Heavy Metals and Trajectories of Anti-Müllerian Hormone During the Menopausal Transition

Abstract

Background

Experimental and epidemiological studies have linked metals with women's reproductive aging, but the mechanisms are not well understood. Disrupted ovarian folliculogenesis and diminished ovarian reserve could be a pathway through which metals impact reproductive hormones and outcomes.

Objective

The study aimed to evaluate the associations of heavy metals with anti-Müllerian hormone (AMH), a marker of ovarian reserve.

Methods

The study included 549 women from the Study of Women's Health Across the Nation with 2252 repeated AMH measurements from 10 to 0 years before the final menstrual period (FMP). Serum AMH concentrations were measured using picoAMH ELISA. Urinary concentrations of arsenic, cadmium, mercury, and lead were measured using high-resolution inductively coupled plasma mass spectrometry. Multivariable linear mixed regressions modeled AMH as a function of time before the FMP interaction terms between metals and time to the FMP were also included.

Results

Adjusting for confounders, compared with those in the lowest tertile, women in the highest tertile of urinary arsenic or mercury concentrations had lower AMH concentrations at the FMP (percent change: −32.1%; 95% CI, −52.9 to −2.2, P-trend = .03 for arsenic; percent change: −40.7%; 95% CI, −58.9 to −14.5, P-trend = .005 for mercury). Higher cadmium and mercury were also associated with accelerated rates of decline in AMH over time (percent change per year: −9.0%; 95% CI, −15.5 to −1.9, P-trend = .01 for cadmium; −7.3%; 95% CI, −14.0 to −0.1, P-trend = .04 for mercury).

Conclusion

Heavy metals including arsenic, cadmium, and mercury may act as ovarian toxicants by diminishing ovarian reserve in women approaching the FMP.

arsenic, cadmium, mercury, lead, anti-Müllerian hormone (AMH), women

Anti-Müllerian hormone (AMH) is a member of the TGF-β superfamily produced by granulosa cells of preantral and early antral follicles and is detectable in circulation (1). Over recent decades, serum AMH has received much attention because of its strong and positive correlation with the remaining number of ovarian follicles in human female ovaries, a measure of ovarian aging and fertility potential known as ovarian reserve (2). As a promising serum biomarker of ovarian reserve, measuring AMH has the advantage of being noninvasive and relatively stable over the menstrual cycle, contrary to antral follicle count and ovarian morphology, which are best measured during the follicular phase (3). Although genetic and lifestyle factors, such as smoking, are known to affect ovarian reserve, the impact of environmental exposures remains inconclusive (4).

Heavy metals, such as arsenic, cadmium, mercury, and lead, are ubiquitous in the environment. The general population can be exposed to heavy metals through drinking water, air pollution, and food contamination, so that human exposure to these heavy metals is inevitable (5). Metal ions that enter the body from the environment can bind to many molecules in body tissues and affect human health. Toxicological evidence has demonstrated that heavy metals can promote the production of reactive oxygen species and ovarian follicular atresia (6). Despite evidence from experimental studies that metals can alter ovarian function (7, 8), only a few studies have explored associations of cadmium and lead with AMH, reporting that cadmium may alter AMH concentrations in pregnant women and premenopausal women aged 30 to 45 years (9, 10). The recent development of ultra-sensitive AMH assays allows an observation of the changes in ovarian reserve and women's health nearing the menopausal transition (11). However, little is known about the impact of environmental pollutants on the trajectories of ovarian reserve in this phase of the reproductive lifespan.

We attempted to address this gap by examining the associations of heavy metals including arsenic, cadmium, mercury, and lead, with longitudinal trajectories of serum AMH concentrations as markers of ovarian reserve during the menopausal transition from samples collected in the Study of Women's Health Across the Nation (SWAN). We hypothesized that higher metal concentrations were associated with lower AMH concentrations and accelerated rates of decline in AMH during the menopausal transition.

