{"paper_id":"fa1b405e-caf0-4ebe-b33f-dd6550670a2f","body_text":"Anti-müllerian hormone (AMH) is biological marker of ovarian reserve, as it is proportional to the number of developing follicles in the ovary  1 . Clinically, serum AMH reflects oocyte quantity and is used to estimate the ovarian reserve which informs the dosing of ovarian stimulation  2 . Epidemiologically, AMH levels reflect reproductive age in women and are predictive of timing of menopause  3 - 6 . Prior research has also suggested that low AMH levels may be associated with increased risk for breast cancer, as well as cardiovascular disease  7 - 10 .\nInfertility is defined as 12 months of trying to conceive without success and may be caused by a number of underlying conditions. There is emerging evidence that having a history of infertility may be associated with mortality  11  as well as risk for other chronic diseases later in life  11 - 14 . Infertility may also be associated with premature menopause; however, the existing literature on this association has been mixed  15 - 19 . While AMH is now routinely measured as part of an evaluation for fertility treatment, the association between history of infertility and AMH levels later in life is not clearly established.\nTherefore, the purpose of this study was to evaluate the association between having experienced infertility previously and AMH levels measured later in life. To assess this question, we used data on AMH levels measured on average, at age 40, among a subsample of participants in the Nurses’ Health Study II.\n\nThe NHSII began in 1989 when 116,429 registered female nurses between the ages of 25-42 returned a mailed questionnaire  20 . These participants were then followed every two years and sent a questionnaire that collected detailed information on a variety of health conditions. As has been described in detail previously  21 , in 1996-1999 participants who were premenopausal, who did not have a prior diagnosis of cancer (other than non-melanoma skin cancer), and who had not been pregnant in the last six months were invited to participate in blood collection. If participants were not on hormonal medication (menopausal hormone therapy or oral contraceptives), then they were asked to provide two timed blood collections during the follicular phase and the luteal phase of the menstrual cycle. If participants had irregular cycles, they were asked to collect their luteal sample 22 days after last menses. If the participant was using exogenous sex hormones, they were asked to provide one untimed sample. Menstrual cycle timing was confirmed by mailed postcard indicating the start date of their next menstrual period. The NHSII protocol was approved by the Institutional Review Board (IRB) of the Partners Health Care System, Boston, MA, USA and this analysis was reviewed by the IRB of the University of Arizona.\nWe defined infertility history as having ever reported \"tried to become pregnant for more than one year without success” on any NHSII questionnaire. On every questionnaire, participants reported whether they had become pregnant and the outcome of that pregnancy. Our primary comparison was women with a history of infertility prior to blood draw vs. fertile women, defined as those who had a history of gravidity and who never reported infertility prior to blood draw. Women who reported infertility were asked their age at first reported infertility, which was categorized as ≤25, 26-30, >30 years old. Participants were also asked about the cause of their infertility and were able to report multiple causes: tubal blockage, ovulatory disorder, endometriosis, cervical mucus factors, male factor infertility, not investigated, not found, and/or other. We also categorized our participants as having primary infertility (infertility occurring prior to first live birth) and secondary infertility (infertility occurring after first reported live birth). As has been done in prior analyses given the correlation between maternal age and infertility, infertility status was not updated after the age of 40 12 ,  13 . In past research within our study, participants have been able to validly recall their infertility diagnoses with high accuracy compared to medical records (ovulatory infertility, 95% concordance with medical records  22 , endometriosis, 97% concordance with medical records  23 ).