Advertisement

Anti-Müllerian hormone levels among contraceptive users: evidence from a cross-sectional cohort of 27,125 individuals

Open AccessPublished:June 11, 2021DOI:https://doi.org/10.1016/j.ajog.2021.06.052

      Background

      Anti-Müllerian hormone has become the clinical biomarker-based standard to assess ovarian reserve. As anti-Müllerian hormone testing becomes more common, more individuals are seeking to interpret the values obtained while using contraceptives. To appropriately counsel women, a better understanding of anti-Müllerian hormone levels in women using different contraceptives is needed.

      Objective

      To study the association between different forms of contraceptives and anti-Müllerian levels in women of reproductive age.

      Study Design

      This is a cross-sectional study including 27,125 US-based women aged 20 to 46 years, accessing reproductive hormone results through Modern Fertility and who provided informed consent to participate in the research. Anti-Müllerian hormone levels were collected through dried blood spot card (95.9%) or venipuncture (4.1%), and previous work has shown high correlation between hormone levels collected by these 2 methods. Multiple linear regressions were run to compare anti-Müllerian hormone levels in women using contraceptives with women not on any contraceptive, controlling for age, age of menarche, body mass index, smoking, sample collection method, cycle day, and self-reported polycystic ovary syndrome diagnosis. We also analyzed whether duration of contraceptive use predicted anti-Müllerian hormone levels in users of the hormonal intrauterine device and combined oral contraceptive pill, given the size of these contraceptive groups.

      Results

      Mean anti-Müllerian hormone levels were statistically significantly lower in women using the combined oral contraceptive pill (23.68% lower; coefficient, 0.76; 95% confidence interval, 0.72–0.81; P<.001), vaginal ring (22.07% lower; coefficient, 0.78; 95% confidence interval, 0.71–0.86; P<.001), hormonal intrauterine device (6.73% lower; coefficient, 0.93; 95% confidence interval, 0.88–0.99; P=.014), implant (23.44% lower; coefficient, 0.77; 95% confidence interval, 0.69–0.85; P<.001), or progestin-only pill (14.80% lower; coefficient, 0.85; 95% confidence interval, 0.76–0.96; P=.007) than women not on any contraceptive when controlling for covariates. Anti-Müllerian hormone levels were not significantly different when comparing women not using any contraceptives to those using the copper intrauterine device (1.57% lower; coefficient, 0.98; 95% confidence interval, 0.92–1.05, P=.600). Associations between contraceptive use and anti-Müllerian hormone levels did not differ based on self-reported polycystic ovary syndrome diagnosis. Duration of hormonal intrauterine device use, but not of combined oral contraceptive pill use, was slightly positively associated with anti-Müllerian hormone levels, although this small magnitude effect is likely not clinically meaningful (coefficient, 1.002; 95% confidence interval, 1.0005–1.003; P=.007).

      Conclusion

      Current hormonal contraceptive use is associated with a lower mean anti-Müllerian hormone level than that of women who are not on contraceptives, with variability in the percent difference across contraceptive methods. These data provide guidance for clinicians on how to interpret anti-Müllerian hormone levels assessed while on contraceptives and may facilitate more patients to continue contraceptive use while being evaluated for their ovarian reserve.

      Key words

      Introduction

      Over the last few decades, women have been increasingly delaying childbearing.
      • Martin J.A.
      • Hamilton B.E.
      • Osterman M.J.K.
      • Driscoll A.K.
      Births: final data for 2018.
      ,
      • Mathews T.J.
      • Hamilton B.E.
      Mean age of mothers is on the rise: United States, 2000-2014.
      The average maternal age at first birth has been increasing, whereas the total fertility rate and the total number of children born to a woman in her lifetime keep declining.
      • Martin J.A.
      • Hamilton B.E.
      • Osterman M.J.K.
      • Driscoll A.K.
      Births: final data for 2018.
      ,
      • Mathews T.J.
      • Hamilton B.E.
      Mean age of mothers is on the rise: United States, 2000-2014.
      This trend has been paralleled by a more recent increase in the number of women who proactively seek fertility assessments or elective fertility preservation services.
      • Jones B.P.
      • Saso S.
      • Mania A.
      • Smith J.R.
      • Serhal P.
      • Ben Nagi J.
      The dawn of a new ice age: social egg freezing.
      Of these women, some are on contraceptives at the time they seek to evaluate their ovarian reserve. Stopping contraceptives due to a concern for the temporary suppression of ovarian reserve
      • Girum T.
      • Wasie A.
      Return of fertility after discontinuation of contraception: a systematic review and meta-analysis.
      can be impractical or undesirable, especially if the contraceptive method is treating an underlying condition such as polycystic ovary syndrome (PCOS) or endometriosis. Therefore, better understanding of the impact of contraceptives on reproductive hormones has an important clinical value.

       Why was this study conducted?

      There is variation among previous studies on the relationship between hormonal contraceptives and anti-Müllerian hormone levels because of small sample sizes, collapsing of contraceptive methods, and limited control for confounding factors.

       Key findings

      Anti-Müllerian hormone levels are significantly (all P<.02) lower in combined oral contraceptive, implant, progestin-only pill, and vaginal ring users than those not on contraceptives. Although anti-Müllerian hormone levels are significantly (P=.014) lower in hormonal intrauterine device users, this difference is small and likely not clinically meaningful. The duration of combined oral contraceptive use is not associated with lower anti-Müllerian hormone levels.

       What does this add to what is known?