Materials and Methods

Study Population

The SWAN is a longitudinal, multiracial/multiethnic, community-based cohort study that aims to investigate the natural history of the menopausal transition and subsequent chronic conditions among 3302 pre- and early perimenopausal women who enrolled in 1996 and 1997. A detailed description of SWAN has previously been published (12, 13). Eligibility criteria for SWAN included being aged 42 to 52 years, having an intact uterus, having had at least 1 menstrual period in the previous 3 months, no hormone use over the previous 3 months, and self-identifying with one of the site's designated racial/ethnic groups (Black in southeast Michigan, Chicago, Illinois, Pittsburgh, Pennsylvania, and Boston, Massachusetts; Hispanic in Newark, New Jersey; Chinese in Oakland, California; and Japanese in Los Angeles, California; White in all 7 sites). Of the 2694 participants who participated in SWAN follow-up visit 03 (1999-2000), 368 women from Chicago and 278 from Newark were excluded because these sites did not collect urine, the biospecimen used for heavy metal assessment. We further excluded 648 women with insufficient volumes of serum or urine samples, resulting in 1400 women from 4 racial/ethnic groups (including White, Black, Chinese, and Japanese) for the SWAN Multi-Pollutant Study (MPS). SWAN follow-up visit 3 will be referred to as the SWAN-MPS baseline hereafter. The study design of the SWAN MPS is described elsewhere (5, 14, 15). To be eligible for hormone assays, women had to have a documented date of the final menstrual period (FMP), experienced a natural menopausal transition without hysterectomy and bilateral oophorectomy, and without hormone therapy. After the FMP, AMH became undetectable for most women. Therefore, the final analytic sample included 549 women with 2252 observations from 10 to 0 years before the FMP. Each participating site approved the research protocol. Each participant received a complete description of the study and provided written informed consent before enrollment. A flow chart of the current study is shown in Supplementary Fig. S1 (all supplementary materials are located in a digital research materials repository (16)).

Heavy Metals

Urine samples at the MPS baseline were collected before 9 Am and stored −80 °C without thawing in the SWAN Repository until the metal assessment was conducted. Urinary concentrations of arsenic, cadmium, mercury, and lead were measured using high-resolution, inductively coupled plasma mass spectrometry (Thermo Scientific, Waltham, MA, USA) by the Applied Research Center of NSF International (Ann Arbor, Michigan), a part of the Michigan Children's Health Exposure Analysis Resource Laboratory Hub. The laboratory methods and quality control procedures have been previously reported (5). The limits of detection (LOD) (detection frequencies) were 0.3 μg/L (100%) for arsenic, 0.06 μg/L (94.5%) for cadmium, 0.05 μg/L (99.7%) for mercury, and 0.1 μg/L (97.8%) for lead. For metal concentrations below LODs, a value equal to the LOD divided by the square root of 2 was assigned.

Anti-Mullerian Hormone

Fasting blood samples were collected approximately annually on menstrual cycle days 2 through 5 and stored at −80 °C without thawing until AMH measurements were made. For women with irregular menstrual cycles, blood samples were collected on a random day within the 30 days after the date of the annual visit. AMH was measured with the picoAMH ELISA (AnshLabs, Webster, TX, USA; Catalog # AL-124, RRID: AB_2783675). The intraassay coefficients of variation ranged from 2.5% to 5.1%, and interassay coefficients of variation ranged from 3.4% to 4.9%. The LOD was 1.85 pg/mL (17). For AMH concentrations below the LOD, a value equal to the LOD divided by the square root of 2 was assigned.