\nOur outcome of AMH was measured among a subset of plasma samples previously assayed as part of four prior studies. The prior studies that measured AMH consisted of three nested case-control studies for breast cancer (n=794), ovarian cancer (n=46), and early menopause (n=820), and one study of risk factors for AMH decline (n=800). Samples were excluded if the participant had a history of breast cancer or were quality control samples (n=399), had a history of ovarian cancer (n=23), and was not premenopausal (n=1). As had been done previously, when participants were selected as controls for more than one study (n=24), we randomly selected one AMH measure  24 . Given our comparison group of fertile women, we also excluded samples from nulligravid women without infertility (n=255)\nFrom preexisting samples, AMH (ng/mL) was assayed using the same ultra-sensitive ELISA assay from ANSH Labs (picoAMH, Webster, TX) employing a quantitative sandwich enzyme immunoassay technique at Massachusetts General Hospital (ovarian cancer case-control study), Boston Children's Hospital (early menopause case-control study, AMH decline study) and Ansh Labs (breast cancer case-control study). All assays were blinded to case/control status. Blinded replicates (10% of sample) were included in each batch to allow for calculation of variation across batch and laboratory. Coefficients of variation ranged from 0.6% and 14.5% across locations  24 . If a sample was below the limit of the detection, it was replaced by a value of limit of detection divide by the square root of two  25 .\nAt blood draw, information was collected on fasting status (yes, no), time of blood collection, season of blood collection (winter, spring, fall, summer), luteal day at blood draw, and current exogenous hormone use. On biennial questionnaires, information was measured on participant height, weight, smoking history, pack-years of smoking, exogenous hormone use, oral contraceptive use history, years until menstrual cycles became regular, age at menarche, menstrual cycle regularity at age 18-22, and mid-adulthood menstrual cycle regularity in adulthood. Self-reported information on covariates used in the models was taken from the questionnaire most proximal to blood collection.\nDue to the non-normal distribution of AMH, we log-transformed continuous values of AMH. We then utilized linear regression models to estimate the percent difference in AMH and corresponding 95% confidence intervals (CI) between women with and without a history of infertility ([exp (β)-1] x 100). Using data from the blinded replicate samples, all models were adjusted for batch-to-batch variability 26 . In model 1, we adjusted multivariable models for age and age 2  at blood collection in months given the non-linear association between age and AMH, fasting status (dichotomous), time of blood collection (4 categories), season of blood collection (winter, spring, fall, summer), luteal day (4 categories) at blood collection. Model 2 we additionally adjusted for covariates we  a priori  hypothesized may be associated with infertility history and AMH: pack-years of smoking (continuous), smoking status (never, current, former), BMI (continuous), exogenous hormone use (dichotomous), duration of OC use (never, 1-23, 24-71, 72-119, ≥120 months), years until cycle became regular (never, <1, 1 - 2, 3 – 4, ≥5 years), and age at menarche (≤11, 12, ≥13 years). We also looked separately at specific infertility diagnoses (ovulatory disorders, endometriosis, cervical mucus disorder, tubal blockage, male factor, cause not investigated, cause not found, and other cause) compared to fertile participants. In these analyses participants with the specific fertility diagnosis were compared to fertile participants; participants with other types of infertility were excluded. We also grouped non-ovulatory female factor diagnoses (endometriosis, cervical mucus disorder, tubal blockage, cause not found, and other cause) and investigated them separately.\nWe investigated whether associations varied by age at first reported infertility (≤ 25, 26-30, > 30 years old) by testing for a linear trend among the three age categories. We investigated effect modification for factors associated with polycystic ovary syndrome (PCOS) associated phenotypes. Specifically, we stratified models by menstrual cycle pattern at age 18-22 (regular, irregular) and menstrual cycle pattern in adulthood (regular, irregular). We used likelihood ratio tests to test for statistically significant differences between groups  27 . Lastly, we looked separately at women who reported primary compared with secondary infertility.\nIn sensitivity analyses, to assess the impact of downstream factors closely associated with infertility we additionally adjusted for parity (0, 1, 2, 3+) and total breastfeeding duration (nulliparous or breastfed <1 month, 1-12 months, 12+ months). We also performed sensitivity analyses where we excluded participants who were taking exogenous hormones (either oral contraceptives or menopausal hormone treatments (n=25)). Lastly, we excluded individuals with AMH levels ≥ 10 ng/dL; this threshold has sometimes been used clinically to inform a diagnosis of PCOS (n=57) 28  and therefore also may be a PCOS associated phenotype.\n\nPrior to blood draw, 392 (22%) participants reported having a history of infertility ( Table 1 ). Among those with a history of infertility who received an infertility evaluation for a diagnosis, 32% reported ovulatory disorder infertility, 12% reported endometriosis, 11% reported tubal factor infertility, 18% reported male factor, and 6% reported cervical mucosal factor, 37% reported cause not investigated, 25% reported cause not found, and 14% reported other cause. On average, participants were 40 years old at blood collection. Participants with a history of infertility were more likely to have a BMI (kg/m 2 ) of ≥ 30 at the time of blood collection (17% vs. 12%) and at age 18 (8% vs. 5%), to report having a menstrual cycle that was never regular in adulthood (12% vs. 7%) and age 18-22 (23% vs. 22%), to report 21 or more pack-years of cigarette smoking among the smokers (39% vs. 24%), and to be nulliparous (15% vs. 7%).