      This study investigates contraceptive types independently, providing contraceptive-specific differences, with narrower confidence intervals, which can be used by clinicians to counsel patients. This cohort is more than 10 times as large as the next sized cohort looking at this question.
      Anti-Müllerian hormone (AMH), one of the most common hormones tested to assess fertility potential, is used as a surrogate measure for ovarian reserve that can help predict response to ovarian stimulation.
      • Barbakadze L.
      • Kristesashvili J.
      • Khonelidze N.
      • Tsagareishvili G.
      The correlations of anti-Müllerian hormone, follicle-stimulating hormone and antral follicle count in different age groups of infertile women.
      • Nelson S.M.
      • Klein B.M.
      • Arce J.C.
      Comparison of antiMüllerian hormone levels and antral follicle count as predictor of ovarian response to controlled ovarian stimulation in good-prognosis patients at individual fertility clinics in two multicenter trials.
      • Fanchin R.
      • Schonäuer L.M.
      • Righini C.
      • Guibourdenche J.
      • Frydman R.
      • Taieb J.
      Serum anti-Müllerian hormone is more strongly related to ovarian follicular status than serum inhibin B, estradiol, FSH and LH on day 3.
      • Kevenaar M.E.
      • Meerasahib M.F.
      • Kramer P.
      • et al.
      Serum anti-Müllerian hormone levels reflect the size of the primordial follicle pool in mice.
      • Broekmans F.J.
      • Kwee J.
      • Hendriks D.J.
      • Mol B.W.
      • Lambalk C.B.
      A systematic review of tests predicting ovarian reserve and IVF outcome.
      Compared with other common biomarkers, such as follicle-stimulating hormone and estradiol, AMH is more stable throughout the menstrual cycle,
      • Deb S.
      • Campbell B.K.
      • Clewes J.S.
      • Pincott-Allen C.
      • Raine-Fenning N.J.
      Intracycle variation in number of antral follicles stratified by size and in endocrine markers of ovarian reserve in women with normal ovulatory menstrual cycles.
      • Kissell K.A.
      • Danaher M.R.
      • Schisterman E.F.
      • et al.
      Biological variability in serum anti-Müllerian hormone throughout the menstrual cycle in ovulatory and sporadic anovulatory cycles in eumenorrheic women.
      Practice Committee of the American Society for Reproductive Medicine. Electronic address: [email protected], Practice Committee of the American Society for Reproductive Medicine
      Testing and interpreting measures of ovarian reserve: a committee opinion.
      more strongly correlated with antral follicle count,
      • Fanchin R.
      • Schonäuer L.M.
      • Righini C.
      • Guibourdenche J.
      • Frydman R.
      • Taieb J.
      Serum anti-Müllerian hormone is more strongly related to ovarian follicular status than serum inhibin B, estradiol, FSH and LH on day 3.
      ,
      • Van Rooij I.A.J.
      • Broekmans F.J.M.
      • Scheffer G.J.
      • et al.
      Serum antiMüllerian hormone levels best reflect the reproductive decline with age in normal women with proven fertility: a longitudinal study.
      and cycle independent, thus explaining its wide adoption as the single best predictor of ovarian reserve.
      • Tal R.
      • Seifer D.B.
      Ovarian reserve testing: a user’s guide.
      The widespread use of AMH as a marker for ovarian reserve has led to multiple studies assessing the demographic, lifestyle, and clinical factors that may influence AMH levels.
      • Jung S.
      • Allen N.
      • Arslan A.A.
      • et al.
      Demographic, lifestyle, and other factors in relation to antiMüllerian hormone levels in mostly late premenopausal women.
      • Dólleman M.
      • Verschuren W.M.M.
      • Eijkemans M.J.C.
      • et al.
      Reproductive and lifestyle determinants of anti-Müllerian hormone in a large population-based study.
      • Seifer D.B.
      • Golub E.T.
      • Lambert-Messerlian G.
      • et al.
      Variations in serum Müllerian inhibiting substance between white, black, and Hispanic women.
      In addition, the effect of contraceptives on AMH levels has been of particular interest to clinicians who counsel patients seeking a fertility assessment while on hormonal contraceptives that temporarily suppress ovarian function. The putative effect of current hormonal contraceptive use on AMH has been examined through longitudinal
      • Streuli I.
      • Fraisse T.
      • Pillet C.
      • Ibecheole V.
      • Bischof P.
      • de Ziegler D.
      Serum antiMüllerian hormone levels remain stable throughout the menstrual cycle and after oral or vaginal administration of synthetic sex steroids.
      • Kallio S.
      • Puurunen J.
      • Ruokonen A.
      • Vaskivuo T.
      • Piltonen T.
      • Tapanainen J.S.
      AntiMüllerian hormone levels decrease in women using combined contraception independently of administration route.
      • Landersoe S.K.
      • Birch Petersen K.
      • Sørensen A.L.
      • et al.
      Ovarian reserve markers after discontinuing long-term use of combined oral contraceptives.
      • Van Den Berg M.H.
      • Van Dulmen-Den Broeder E.
      • Overbeek A.
      • et al.
      Comparison of ovarian function markers in users of hormonal contraceptives during the hormone-free interval and subsequent natural early follicular phases.
      • Johnson L.N.C.
      • Sammel M.D.
      • Dillon K.E.
      • Lechtenberg L.
      • Schanne A.
      • Gracia C.R.
      AntiMüllerian hormone and antral follicle count are lower in female cancer survivors and healthy women taking hormonal contraception.
      and cross-sectional
      • Dólleman M.
      • Verschuren W.M.M.
      • Eijkemans M.J.C.
      • et al.
      Reproductive and lifestyle determinants of anti-Müllerian hormone in a large population-based study.
      ,
      • Bentzen J.G.
      • Forman J.L.
      • Pinborg A.
      • et al.
      Ovarian reserve parameters: A comparison between users and non-users of hormonal contraception.
      • Birch Petersen K.
      • Hvidman H.W.
      • Forman J.L.
      • et al.
      Ovarian reserve assessment in users of oral contraception seeking fertility advice on their reproductive lifespan.
      • Landersoe S.K.
      • Forman J.L.
      • Birch Petersen K.
      • et al.
      Ovarian reserve markers in women using various hormonal contraceptives.
      studies, and estimates have varied widely from as low as 55% below noncontraceptive users
      • Johnson L.N.C.
      • Sammel M.D.
      • Dillon K.E.
      • Lechtenberg L.
      • Schanne A.
      • Gracia C.R.
      AntiMüllerian hormone and antral follicle count are lower in female cancer survivors and healthy women taking hormonal contraception.
      to no different than noncontraceptive users.
      • Deb S.
      • Campbell B.K.
      • Pincott-Allen C.
      • Clewes J.S.
      • Cumberpatch G.
      • Raine-Fenning N.J.
      Quantifying effect of combined oral contraceptive pill on functional ovarian reserve as measured by serum anti-Müllerian hormone and small antral follicle count using three-dimensional ultrasound.
      Many studies are limited by sample size, the number and types of contraceptives included, and the ability to control for confounders.
      • Dólleman M.
      • Verschuren W.M.M.
      • Eijkemans M.J.C.
      • et al.
      Reproductive and lifestyle determinants of anti-Müllerian hormone in a large population-based study.
      ,
      • Streuli I.
      • Fraisse T.
      • Pillet C.
      • Ibecheole V.
      • Bischof P.
      • de Ziegler D.
      Serum antiMüllerian hormone levels remain stable throughout the menstrual cycle and after oral or vaginal administration of synthetic sex steroids.
      ,
      • Kallio S.
      • Puurunen J.
      • Ruokonen A.
      • Vaskivuo T.
      • Piltonen T.
      • Tapanainen J.S.
      AntiMüllerian hormone levels decrease in women using combined contraception independently of administration route.
      ,
      • Bentzen J.G.
      • Forman J.L.
      • Pinborg A.
      • et al.
      Ovarian reserve parameters: A comparison between users and non-users of hormonal contraception.
      ,
      • Birch Petersen K.
      • Hvidman H.W.
      • Forman J.L.
      • et al.
      Ovarian reserve assessment in users of oral contraception seeking fertility advice on their reproductive lifespan.
      ,
      • Bas-Lando M.
      • Rabinowitz R.
      • Farkash R.
      • et al.
      Prediction value of anti-Müllerian hormone (AMH) serum levels and antral follicle count (AFC) in hormonal contraceptive (HC) users and non-HC users undergoing IVF-PGD treatment.
      This wide range of reported effects makes it challenging to adequately counsel patients.
      Given the rise of fertility assessments among current contraceptive users, it is imperative to understand the impact of contraceptives on serum AMH values. Our study sought to evaluate the difference in AMH levels between a diverse set of contraceptive types and no contraceptive use in a large cohort of reproductive age women.