Covariates

Age, race/ethnicity, education, financial hardship, and a food frequency questionnaire (FFQ) were obtained at SWAN's baseline visit. Information on all other variables were obtained at each follow-up visit. Education was classified into less than high school graduation, high school graduate, some college, and college graduate/postcollege. Financial hardship was asked with the question, “How hard is it for you to pay for the very basics?” and classified into very hard, somewhat hard, and not hard at all. The FMP was defined retrospectively as the date of the last bleeding episode followed by 12 months of amenorrhea without hysterectomy, bilateral oophorectomy, or hormone therapy. We then determined age at the FMP. Because reproductive age was used as the time scale instead of calendar time, time in the current analysis was defined on the reproductive age scale as age at each visit relative to age at the FMP. For parity, women were categorized as either parous or nulliparous. Smoking status was classified into never smokers, former smokers, and current smokers. Body mass index (BMI) was determined as the ratio of measured weight to height squared (kg/m²). Rice and seafood consumption, major sources of heavy metals, were collected using a detailed semiquantitative FFQ adapted from the Block FFQ (18). To account for measurement error in urinary concentrations of metals resulting from urine dilution, urinary creatinine was measured using the Cobas Mira analyzer (Horiba ABX, Montpellier, France). We used a Directed Acyclic Graph to show the hypothesized relations between heavy metals, confounders, and AMH (Supplementary Fig. S2) (5, 16, 19-21).

Statistical Analyses

Medians (interquartile ranges [IQRs]) for continuous measures and percentages for categorical measures of participant characteristics were calculated. Generalized additive mixed effects models with penalized spline were constructed to capture the shapes of AMH trajectories during the menopausal transition. AMH was log-transformed to normalize the distribution. Because the trajectories are approximately linear (Supplementary Fig. S3) (16), linear mixed effects models with a linear term for time to the FMP were used in the analyses. Linear mixed models are well-suited for handling unbalanced longitudinal data by accounting for both within-person and between-person variations (22, 23). We adjusted for age at the FMP, race/ethnicity, study site, education, smoking status, parity, BMI, and time of metal measurement. Urinary creatinine was also adjusted in all models to control for individual variations in urinary dilution (24). We chose education, as opposed to financial hardship, as a proxy measure of socioeconomic status because education had less missing data and exhibited stronger associations with metals compared with financial hardship (5). Random intercepts were included to account for correlations in AMH within each participant.

Metal concentrations were categorized into tertiles to capture potential nonlinear dose-response relationships. To estimate the associations between metals and the rates of change in AMH concentrations, we included interaction terms of metals with linear terms of time to the FMP in the models. Differences in AMH concentrations at the FMP and rates of change in AMH concentrations were calculated comparing the second and third tertiles to the first tertile (the reference group). Percent changes and 95% CIs were computed by back-transformed (exponentiating) β coefficients. As a sensitivity analysis, we further adjusted for rice and seafood consumption, which have been identified as important determinants of heavy metal concentrations and may be linked to AMH concentrations through related lifestyle factors such as smoking (5, 20). We also conducted a sensitivity analysis with further adjustment for financial hardship to account for potential residual confounding by socioeconomic status. All statistical tests were performed at the 2-sided significance level of 0.05. The statistical analyses were performed using SAS version 9.4 (SAS Institute, Inc.) and R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/).

Results

Participant Characteristics

The age range at SWAN-MPS baseline was 45 to 56 years. On average, participants completed 4 visits with available AMH measurements. The median age at the FMP was 53.0 (IQR: 51.3, 54.4) years, with the median time to the FMP of −2.1 years (or 2.1 years before the FMP) at SWAN-MPS baseline (Table 1). Of 549 participants included in the present study, 246 (44.8%) were White, 118 (21.5%) were Black, 82 (14.9%) were Chinese, and 103 (18.8%) were Japanese. Participants tended to have a college education or higher (50.2%), were never smokers (64.2%), and gave birth to at least 1 child (78.1%). AMH was generally detectable until the FMP, with a significant decline observed in AMH concentrations when women approached their FMP (Table 2, Supplementary Fig. S3) (16). Median AMH concentration was 1994.0 (IQR: 892.4, 3056.9) pg/mL in women at approximately 10 years before the FMP, whereas it decreased to 2.4 (IQR: <LOD, 14.5) pg/mL at the FMP (Table 2). Table 3 presents the distribution of urinary metal concentrations by tertiles.

Table 1.

