among women with known MASLD, which was independent of metabolic risk factors.¹⁰ Our results build on these findings by showing that low AMH is also associated with the risk of future MASLD. Consistent with the data from the NASH CRN, this relationship was independent of BMI, in contrast to studies of other metabolic outcomes, in which the association with AMH was largely attributable to elevated BMI.³⁶

The mechanisms connecting low AMH levels to MASLD are likely multifactorial, though one likely pathway is through metabolic dysregulation. Women with lower AMH levels are prone to insulin resistance and central adiposity, established risk factors for MASLD.⁴³,⁴⁴ Indeed, our analysis showed that the association between low AMH and MASLD was attenuated but remained significant after adjusting for insulin resistance. This suggests that low AMH may reflect an underlying metabolic phenotype that predisposes women to liver disease. It is also plausible that the observed association between low AMH and MASLD reflects the known association between menopause and steatosis.⁴⁵ As expected, participants with low AMH were more likely to transition to menopause during the study period. However, the findings are not entirely explained by the menopausal transition, as the prevalence of MASLD was not higher among low AMH participants who transitioned to menopause during the study than among low AMH participants who did not transition to menopause. Therefore, our results support the increased risks of MASLD with menopausal transition, while also supporting a potential role of low AMH as an early biomarker of MASLD risk.

The current study has some important strengths and limitations. Leveraging a large, well-characterized cohort with long-term follow-up enabled us to examine the relationship between AMH levels and subsequent MASLD. The use of CT imaging provides an objective measure of MASLD, lowering the risk of misclassification compared to self-reported outcomes or administrative coding. While we adjusted for multiple potential confounders, residual confounding by unmeasured factors, such as diet or physical activity, remains possible. There was also potential for misclassification of PCOS and menopausal onset, as these measures were self-reported. Notably, the prevalence of MASLD in the study (8.5%) may be lower than expected compared to the general population, though this can likely be explained by the younger age of this cohort and the proportion of participants who were premenopausal at the time of CT assessment. AMH measures were also obtained from a single time point, precluding our ability to evaluate the influence of AMH trajectories on MASLD risk. AMH measurement was also performed at a median age of 42 years in the CARDIA cohort; therefore, the wider spectrum of AMH levels that would be apparent in younger participants could not be assessed.¹⁶ Finally, the CARDIA cohort did not capture Hispanic ethnicity, limiting generalizability to this population with a higher risk for early onset MASLD.⁴⁶

In summary, lower AMH levels in reproductive-aged women conferred a more than 2-fold higher odds of prevalent MASLD in midlife, independent of metabolic comorbidities. While AMH is commonly used in reproductive health assessments, this test may serve as a valuable biomarker of MASLD risk in young women. Early identification of women with low AMH levels may inform targeted interventions to prevent the progression to MASLD, particularly in those with concurrent metabolic risk factors. Future studies should explore whether trajectories of AMH may predict incident MASLD or clinically significant fibrosis, as well as age-specific cutoffs to risk-stratify for liver disease in women.