      Material and Methods

       Subjects

      Participants in this study purchased a fertility hormone test from Modern Fertility (San Francisco, CA) between June 1, 2018, and February 26, 2021. Women in the reproductive age group, between 20 and 46 years, who received their hormone results, and provided informed consent to participate in the research were included in this study. All data were de-identified before analysis. This study was approved by the Western Institutional Review Board, an ethics committee that ensures proper research consent and methodology are followed.

       Sample collection and anti-Müllerian hormone quantification

      Blood samples were collected by means of 1 of the following 2 methods: (1) dried blood spot collection card and processed by US Specialty Labs (San Diego, CA); or (2) venipuncture and processed at Quest Diagnostics (Secaucus, NJ). Both laboratories used Access AMH immunoassay by Beckman-Coulter (Brea, CA) that has a reportable measuring range of 0.08 to 24 ng/mL; limit of detection ≤0.02 ng/mL; and limit of quantitation ≤0.08 ng/mL). All samples were analyzed continuously throughout the study period.
      For the dried blood spot samples, 4 large drops of blood were collected on serum separator cards. Participants using a hormonal intrauterine device (IUD), copper IUD, progestin-only pill (POP), or no contraceptive, who self-report getting a period, were instructed to collect their sample on the third day of their menstrual cycle (range, day 2–day 4). All other participants were instructed to collect on any day of their menstrual cycle. The proportions of participants in each group collecting their sample on day 3 of the cycle are presented in the Table.
      TableSample characteristics and demographic information
      Characteristics and demographicsOverall (n=27,125)None (n=18,527)Copper IUD (n=871)Hormonal IUD (n=2779)POP (n=255)Ring (n=530)Implant (n=442)COC (n=3721)
      Age (y)31.5 (4.52)32.1 (4.65)30.7 (4.05)30.2 (3.83)30.7 (4.31)31.2 (3.68)29.0 (3.71)29.9 (3.93)
      AMH (ng/mL)3.80 (3.15)3.80 (3.16)3.80 (2.75)4.06 (3.19)3.79 (3.29)3.67 (3.14)3.83 (3.13)3.67 (3.14)
      Age of menarche (y)12.5 (1.58)12.5 (1.58)12.6 (1.51)12.6 (1.56)12.6 (1.80)12.4 (1.61)12.4 (1.68)12.5 (1.56)
      BMI (kg/m2)26.6 (6.82)27.0 (7.11)25.2 (5.79)25.9 (6.20)27.0 (6.93)26.0 (5.71)27.5 (7.00)25.5 (5.76)
      Cigarettes/mo4.36 (26.3)5.16 (28.8)3.12 (20.8)2.46 (19.4)3.85 (24.2)1.23 (14.0)5.79 (30.3)2.40 (19.0)
      Self-reported PCOS
       No PCOS24,962 (92.0)16,938 (91.4)824 (94.6)2610 (93.9)221 (86.7)494 (93.2)409 (92.5)3466 (93.1)
       PCOS2163 (8.0)1589 (8.6)47 (5.4)169 (6.1)34 (13.3)36 (6.8)33 (7.5)255 (6.9)
      Sample collection day
       Day 319,657 (72.5)18,167 (98.1)848 (97.4)558 (20.1)102 (40.0)000
       Sample day unspecified7450 (27.5)360 (1.9)23 (2.6)2221 (79.9)153 (60.0)530 (100)442 (100)3721 (100)
      Sample collection method
       Dried blood spot26,009 (95.9)17,690 (95.5)838 (96.2)2705 (97.3)242 (94.9)513 (96.8)424 (95.9)3597 (96.7)
       Venipuncture1116 (4.1)837 (4.5)33 (3.8)74 (2.7)13 (5.1)17 (3.2)18 (4.1)124 (3.3)
      Race and ethnicity
       White14,506 (53.5)9800 (52.9)467 (53.6)1577 (56.7)125 (49.0)271 (51.1)240 (56.7)2026 (54.4)
       Hispanic1331 (4.9)1016 (5.5)28 (3.2)87 (3.1)10 (3.9)22 (4.2)19 (4.3)149 (4.0)
       Asian861 (3.2)599 (3.2)25 (2.9)79 (2.8)10 (3.9)21 (4.0)8 (1.8)119 (3.2)
       Black678 (2.5)553 (3.0)11 (1.3)32 (1.2)9 (3.5)9 (1.7)13 (2.9)51 (1.4)
       Other or multiracial1906 (7.0)1336 (7.2)77 (8.8)186 (6.7)23 (9.0)43 (8.1)34 (7.7)205 (5.5)
      BMI, body mass index; COC, combined oral contraceptive; PCOS, polycystic ovary syndrome; POP, progestin-only pill.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.
      After drying completely, the cards were mailed to the laboratory for processing. Coefficient of variation varied from 3.3% to 4.5%.
      • Burke E.E.
      • Beqaj S.
      • Douglas N.C.
      • Luo R.
      Concordance of fingerstick and venipuncture sampling for fertility hormones.
      Because previous work has demonstrated that AMH levels are stable only for up to 14 days, any samples received after the 14-day window were not processed and a new sample was collected.
      • McDade T.W.
      • Woodruff T.K.
      • Huang Y.Y.
      • et al.
      Quantification of anti-Müllerian hormone (AMH) in dried blood spots: validation of a minimally invasive method for assessing ovarian reserve.
      Validation studies suggest excellent concordance between dried blood spot and venipuncture sampling for AMH (r>.97) and no statistically significant bias, suggesting that AMH values from these 2 methods can be used and interpreted interchangeably.
      • Burke E.E.
      • Beqaj S.
      • Douglas N.C.
      • Luo R.
      Concordance of fingerstick and venipuncture sampling for fertility hormones.
      ,
      • McDade T.W.
      • Woodruff T.K.
      • Huang Y.Y.
      • et al.
      Quantification of anti-Müllerian hormone (AMH) in dried blood spots: validation of a minimally invasive method for assessing ovarian reserve.
      For the analysis, values below the assay limit of quantitation (henceforth below quantifiable limit [BQL]) were replaced by the assay upper limit of quantitation (0.079 ng/mL). We handled BQL replacement in 2 additional ways to determine the robustness of our selected approach (Supplemental Table 1). After BQL values were recoded, all AMH values were natural log transformed to reduce skew and allow for normality assumptions to be met in the linear regression.

       Data collection

      After providing consent to the research, participants filled out an online questionnaire about lifestyle factors, demographics, and previous medical diagnoses. Participants did not receive compensation for responding to questionnaires, could skip any questions they chose, and could cease participation at any time. Collected data included height, weight, smoking frequency, contraceptive type, contraceptive duration of use, age at menarche, and previous diagnoses of primary ovarian insufficiency (POI) and PCOS. Participants reported their frequency of smoking on a scale ranging from 0 (never) to 7 (>6 times per day), and responses were transformed to the number of cigarettes smoked per month. The body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Demographic information on self-identified race and ethnicity, household income, and education was also collected. Participants self-identified as American Indian or Alaskan, White, Black or African American, Hispanic or Latino, Asian, Native Hawaiian or Pacific Islander, or “other.”

       Statistical analysis

      A multiple linear regression was fitted with all potential covariates of interest to predict natural log transformed AMH values. This regression included contraceptive type, age, the linear effect of BMI, number of cigarettes smoked per month, age of menarche, method of sample collection, day of sample collection (cycle day 3 vs not specified), self-reported PCOS diagnosis, and the interaction between contraceptive type and PCOS diagnosis to predict AMH. Contraceptive type was treated as a 7-level categorical variable with no contraceptive use as the reference category, yielding estimates for the effect of each contraceptive type on AMH relative to no contraceptive use. Complete data for all covariates were available for 27,125 participants (Figure 1). Compared with participants for whom complete covariate data were not available, participants who provided complete covariate data were relatively similar. Small but statistically significant differences included being approximately 6 months younger, had a BMI about 1 unit higher, experienced menarche about 1 month earlier, smoked one more cigarette per month in the participants without covariate data. They were also more likely to identify as White (75% vs 72%), have a bachelor’s degree or higher (84% vs 82%), and collect their sample through dried blood spot (96% vs 93%). No other group differences were significant (all P>.05).
      Figure thumbnail gr1
      Figure 1Participant inclusion criteria
      AMH, anti-Müllerian hormone; POI, primary ovarian insufficiency.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.
      We examined the relationship between the duration of contraceptive use and AMH across contraceptive types through visual inspection of the values and regressions. Length of contraceptive use was reported as 1 to 3 months, 4 to 6 months, 7 to 9 months, 10 to 11 months, 1 to 3 years, 4 to 5 years, 6 to 10 years, 11 to 15 years, 16 to 20 years, and >20 years. We ran a linear regression including both main and interactive effects of duration of contraceptive use and contraceptive type in addition to age, BMI, age of menarche, number of cigarettes smoked per month, PCOS diagnosis, method of sample collection, and day of sample collection. Our main analysis included the effect of the duration of contraceptive use on AMH levels for the following 2 contraceptive methods with the largest sample sizes: combined oral contraceptive (COC; n=1952), and hormonal IUD (n=1504). Visual inspection of values and regressions for contraceptives with smaller samples sizes are presented in the Supplemental Figure and in Supplemental Table 2.
      Because AMH values were natural log transformed before analysis, estimates of difference are calculated by exponentiating unstandardized beta coefficients from regressions. All analyses were performed in R (R Foundation for Statistical Computing, Austria), and all code files have been uploaded as supplementary material. Statistical significance was set a priori at α=0.05. The Bonferroni-corrected threshold, considering 6 contraceptive types to no contraceptive use, would be α=0.05/6=0.008.