Participant characteristics at the Multi-pollutant Study baseline (n = 549)

Characteristics	Median (IQR) or N (%)
Urinary concentrations, μg/L
Arsenic	15.0 (6.8, 38.8)
Cadmium	0.4 (0.2, 0.8)
Lead	1.2 (0.7, 2.4)
Mercury	0.8 (0.5, 1.3)
Age at FMP, y	53.0 (51.3, 54.4)
Time to the FMP, y	-2.1 (-4.2, -0.5)
Body mass index, kg/m2	25.9 (22.4, 31.7)
Race/ethnicity
White	246 (44.8%)
Black	118 (21.5%)
Chinese	82 (14.9%)
Japanese	103 (18.8%)
Study site
Michigan	107 (19.5%)
Boston	96 (17.5%)
Oakland	127 (23.1%)
Los Angeles	147 (26.8%)
Pittsburgh	72 (13.1%)
Education
<High school	102 (18.7%)
High school graduation	170 (31.1%)
Some college	136 (24.9%)
College graduate/postcollege	138 (25.3%)
Smoking status
Never smoked	352 (64.2%)
Past smoker	148 (21.0%)
Current smoker	48 (8.8%)
Parity
Nulliparous	120 (21.9%)
Parous	429 (78.1%)

Abbreviations: FMP, final menstrual period; IQR, interquartile range.

Table 2.

Serum concentrations, median (IQR) and detection rates of AMH from women without hormone therapy during the menopausal transition and postmenopause

Year to FMP	AMH
	N	>LODa, %	Median (IQR), pg/mL
-10	28	100	1994.0 (892.4-3056.9)
-9	47	100	1039.0 (504.0-1729.2)
-8	78	97.9	821.5 (378.9-1363.8)
-7	148	96.2	545.7 (256.5-1065.1)
-6	193	95.3	346.0 (175.3-698.4)
-5	263	93.3	261.7 (109.5-445.3)
-4	323	94.3	161.4 (45.61-308.3)
-3	369	87.0	99.2 (16.6-224.3)
-2	446	81.6	38.2 (6.0-126.5)
-1	439	70.9	11.3 (<LOD-50.3)
0	402	52.4	2.4 (<LOD-14.5)
+1	37	27.4	<LOD (<LOD-2.2)
+2	31	10.8	<LOD
+3	12	3.2	<LOD
+4	4	8.3	<LOD
+5	2	0	<LOD
+6	2	0	<LOD
+7	2	0	<LOD

Abbreviations: AMH, anti-Mullerian hormone; FMP, final menstrual period; IQR, interquartile range; LOD, limit of detection.

^aThe LOD was 1.85 pg/mL for AMH.

Table 3.

Descriptive statistics of urinary metal concentrations (ng/mL) at SWAN visit 03 (1999-2000)

Metals	LOD	% > LOD	Median (IQR)	Tertiles
				Tertile 1	Tertile 2	Tertile 3
				Min-Max, μg/L	Min-Max, μg/L	Min-Max, μg/L
Arsenic	0.3	100	15.0 (6.8, 38.8)	0.42-9.65	9.79-30.83	31.28-2983.79
Cadmium	0.06	94.5	0.4 (0.2, 0.8)	<LOD-0.27	0.27-0.61	0.61-22.96
Mercury	0.05	99.7	1.2 (0.7, 2.4)	<LOD-0.86	0.87-1.87	1.88-26.60
Lead	0.1	97.8	0.8 (0.5, 1.3)	<LOD-0.52	0.52-0.93	0.94-21.23

Abbreviations: IQR, interquartile range; LOD, limit of detection; SWAN, Study of Women’s Health Across the Nation.

Metals in Relation to AMH Trajectories

Higher urinary concentrations of arsenic and mercury were associated with lower serum AMH concentrations at the FMP (Table 4). After adjusting for age at the FMP, race/ethnicity, study site, education, smoking status, parity, BMI, time of metal measurement, and urinary creatinine, compared with those in the lowest tertile, women in the highest tertile of metal concentrations had 32.1% (95% CI, −52.9 to −2.2) lower AMH for arsenic (P-trend = .03), and 40.7% (95% CI, −58.9 to −14.5) lower AMH for mercury (P-trend = .005) at the FMP. By contrast, no main effects of cadmium or lead were found with AMH concentrations.