\nWe observed no statistically significant association between the history of infertility overall and levels of AMH (% difference in AMH: −8.1%, 95% CI: −19.4 to 4.8) ( Table 2 ). In subgroup analyses, when we investigated AMH levels by specific infertility diagnoses, we did not observe statistically significantly different patterns of association based on infertility diagnostic groups ( Table 3 ). History of ovulatory infertility was associated with higher AMH levels (5.7%, 95% CI: −14.7 to 30.9); while endometriosis-related infertility (−8.8%, 95% CI: −35.2 to 28.4), cervical-mucosal infertility (−22.6%, 95% CI: −51.3 to 22.9), tubal blockage (−18.3%, 95% CI: −42.5 to 16.1), and male factor (−15.3%, 95% CI:−35.5 to 11.2) had a suggestion of lower AMH levels compared to fertile participants, however none of these associations reached statistical significance. When we combined all non-ovulatory female factor infertility diagnoses, a diagnosis of infertility attributed to a female factor that was not an ovulatory disorder was associated with 11.9% lower AMH values (95% CI: −25.7 to 4.4%), however this also was not statistically significant.\nWe observed that the association with AMH varied by age at first reported infertility (p-value, test for linear trend: 0.06) ( Table 4 ). The association between history of infertility and AMH levels was strongest among those who reported age at first reported infertility > 30 years old (−17.7, 95% CI: −32.1 to −0.3) and we observed no association among those who reported age at first reported infertility <25 years old (9.1%, 95% CI: −13.6 to 37.7).\nThere was no statistically significant difference in infertility history and AMH levels by menstrual cycle regularity pattern ( Supplemental Table 1 ). Among women who reported regular menstrual cycles in adulthood, there was a suggestion that infertility history was associated with lower AMH levels (−8.8%, 95% CI: −20.6 to 4.7), but this was not statistically significant. We observed no differences in the association between primary and secondary infertility and AMH levels ( Supplemental Table 2 ).\nIn sensitivity analyses where we additionally adjusted for parity and total breastfeeding, the results attenuated slightly (−9.7%, 95% CI: −22.7 to 5.6). In sensitivity analyses where participants taking exogenous hormones at blood collection (n=25) were excluded the results did not appreciably change (−13.2%, 95% CI: −25.6 to 1.26). When we excluded 57 participants with AMH ≥ 10 ng/mL a possible PCOS associated phenotype, we observed that participants who reported a history of infertility had −16.0% lower AMH levels than gravid participant without a history of infertility (95% CI: −28.0 to −2.1).\n\nOverall, we did not observe a statistically significant association between infertility history and AMH levels. In subgroup analyses, we observed lower AMH among those who reported their first experience of infertility after age 30 and in sensitivity analyses restricted to participants with AMH <10.\nWhile AMH is routinely measured as part of an infertility evaluation, to our knowledge there has been limited research on the impact of infertility history and AMH levels outside of reproduction. Moreover, the existing literature on overall infertility history and its influence on menopausal timing and/or premature menopause has been conflicting  17  with some studies reporting that infertility history was associated with earlier age at menopause  16  and other studies reporting no association between infertility history and menopausal timing  19 . One complexity when comparing results across studies is heterogeneity in the definition of infertility. Prior research has used access to IVF treatment, infertility diagnoses, and self-reported infertility all as definitions for infertility. We know that in the NHSII approximately 65% of participants who reported infertility had their infertility clinically investigated  29  and even fewer accessed fertility treatment  30 . Recent pooled analyses of 51,000 women from nine observational studies in the InterLACE consortium observed that nulliparity was associated with a 2.26-fold increased risk of premature menopause  18 . While research from our team observed that while parity history was associated with lower risk of premature menopause in the NHSII even after adjustment for breastfeeding  31 , parity was not observed to be associated with AMH levels after adjusting for breastfeeding  21 . Prior research which pooled data across nine studies also failed to show an association between parity history and AMH levels  32 . In sum, our data suggest that the underlying cause of infertility may be a more important driver of AMH levels than omnibus information on infertility history.