      Results

       Study population

      Participants were customers of Modern Fertility who consented to participate in the research. Our main analysis was restricted to those who did not explicitly report a diagnosis of POI, for whom we had reliable data on contraceptive type, who reported using a contraceptive type reported by at least 100 women, and who provided complete covariate information (n=27,125; Figure 1). Omnibus analysis of variances (for continuous variables) and chi-square tests of independence (for categorical variables) revealed differences in age, AMH, age of menarche, self-reported PCOS diagnosis, BMI, cigarette smoking, sample collection method, day of sample collection, and race and ethnicity across contraceptive groups (Table). Because mean ages differ across contraceptive groups, it is not surprising that unadjusted AMH levels differ across contraceptive groups as well.

       Association of contraceptives with anti-Müllerian hormone

      In a regression predicting log-transformed AMH with data from 27,125 participants, age (coefficient, 0.923; 95% confidence interval [CI], 0.920–0.925; P<.001), age of menarche (coefficient, 1.0295; 95% CI, 1.01–1.03; P<.001), BMI (coefficient, 0.995; 95% CI, 0.994–0.997; P<.001), number of cigarettes smoked per month (coefficient, 0.999; 95% CI, 0.9991–0.9999; P=.009), and self-reported PCOS diagnosis (coefficient, 1.58; 95% CI, 1.51–1.65; P<.001) were significantly predictive of AMH levels. Sample collection method (coefficient, 1.04; 95% CI, 0.99–1.10; P=.105) and day of sample collection (coefficient for difference in samples obtained irrespective of cycle day relative to those obtained around cycle day 3=1.05; 95% CI, 0.99–1.12; P=.061) did not significantly predict AMH.
      In this regression, relative to women not on contraceptives, AMH levels were 7% lower in women on hormonal IUD (coefficient, 0.93; 95% CI, 0.88–0.99; P=.014), 15% lower in those on POP (coefficient, 0.85; 95% CI, 0.76–0.96; P=.007), 22% lower in vaginal ring users (coefficient, 0.78; 95% CI, 0.71–0.86; P<.001), 23% lower in those on implant (coefficient, 0.77; 95% CI, 0.69–0.85; P<.001), and 24% lower in those on COC (coefficient, 0.76, 95% CI, 0.72–0.81; P<.001). AMH in women using copper IUD did not differ significantly from AMH in women who were not on contraceptives (coefficient, 0.98; 95% CI, 0.92–1.05; P=.600) (Figure 2).
      Figure thumbnail gr2
      Figure 2AMH levels across contraceptive groups
      A, Individual values, means, and 95% confidence intervals for natural log transformed AMH values across contraceptive groups, adjusted for covariates. B, Percent difference estimates and 95% confidence intervals in AMH across contraceptive groups relative to women not on contraceptives. The asterisks represent significant differences from women not on contraceptives at P<.05. Combined oral contraceptives (COC); progestin-only pill (POP); None refers to participants who were not using contraceptives at the time of AMH measurement.
      AMH, anti-Müllerian hormone.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.
      We examined the interaction in terms of contraceptive type and self-reported PCOS diagnosis to elucidate whether the percent difference in AMH between contraceptive users and women not on contraceptives varied based on self-reported PCOS diagnosis. None of these interaction terms were significant (all P>.17).
      We performed several analyses to probe the sensitivity of these findings. First, we examined the effect of differential replacement of BQL AMH levels, by replacing BQL values with either 0.04 (the midpoint between 0 and the assay limit of quantitation) or zero. These alternate treatments of BQL values did not meaningfully affect our pattern of results (Supplemental Table 1). Next, we repeated the regression described here, adding self-reported race and ethnicity, household income, and education as covariates. Complete data were available for 18,555 participants. The inclusion of these additional covariates did not meaningfully affect our pattern of results (data not presented).

       Duration of contraceptive use

      The primary analysis on duration of contraceptive use was limited to the users of hormonal IUD (n=1504) and COC (n=1952) due to the sample sizes for these contraceptive types (exploratory analyses on duration of contraceptive use for all contraceptive types are presented in the Supplemental Material). In a regression controlling for PCOS diagnosis, age, BMI, age of menarche, cigarettes smoked per month, sample collection method, and day of sample collection, there was a significant main effect of birth control length (P=.005) and a significant interaction between contraceptive use group and birth control length (P=.001), suggesting different relationships between duration of contraceptive use and AMH levels across contraceptive use groups. We then ran follow-up regressions separately for hormonal IUD users and for COC users to probe this interaction. The relationship between duration of contraceptive use and AMH was not significant among users of the COC (P=.122) but was significant among users of the hormonal IUD (coefficient, 1.002; 95% CI, 1.0005–1.003; P=.007) (Figure 3).
      Figure thumbnail gr3
      Figure 3Duration of contraceptive use and AMH levels
      A, When controlling for covariates, AMH levels were not associated with the duration of COC use across the range from 1 month to 20 years. B, However, among users of the hormonal IUD, longer duration of use was associated with increased AMH (coefficient, 1.002; 95% CI, 1.0005–1.003; P=.007).
      AMH, anti-Müllerian hormone; COC, combined oral contraceptives; IUD, intrauterine device.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.