Table 4.

Percent changes (95% CIs) in AMH concentrations and rates of changes by tertiles of urinary metal concentrations from multivariable linear mixed models

Metals	Main effects				Interaction effects
	Tertile 1	Tertile 2	Tertile 3	P-trend	Tertile 1	Tertile 2	Tertile 3	P-trend
	Percent change (95% CI)	Percent change (95% CI)	Percent change (95% CI)		Percent change per year (95% CI)	Percent change per year (95% CI)	Percent change per year (95% CI)
Arsenic	Ref	−8.3 (-35.5 to 30.4)	−32.1 (-52.9 to -2.2)	.03	Ref	0.09 (-7.0 to 7.7)	0.8 (-6.5 to 8.6)	.84
Cadmium	Ref	−11.1 (-37.3 to 25.9)	−19.4 (-44.7 to 17.5)	.25	Ref	−7.0 (-13.6 to 0.0)	−9.0 (-15.5 to -1.9)	.01
Mercury	Ref	−27.0 (-48.2 to 2.9)	−40.7 (-58.9 to -14.5)	.005	Ref	−3.7 (-10.6 to 3.6)	−7.3 (-14.0 to -0.1)	.04
Lead	Ref	−6.9 (-34.6 to 32.5)	−16.2 (-43.3 to 24.0)	.37	Ref	−0.8 (-7.8 to 6.7)	−1.2 (-8.3 to 6.4)	.75

Models were adjusted for age at the FMP, race/ethnicity, study site, education, smoking status, parity at baseline, BMI, time to the FMP, urinary creatinine, and time of metal measurement. Interaction terms between metal and time to the FMP were also included in the models.

Abbreviations: AMH, anti-Mullerian hormone; BMI, body mass index; FMP, final menstrual period.

Higher urinary concentrations of cadmium and mercury were associated with accelerated rates of decline in AMH during the menopausal transition. Compared with those in the lowest tertile, women in the highest tertile of metal concentrations had 9.0% per year (95% CI, −15.5 to −1.9) higher rates of decline for cadmium (P-trend = .01) and 7.3% per year (95% CI, −14.0 to −0.1) higher rates of decline for mercury (P-trend = .04). No significant interaction existed for arsenic or lead. Results from sensitivity analyses with additional adjustments for rice and seafood consumption (Supplementary Table S1) and financial hardship (Supplementary Table S2) were similar to those presented in our main analysis (16).

Discussion

To our knowledge, this is the first study that has examined exposure to heavy metals in relation to longitudinal trajectories of AMH concentrations in midlife women during the menopausal transition. We used the ultrasensitive picoAMH ELISA to detect AMH concentrations as a marker of ovarian reserve, an estimate of the number of oocytes remaining in the ovaries and thus the change in ovarian function that occurs as this number declines, approaching the final menstrual period (19). A progressive decrease in oocyte number and decline in oocyte functional capacity together constitute ovarian aging, a potential independent risk factor for health outcomes during midlife and beyond. Therefore, trajectories of AMH decline can serve as a biochemical proxy for ovarian aging with potential associations of decreasing ovarian function with these health outcomes.

Ovarian aging as estimated by AMH levels during the menopausal transition have been linked with women's health in mid- and late-life, particularly metabolic disturbances. A study of 697 parous women (mean age, 36.7 years; 27% of women aged ≥40 years) from Project Viva reported that low AMH (median AMH, 1.97 ng/mL) at baseline was associated with greater adiposity concurrently and across approximately 9 years of follow-up, suggesting low AMH was a useful marker of metabolic risk across midlife (25). Recent analyses of SWAN data revealed that lower premenopausal AMH levels and greater declines in AMH over the menopausal transition were associated with greater total cholesterol and high-density lipoprotein cholesterol, which might cause a greater atherosclerotic risk after menopause (26). A cross-sectional study with 3841 healthy premenopausal women also showed significant associations between AMH levels and breast cancer risk factors, including obesity in women aged ≥40 years whose geometric mean levels of AMH were 3.99, 1.28, and 0.26 pg/mL for age groups 40 to 44, 45 to 49, and 50 to 57 years, respectively (27).

