Testosterone and Androstenedione Are Positively Associated With Anti-Müllerian Hormone in Premenopausal Women

Rakibul M. Islam; Robin J. Bell; Marina A. Skiba; Susan R. Davis


Clin Endocrinol. 2021;95(5):752-759. 

In This Article

Abstract and Introduction


Objective: To document associations between anti-Müllerian hormone (AMH) and circulating androgens in nonhealthcare-seeking premenopausal women.

Design: Community-based, cross-sectional study.

Setting: Eastern states of Australia.

Participants: Women aged 18–39 years not using systemic hormones, not pregnant or breastfeeding within 3 months, and not postmenopausal.

Measurements: AMH, measured by the Beckman Access 2, 2 site immunometric assay from fresh samples, and testosterone, androstenedione, dehydroepiandrosterone (DHEA) and 11-oxygenated C19 steroids, measured by liquid chromatography–tandem mass spectrometry.

Results: Data were available for 794 women, median age of 33 years (range: 18–39). 76.1% were of European ancestry and 48.2% were parous. Serum AMH was positively associated with testosterone (rho = .29, p < .001) androstenedione (rho = .39, p < .001) and DHEA (rho = .10, p = .005) but not 11-ketoandrostenedione or 11-ketotestosterone. When adjusted for age, body mass index and smoking, using quantile regression, independent positive associations remained between AMH and testosterone (β coefficient: 20.90, 95% confidence interval [CI]: 13.79–28.03; p < .001) and androstenedione (β coefficient: 5.90, 95% CI: 3.76–8.03; p < .001). The serum concentration of testosterone was greater at the top AMH quintile than other quintiles (0.56 nmol/L [range: 0.21–1.90] vs. 0.36 nmol/L [range: 0.13–0.87]; p = .001) in women with self-reported polycystic ovary syndrome.

Conclusions: The positive associations between serum testosterone and androstenedione and AMH in premenopausal women is consistent with androgens directly or indirectly influencing AMH production during follicular development. As the highest AMH concentrations are most likely to be seen in women with multifollicular ovaries, it would be expected that women with multifollicular ovaries would have higher serum testosterone. Therefore, whether hyperandrogenemia and multifollicular ovaries should be considered independent characteristics of polycystic ovary syndrome warrants review.


In premenopausal women, blood testosterone is derived from direct ovarian testosterone secretion and peripheral production from preandrogens, androstenedione and dehydroepiandrosterone (DHEA), produced by the ovaries and adrenals, and DHEA sulphate (DHEAS) which is primarily of adrenal origin. In addition, the adrenals produce 11 oxygenated steroids, which have been reported to bind and activate androgen receptors, and exhibit similar androgenic potency to testosterone.[1,2] Ovarian testosterone, from androstenedione and DHEA by the thecal cells under the control of luteinizing hormone, is important for follicular maturation.[3]

Testosterone concentrations decline with age in premenopausal women with regular menstrual cycles (eumenorrhea)[4] in parallel with declining fertility.[5] Accordingly, women with premature ovarian insufficiency have low testosterone levels[6] and androgen insufficiency has been associated with diminished ovarian reserve in premenopausal women.[7] Although it has been suggested that a decline in ovarian androgen production might contribute to infertility,[7] this has not been established.[8] Available data are primarily from studies of women with diminished ovarian reserve, who are likely to have low blood testosterone which cannot be measured with precision by the immunoassays used.[7,9] Currently the associations between ovarian testosterone production and circulating testosterone and other androgens remain poorly understood, particularly for unselected premenopausal women.

Anti-Müllerian hormone (AMH) is a glycoprotein produced by ovarian granulosa cells of recruited follicles and the serum concentration of AMH is correlated with the ovarian antral follicle count (AFC).[10] While low AMH concentrations are indicative of diminished ovarian reserve, higher concentrations are associated with multifollicular ovaries.[11] Thus, AMH can be used as a surrogate measure of AFC to further explore the associations between circulating androgens and the number of developing follicles. We have examined the associations between serum AMH and testosterone and the preandrogens and the newly identified adrenal 11-oxygenated C19 steroids, measured by liquid chromatography, tandem mass spectrometry (LC-MS/MS) in a community-based sample of premenopausal women. LC-MS/MS enables the measurement of testosterone with precision at low concentrations and accurate measurement of several steroids simultaneously in a single sample.[12] We purposefully did not limit our analysis to eumenorrheic women, or women with menstrual dysfunction, to examine the associations between AMH and androgens across a broad spectrum of each analyte.