      Comment

       Principal findings

      This study represents the largest cross-sectional cohort ever used to evaluate the effects of current contraceptive method on AMH levels. We showed that women using the COC, implant, and vaginal ring have the largest percent difference in AMH levels than those not using contraceptives (−24%, −23%, and −22%, respectively). Women using POP and hormonal IUD had smaller mean differences in AMH levels than those who were not on contraceptives (−15% and −7%, respectively). Among COC users, duration of contraceptive use (ranging from 1 month to 20 years) was not associated with further lower AMH levels. However, duration of hormonal IUD use was positively associated with AMH levels, although the magnitude of this effect was small and likely clinically insignificant (0.19% per month).

       Results

      These results add to the previous work by providing estimates for each contraceptive type individually rather than grouping contraceptive types together or limiting analyses to a single contraceptive type. In addition, the estimates provided here have the smallest confidence intervals of previous work due to the large size of the data set.

       Clinical implications

      An increasing number of patients are testing AMH levels while on contraceptives for the purpose of oocyte cryopreservation or proactive testing of their ovarian reserve. Despite the fact that previous studies have found AMH to be of limited predictive value for estimating time to pregnancy,
      • Depmann M.
      • Broer S.L.
      • Eijkemans M.J.C.
      • et al.
      Anti-Müllerian hormone does not predict time to pregnancy: results of a prospective cohort study.
      ,
      • Steiner A.Z.
      • Pritchard D.
      • Stanczyk F.Z.
      • et al.
      Association between biomarkers of ovarian reserve and infertility among older women of reproductive age.
      providers are often asked to interpret AMH levels measured while on contraceptives for patients. Given that these women do not wish to conceive immediately, it is important that AMH levels can be measured and interpreted without interrupting safe and effective contraceptive use. Therefore, accounting for the effect of different contraceptive methods on AMH levels is not only clinically useful but necessary. Although some may infer from our findings that AMH levels should not be assessed in current contraceptive users, we suggest that our findings can increase confidence in understanding and interpreting AMH levels among current contraceptive users. In the future, contraceptive-specific AMH reference ranges can be developed to further enhance the utility of AMH levels assessed on contraceptives.
      Although statistically significant differences were reported for all hormonal contraceptives analyzed, it is worth noting that not all effects may be clinically significant. For example, the impact of the hormonal IUD (∼7%) is similar to the intra- and inter-assay variability,
      • Demirdjian G.
      • Bord S.
      • Lejeune C.
      • et al.
      Performance characteristics of the Access AMH assay for the quantitative determination of anti-Müllerian hormone (AMH) levels on the Access∗ family of automated immunoassay systems.
      and is unlikely to be meaningful in clinical decision making.
      For women of reproductive age, our observations should help with the interpretation of AMH levels while on contraceptives. Although this cross-sectional study design could not identify causal relationships, our data are suggestive of a causal relationship between ovarian suppression and hormonal contraceptive use. If results come back on the very low end of the range, clinicians may choose to discontinue contraceptive use (for 3 to 6 months)
      • Fox C.W.
      • Stanhiser J.
      • Quaas A.M.
      Evidence of profound ovarian suppression on combined hormonal contraception resulting in dramatically different ovarian reserve testing and oocyte retrieval outcomes: case report and review of the literature.
      and retest AMH levels. This is especially important if out-of-range results will change clinical management. As individual responses to contraceptive removal on AMH may vary, noncontraceptive results may be helpful for discussing oocyte cryopreservation outcomes or overall reproductive timelines in these cases. Furthermore, elevated AMH levels are associated with diagnosis of PCOS.
      • Teede H.J.
      • Misso M.L.
      • Tassone E.C.
      • et al.
      Anti-Müllerian hormone in PCOS: a review informing international guidelines.
      ,
      • Quinn M.M.
      • Kao C.N.
      • Ahmad A.K.
      • et al.
      Age-stratified thresholds of anti-Müllerian hormone improve prediction of polycystic ovary syndrome over a population-based threshold.
      Although there is no internationally accepted use for AMH in the diagnosis of PCOS at this time, incorporating AMH into the screening criteria has been discussed. If this materializes, our study may prove useful in interpreting AMH values for PCOS patients on contraception.

       Research implications

      Future research should focus on designing large-scale, prospective, longitudinal studies that include AMH levels across changes in contraceptive use because these are necessary to better investigate any causal relationships between contraceptives use and AMH. Such studies should also employ dense sampling to assess the effects of length of contraceptive use on AMH, because increasingly dense sampling will enable us to better understand the time frames and trajectories in which contraceptives may affect AMH.

       Strengths and limitations

      The strengths of this study include its large sample size, standardized AMH testing, large contraceptive subgroups, and ability to control for many confounding factors. Our study, not dissimilar to other studies, was primarily composed of women who self-identify as White, are of higher socioeconomic status, and do not smoke. Extrapolating our results to different population should be done with caution. Although it is plausible that other hormonal contraceptive methods not included in this study, such as Depo-Provera and the contraceptive patch, would follow a similar pattern, we did not have sufficient sample size to analyze these individually. Furthermore, it is possible that different formulations of hormonal contraception within a given group, such as different types of hormonal IUDs, would have different effects on AMH. Our data set did not contain this information and we are unable to study differences beyond the aggregate groups.

       Conclusions

      This study represents the largest cross-sectional cohort, to date, investigating the effect of current contraceptive use on AMH levels. Our findings show that current hormonal contraceptive use is associated with lower AMH levels.
      These results can be used by clinicians interpreting AMH levels in women on contraceptives in setting expectations for how AMH levels compare with noncontraceptive users. In the future, contraceptive-specific reference ranges can be developed to further increase the clinical utility of AMH levels measured on contraceptives.