The approach of using reproductive aging also offers a broader perspective on the potential impact of environmental exposure to heavy metals during this critical life stage. We report that higher urinary arsenic and mercury concentrations were associated with lower AMH concentrations at the FMP. Consistent with this, higher cadmium and mercury concentrations were also related to accelerated rates of decline in AMH over time. Our findings suggest that these heavy metals may diminish ovarian reserve in midlife women during the menopausal transition. Furthermore, the observed magnitude of associations between heavy metals and AMH was stronger than the association between smoking and AMH, which is a known risk factor for depleted ovarian reserve (28). Given that heavy metals are widespread in the general population and urinary metal concentrations measured in our study were comparable to the general female population across the United States (5), the potential adverse effects of heavy metals on ovarian function should be of significant public health concern.

Although these data reveal a negative association of heavy metals with AMH, it remains unclear whether it is due to a direct effect on oocytes with resultant loss of ovarian follicles or an indirect effect on AMH production or secretion from granulosa cells in those follicles. Moreover, whereas decline in AMH has been associated with diminished reproductive outcomes, it has not been directly associated with other health outcomes associated with aging (29, 30). These questions will require more targeted studies of the direct effects of heavy metals on ovarian function and the long-term effects of heavy metal exposure and earlier decline in ovarian reserve on adverse health outcomes.

The association between arsenic and lower AMH is consistent with a previous case-control study that reported urinary arsenic to be associated with lower AMH in younger women that included 169 cases with primary ovarian insufficiency and 209 healthy controls (age range, < 20-45 years) (31). Our previous research in SWAN revealed that arsenic was associated with earlier natural menopause in 1082 premenopausal women during a median follow-up of 4.1 years (32); another cross-sectional study of 210 women in Bangladesh reported that those with arsenic skin lesions experienced menopause approximately 2 years earlier than those without (33). Collectively, these findings underscore the potential role of arsenic in advancing ovarian aging.

Previous cross-sectional studies evaluating the associations of cadmium with AMH yielded inconsistent results (9, 10). One cross-sectional study found that cadmium was negatively correlated with AMH in 283 younger Asian women aged 30 to 45 years (10). However, in a sample of 117 pregnant women (mean age, 26.5 years) from Ukraine and Greenland, Christensen et al reported a positive association between cadmium and AMH (9). The inconsistent findings regarding the association between cadmium and AMH concentrations in cross-sectional studies may be attributed to differences in age distributions among study populations. It is also plausible that cadmium is associated with the rate of change in AMH concentrations rather than the concentrations themselves, which is supported by results of our longitudinal cohort study. Further research, encompassing diverse age ranges and using longitudinal designs, is necessary to elucidate better the relationship between cadmium and AMH.

In this study, we found no association between lead and AMH concentrations; however, it is noteworthy that our previous findings in SWAN revealed associations of lead with lower estradiol concentrations and higher FSH during the menopausal transition (34), as well as earlier timing of menopause in midlife women (32). These findings suggest that, although lead exposure may not directly affect AMH concentrations, it could still play a role in other aspects of ovarian function and aging. A deeper understanding of the mechanisms underlying these associations and potential interactions of such exposures may help better inform the adverse effects of lead exposure on ovarian aging.

Last, we observed a novel association between mercury exposure and both lower AMH concentrations and a faster rate of decline of AMH during the menopausal transition. To our knowledge, this association has not been explored in prior studies. Our results suggest that, if mercury exposure remains consistent over time, it is possible that even more significant differences in AMH concentrations could be observed at younger ages. These findings warrant further longitudinal studies that include younger women to elucidate the relationship between mercury exposure and AMH concentrations across various stages of reproductive aging.