\nIt is well established in prior research that PCOS is associated with infertility and PCOS increases AMH levels. Indeed, we observed that participants with a history of ovulatory infertility, which included PCOS, had a suggestion of higher AMH levels than fertile participants although this association was not statistically significant. Participants categorized as having ovulatory infertility in our study would have been diagnosed prior to their blood collection which occurred between 1996-1999. During this time, most ovulatory infertility diagnoses would have been attributed to PCOS. Higher AMH for a given age has been considered a diagnostic for PCOS. Indeed, in 2023 the International Evidence-based Guideline for the Assessment and Management of PCOS added AMH as a diagnostic of polycystic ovarian morphology instead of ultrasound in adults  33 . Therefore, given the characteristics of the study population, it is possible that the suggestive positive trend between ovulatory disorder infertility and AMH is reflective of the known relation between PCOS and AMH levels.\nWe conducted several subgroup sensitivity analyses aimed at excluding participants presenting with phenotypic characteristics of PCOS. Irregular menstruation is a symptom of PCOS; therefore, we stratified our analysis by menstrual cycle regularity. Among participants who reported regular menstrual cycles in adulthood, we observed that experiencing a history of infertility was associated with a 8% lower AMH, however this difference was not statistically significant. Similarly, a statistically significant association between infertility and AMH levels was observed when we excluded women with AMH levels ≥ 10 ng/mL, which is a threshold sometimes used clinically to inform a diagnosis of PCOS  28 . After we excluded women with AMH ≥ 10 ng/mL, we observed that women with a history of infertility had a 16% lower AMH compared to fertile women. These findings from subgroup analyses suggest that once women with PCOS associated phenotypes are excluded from the analysis, overall infertility history may be associated with lower AMH levels.\nIn addition to PCOS being associated with higher AMH levels, other types of infertility may be associated with lower AMH levels. Given the timing of our data collection, we were unable to separately investigate diminished ovarian reserve (DOR) as these diagnoses would have been combined in ovulatory infertility with PCOS. However, interestingly, in multivariable-adjusted models we observed infertility was attributed to “cause not found” was associated with a suggestive trend of 15% lower AMH levels than fertile women. While this finding was not statistically significant, it does suggest that this group may be comprised of participants with undiagnosed diminished ovarian reserve. Prior research from our team has also observed that women with laparoscopically confirmed endometriosis, regardless of infertility status, had ~30% lower AMH levels than women without a history of endometriosis 34 . This is consistent with prior literature on endometriosis and menopausal timing, which observed that women with a history of endometriosis were at greater risk for earlier age at menopause  35 - 37 .\nOur study has many strengths including the fact that it is situated within a large prospective cohort study with detailed information on a number of health factors prior to blood draw. However, there also are limitations that should be considered. Specifically, there are inherent complexities in defining infertility. The standard definition for infertility is trying to become pregnant for 12 or more months without success. There are participants in the NHSII who did not try to get pregnant and therefore were not at risk of infertility. In order to overcome this potential misclassification, we restricted our analytic comparison group to fertile women with a history of gravidity and who did not report having experienced infertility. In prior research, participants with a history of infertility have been able to validly  22 ,  23  and reliably  38  recall their infertility diagnoses, further reducing the possibility of misclassification. Given the standard infertility diagnosis categorizations at the time, we were unable to fully disentangle women diagnosed with DOR from diagnoses of PCOS which may have differing influences on AMH. Additionally, our study utilized existing data measured on AMH assayed for other studies within the Nurses’ Health Study II. Cases from cancer case-control studies were excluded to reduce the potential for bias caused by this selection. While all analyses were conducted based on  a priori  scientific rationale 39 , several statistical comparisons were made as part of this analysis.\n\nGiven AMH’s association with chronic disease risk and timing of menopause, understanding AMH trajectory across the life course among those with infertility is vital to better understand how infertility may influence long-term health.","source_license":"public-domain-us","license_restricted":false}