      Acknowledgments

      The authors would like to thank all Modern Fertility customers who agreed to participate in research, which made this study possible.

      Supplemental Material

      Figure thumbnail fx1
      Supplemental FigureDuration of contraceptive use and adjusted AMH
      AMH, anti-Müllerian hormone; IUD, intrauterine device.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.
      Supplemental Table 1BQL sensitivity analysis
      ContraceptiveBQL=0.079 (main analysis)BQL=0.001BQL=0.04
      Copper IUD0.9840.9830.984
      Hormonal IUD0.932
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.9620.936
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      Implant0.766
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.782
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.766
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      POP0.852
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.8880.856
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      COC0.763
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.776
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.764
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      Ring0.779
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.819
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      0.784
      Differences relative to women not using contraceptives that were statistically significant at P<.05.
      Coefficients representing the difference in AMH levels between contraceptive users and women not using contraceptives.
      AMH, anti-Müllerian hormone; BQL, before quantifiable limit; COC, combined oral contraceptives; IUD, intrauterine device; POP, progestin-only pill.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.
      a Differences relative to women not using contraceptives that were statistically significant at P<.05.
      Supplemental Table 2Coefficients for exploratory analyses on duration of AMH and contraceptive use for contraceptive types not reported in the main text
      ContraceptiveEstimate
      Copper IUD1.002
      Implant0.999
      POP0.998
      Ring1.0002
      Coefficients represent the exponentiated linear term for duration of contraceptive use. No linear effects were statistically significant at P<.05.
      AMH, anti-Müllerian hormone; IUD, intrauterine device; POP, progestin-only pill.
      Hariton et al. Impact of contraceptives on anti-Müllerian hormone. Am J Obstet Gynecol 2021.