Although the precise mechanisms of heavy metals as potential ovarian toxicants have yet to be elucidated, mechanistic studies support the biological plausibility. Arsenic exposure may cause oxidative stress by generating reactive oxygen species and disrupting ovarian folliculogenesis (35-37). For instance, rats exposed daily to 0.4 and 4 ppm of arsenic via drinking water for 28 days had increased oxidative stress, induced ovarian cell apoptosis, and diminished ovarian steroidogenic enzyme activities (36, 37). Ommati et al also found a dose-response relationship of arsenic exposure (0, 0.2, 2, and 20 ppm of drinking water) with increased levels of autophagy in the oocytes in female rats (35). Furthermore, animal studies also demonstrated an association of chronic mercury exposure with prolonged estrous cycles and increases in atretic and cystic ovarian antral follicles in female rats, possibly through ovarian oxidative stress (38, 39).

Cadmium exposure may also promote free radical generation in ovaries, decrease follicular growth, and induce follicle atresia in rats (40, 41). Although toxicological evidence supports an association between lead exposure and diminished ovarian reserve (42, 43), we did not observe significant results for urinary lead. Given that lead is predominantly stored in bone, bone lead has been considered a biomarker of cumulative exposure to lead (44). Thus, our results of urinary lead and AMH need to be interpreted with caution, and future studies using bone lead concentration as a long-term exposure biomarker are warranted.

The present study had several strengths. First, to our knowledge, this is the largest study to date on this topic. Second, the longitudinal measurements of AMH and the use of time to the FMP as the time scale enabled us to make inferences based on reproductive aging instead of chronologic aging. Because ovarian reserve is a measure of reproductive aging rather than chronologic aging, it is essential to recognize that inferences based on a reproductive time scale provide valuable information regarding the independent effects on the change in gonadal function (45-47). Additionally, we focused on the associations in the years before the FMP, which is a critical window for healthy aging in women. Finally, the generalizability of our findings is enhanced by the ethnically diverse population and the metals concentrations in the SWAN cohort, which are comparable to women of the same age in the U.S. general population (5).

Several limitations should also be acknowledged. First, the findings from generally healthy women may not be applicable to women with health problems. Women who received hysterectomy or oophorectomy or those who received hormone therapy were not included in this cohort because the possible influence of these or other disease conditions on AMH concentrations cannot be excluded. Additionally, these data may not be applicable to younger women, given the relatively low AMH concentrations in midlife women approaching their FMP. Other limitations to be considered are the diverse group of heavy metals we measured in urine samples. Urinary arsenic mainly reflects recent exposures to inorganic and organic arsenic. Additional measurements of arsenic species will be critical to providing a better understanding of arsenic metabolism in relation to ovarian function. Furthermore, sensitivity of the AMH assay is a potential limitation because AMH concentrations decline markedly with the approach of the FMP. Nonetheless, the picoAMH ELISA test performed well at determining concentrations of AMH in women reaching their last menstrual period, making it feasible to capture the role of metals in ovarian aging at low AMH concentrations. Finally, because of the observational nature of our study, residual confounding cannot be entirely ruled out, even though we controlled for numerous potential confounders. For instance, although “education” was used as a proxy for socioeconomic status in our analysis, it does not capture all aspects of socioeconomic status. Future research is warranted to include more comprehensive measures.

In summary, we found arsenic, cadmium, and mercury to be associated with lower AMH concentrations and/or accelerated rates of AMH decline during the menopausal transition. This information may enable researchers to address adverse health outcomes known to be associated with metals and with reproductive hormone changes (eg, premature menopause, bone loss and osteoporosis, increased risks of cardiovascular disease, cognitive decline, vasomotor symptoms). Our findings require further investigation, particularly in a younger population, to understand fully the role of heavy metals as potential ovarian toxicants that diminish ovarian reserve. Future studies are needed to confirm the reliability of our findings and explore these associations in a more diverse and extensive population.

Acknowledgments

The authors thank the study staff at each site and all the women who participated in SWAN.