      References

        • Martin J.A.
        • Hamilton B.E.
        • Osterman M.J.K.
        • Driscoll A.K.
        Births: final data for 2018.
        Natl Vital Stat Rep. 2019; 68: 1-47
        • Mathews T.J.
        • Hamilton B.E.
        Mean age of mothers is on the rise: United States, 2000-2014.
        NCHS Data Brief. 2016; 232: 1-8
        • Jones B.P.
        • Saso S.
        • Mania A.
        • Smith J.R.
        • Serhal P.
        • Ben Nagi J.
        The dawn of a new ice age: social egg freezing.
        Acta Obstet Gynecol Scand. 2018; 97: 641-647
        • Girum T.
        • Wasie A.
        Return of fertility after discontinuation of contraception: a systematic review and meta-analysis.
        Contracept Reprod Med. 2018; 3: 9
        • Barbakadze L.
        • Kristesashvili J.
        • Khonelidze N.
        • Tsagareishvili G.
        The correlations of anti-Müllerian hormone, follicle-stimulating hormone and antral follicle count in different age groups of infertile women.
        Int J Fertil Steril. 2015; 8: 393-398
        • Nelson S.M.
        • Klein B.M.
        • Arce J.C.
        Comparison of antiMüllerian hormone levels and antral follicle count as predictor of ovarian response to controlled ovarian stimulation in good-prognosis patients at individual fertility clinics in two multicenter trials.
        Fertil Steril. 2015; 103: 923-930.e1
        • Fanchin R.
        • Schonäuer L.M.
        • Righini C.
        • Guibourdenche J.
        • Frydman R.
        • Taieb J.
        Serum anti-Müllerian hormone is more strongly related to ovarian follicular status than serum inhibin B, estradiol, FSH and LH on day 3.
        Hum Reprod. 2003; 18: 323-327
        • Kevenaar M.E.
        • Meerasahib M.F.
        • Kramer P.
        • et al.
        Serum anti-Müllerian hormone levels reflect the size of the primordial follicle pool in mice.
        Endocrinology. 2006; 147: 3228-3234
        • Broekmans F.J.
        • Kwee J.
        • Hendriks D.J.
        • Mol B.W.
        • Lambalk C.B.
        A systematic review of tests predicting ovarian reserve and IVF outcome.
        Hum Reprod Update. 2006; 12: 685-718
        • Deb S.
        • Campbell B.K.
        • Clewes J.S.
        • Pincott-Allen C.
        • Raine-Fenning N.J.
        Intracycle variation in number of antral follicles stratified by size and in endocrine markers of ovarian reserve in women with normal ovulatory menstrual cycles.
        Ultrasound Obstet Gynecol. 2013; 41: 216-222
        • Kissell K.A.
        • Danaher M.R.
        • Schisterman E.F.
        • et al.
        Biological variability in serum anti-Müllerian hormone throughout the menstrual cycle in ovulatory and sporadic anovulatory cycles in eumenorrheic women.
        Hum Reprod. 2014; 29: 1764-1772
        • Practice Committee of the American Society for Reproductive Medicine. Electronic address: [email protected], Practice Committee of the American Society for Reproductive Medicine
        Testing and interpreting measures of ovarian reserve: a committee opinion.
        Fertil Steril. 2020; 114: 1151-1157
        • Van Rooij I.A.J.
        • Broekmans F.J.M.
        • Scheffer G.J.
        • et al.
        Serum antiMüllerian hormone levels best reflect the reproductive decline with age in normal women with proven fertility: a longitudinal study.
        Fertil Steril. 2005; 83: 979-987
        • Tal R.
        • Seifer D.B.
        Ovarian reserve testing: a user’s guide.
        Am J Obstet Gynecol. 2017; 217: 129-140
        • Jung S.
        • Allen N.
        • Arslan A.A.
        • et al.
        Demographic, lifestyle, and other factors in relation to antiMüllerian hormone levels in mostly late premenopausal women.
        Fertil Steril. 2017; 107: 1012-1022.e2
        • Dólleman M.
        • Verschuren W.M.M.
        • Eijkemans M.J.C.
        • et al.
        Reproductive and lifestyle determinants of anti-Müllerian hormone in a large population-based study.
        J Clin Endocrinol Metab. 2013; 98: 2106-2115
        • Seifer D.B.
        • Golub E.T.
        • Lambert-Messerlian G.
        • et al.
        Variations in serum Müllerian inhibiting substance between white, black, and Hispanic women.
        Fertil Steril. 2009; 92: 1674-1678
        • Streuli I.
        • Fraisse T.
        • Pillet C.
        • Ibecheole V.
        • Bischof P.
        • de Ziegler D.
        Serum antiMüllerian hormone levels remain stable throughout the menstrual cycle and after oral or vaginal administration of synthetic sex steroids.
        Fertil Steril. 2008; 90: 395-400
        • Kallio S.
        • Puurunen J.
        • Ruokonen A.
        • Vaskivuo T.
        • Piltonen T.
        • Tapanainen J.S.
        AntiMüllerian hormone levels decrease in women using combined contraception independently of administration route.
        Fertil Steril. 2013; 99: 1305-1310
        • Landersoe S.K.
        • Birch Petersen K.
        • Sørensen A.L.
        • et al.
        Ovarian reserve markers after discontinuing long-term use of combined oral contraceptives.
        Reprod Biomed Online. 2020; 40: 176-186
        • Van Den Berg M.H.
        • Van Dulmen-Den Broeder E.
        • Overbeek A.
        • et al.
        Comparison of ovarian function markers in users of hormonal contraceptives during the hormone-free interval and subsequent natural early follicular phases.
        Hum Reprod. 2010; 25: 1520-1527
        • Johnson L.N.C.
        • Sammel M.D.
        • Dillon K.E.
        • Lechtenberg L.
        • Schanne A.
        • Gracia C.R.
        AntiMüllerian hormone and antral follicle count are lower in female cancer survivors and healthy women taking hormonal contraception.
        Fertil Steril. 2014; 102: 774-781.e3
        • Bentzen J.G.
        • Forman J.L.
        • Pinborg A.
        • et al.
        Ovarian reserve parameters: A comparison between users and non-users of hormonal contraception.
        Reprod Biomed Online. 2012; 25: 612-619
        • Birch Petersen K.
        • Hvidman H.W.
        • Forman J.L.
        • et al.
        Ovarian reserve assessment in users of oral contraception seeking fertility advice on their reproductive lifespan.
        Hum Reprod. 2015; 30: 2364-2375
        • Landersoe S.K.
        • Forman J.L.
        • Birch Petersen K.
        • et al.
        Ovarian reserve markers in women using various hormonal contraceptives.
        Eur J Contracept Reprod Health Care. 2020; 25: 65-71
        • Deb S.
        • Campbell B.K.
        • Pincott-Allen C.
        • Clewes J.S.
        • Cumberpatch G.
        • Raine-Fenning N.J.
        Quantifying effect of combined oral contraceptive pill on functional ovarian reserve as measured by serum anti-Müllerian hormone and small antral follicle count using three-dimensional ultrasound.
        Ultrasound Obstet Gynecol. 2012; 39: 574-580
        • Bas-Lando M.
        • Rabinowitz R.
        • Farkash R.
        • et al.
        Prediction value of anti-Müllerian hormone (AMH) serum levels and antral follicle count (AFC) in hormonal contraceptive (HC) users and non-HC users undergoing IVF-PGD treatment.
        Gynecol Endocrinol. 2017; 33: 797-800
        • Beckman Coulter
        Access AMH: instructions for use.
        (Available at:) (Accessed Jun. 29, 2021)
        • Burke E.E.
        • Beqaj S.
        • Douglas N.C.
        • Luo R.
        Concordance of fingerstick and venipuncture sampling for fertility hormones.
        Obstet Gynecol. 2019; 133: 343-348
        • McDade T.W.
        • Woodruff T.K.
        • Huang Y.Y.
        • et al.
        Quantification of anti-Müllerian hormone (AMH) in dried blood spots: validation of a minimally invasive method for assessing ovarian reserve.
        Hum Reprod. 2012; 27: 2503-2508
        • Depmann M.
        • Broer S.L.
        • Eijkemans M.J.C.
        • et al.
        Anti-Müllerian hormone does not predict time to pregnancy: results of a prospective cohort study.
        Gynecol Endocrinol. 2017; 33: 644-648
        • Steiner A.Z.
        • Pritchard D.
        • Stanczyk F.Z.
        • et al.
        Association between biomarkers of ovarian reserve and infertility among older women of reproductive age.
        JAMA. 2017; 318: 1367-1376
        • Demirdjian G.
        • Bord S.
        • Lejeune C.
        • et al.
        Performance characteristics of the Access AMH assay for the quantitative determination of anti-Müllerian hormone (AMH) levels on the Access∗ family of automated immunoassay systems.
        Clin Biochem. 2016; 49: 1267-1273
        • Fox C.W.
        • Stanhiser J.
        • Quaas A.M.
        Evidence of profound ovarian suppression on combined hormonal contraception resulting in dramatically different ovarian reserve testing and oocyte retrieval outcomes: case report and review of the literature.
        Fertil Steril Rep. 2020; 1: 94-98
        • Teede H.J.
        • Misso M.L.
        • Tassone E.C.
        • et al.
        Anti-Müllerian hormone in PCOS: a review informing international guidelines.
        Trends Endocrinol Metab. 2019; 30: 467-478
        • Quinn M.M.
        • Kao C.N.
        • Ahmad A.K.
        • et al.
        Age-stratified thresholds of anti-Müllerian hormone improve prediction of polycystic ovary syndrome over a population-based threshold.
        Clin Endocrinol (Oxf). 2017; 87: 733-740