Clinical Centers: University of Michigan, Ann Arbor—Carrie Karvonen-Gutierrez, principal investigator (PI) 2021-present, Siobán Harlow, PI 2011-2021, MaryFran Sowers, PI 1994-2011; Massachusetts General Hospital, Boston, MA—Sherri-Ann Burnett-Bowie, PI 2020-present; Joel Finkelstein, PI 1999-2020; Robert Neer, PI 1994-1999; Rush University, Rush University Medical Center, Chicago, IL—Imke Janssen, PI 2020-present; Howard Kravitz, PI 2009-2020; Lynda Powell, PI 1994-2009; University of California, Davis/Kaiser—Elaine Waetjen and Monique Hedderson, PIs 2020-present; Ellen Gold, PI 1994-2020; University of California, Los Angeles—Arun Karlamangla, PI 2020-present; Gail Greendale, PI 1994-2020; Albert Einstein College of Medicine, Bronx, NY—Carol Derby, PI 2011-present, Rachel Wildman, PI 2010-2011; Nanette Santoro, PI 2004-2010; University of Medicine and Dentistry—New Jersey Medical School, Newark—Gerson Weiss, PI 1994-2004; and the University of Pittsburgh, Pittsburgh, PA—Rebecca Thurston, PI 2020-present; Karen Matthews, PI 1994-2020.

NIH Program Office: National Institute on Aging, Bethesda, MD—Rosaly Correa-de-Araujo 2020-present; Chhanda Dutta 2016- present; Winifred Rossi 2012-2016; Sherry Sherman 1994-2012; Marcia Ory 1994-2001; National Institute of Nursing Research, Bethesda, MD—Program Officers.

Central Laboratory: University of Michigan, Ann Arbor—Daniel McConnell (Central Ligand Assay Satellite Services).

NIA Biorepository: Rosaly Correa-de-Araujo 2019-present; SWAN Repository: University of Michigan, Ann Arbor—Siobán Harlow 2013-2018; Dan McConnell 2011-2013; MaryFran Sowers 2000-2011.

Coordinating Center: University of Pittsburgh, Pittsburgh, PA—Maria Mori Brooks, PI 2012-present; Kim Sutton-Tyrrell, PI 2001-2012; New England Research Institutes, Watertown, MA—Sonja McKinlay, PI 1995-2001.

Steering Committee: Susan Johnson, Current Chair

Chris Gallagher, Former Chair

Funding

The Study of Women's Health Across the Nation (SWAN) has grant support from the National Institutes of Health (NIH), Department of Health and Human Services, through the National Institute on Aging (NIA), the National Institute of Nursing Research (NINR), and the NIH Office of Research on Women's Health (ORWH) (Grants U01NR004061, U01AG012505, U01AG012535, U01AG012531, U01AG012539, U01AG012546, U01AG012553, U01AG012554, U01AG012495, and U19AG063720). The study was also supported by the SWAN Repository (U01AG017719). This publication was supported in part by the National Center for Research Resources and the National Center for Advancing Translational Sciences, NIH, through UCSF-CTSI Grant Number UL1 RR024131.

This study was also supported by grants from the National Institute of Environmental Health Sciences (NIEHS) R01-ES026578, R01-ES026964, and P30-ES017885, and by the Centers for Disease Control and Prevention (CDC)/National Institute for Occupational Safety and Health (NIOSH) grant T42-OH008455.

The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the NIA, NINR, ORWH, or the NIH.

Disclosures

The authors declare no competing interest.

Data Availability

Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

Abbreviations

 
• AMH

  anti-Müllerian hormone

 
• BMI

  body mass index

 
• FFQ

  Food Frequency Questionnaire

 
• FMP

  final menstrual period

 
• IQR

  interquartile ratio

 
• LOD

  limit of detection

 
• MPS

  Multi-Pollutant Study

 
• SWAN

  Study of Women's Health Across the Nation

Author notes

Ning Ding and Xin Wang contributed equally to this work.