O-24: Diphasic pattern of anti-mullerian hormone (AMH) in ovulatory cycles as evidenced by means of an ultra-sensitive assay: New insights into ovarian function

Abstract

AMH levels have been recognized for a decade as a reliable index of ovarian reserve. However the assay initially developed for testis investigation by Josso et al (JCEM 1990) was designed for the determination of high AMH levels in diluted serum, but was not sensitive enough for the measurement of low AMH levels in diluted samples. On the other hand, a significant matrix effect was observed with neat serum, resulting in underestimation of low AMH levels. As a consequence, the first generation AMH assay was unable to find any significant change in AMH levels during normal menstrual cycles. The aim of the study was to reassess ovarian secretion of AMH by means of a new ultrasensitive assay (DSL, Houston, Texas), suitable for precise and free of matrix effect measurements on diluted female serum samples. AMH as well as other ovarian hormone and gonadotropin levels were measured in women enrolled in a multicenter international trial. They received daily for 3 months either a placebo or the progesterone receptor antagonist VA 2914 (HRA Pharma, Paris, France). The trial was approved by the local Ethics Committee and conducted in compliance with the Declaration of Helsinki The subjects were sampled every 3 days (D) throughout the 3rd cycle of treatment. FSH and LH were measured by means of AutoDelfia reagents (PerkinElmer France), estradiol and progesterone by RIA (DiaSorin-France and CIS biointernational, France, respectively,) and inhibin A by Elisa (Oxford Bioinovation-DSL-France) In 12 placebo treated ovulatory cycles, AMH levels (pmol / l, mean ± sem) exhibited a diphasic pattern, with a peak on D -7 relative to LH peak : 26 ± 3.2, a nadir on D0 : 19.1 ± 3.5 , and a second peak in luteal phase on D10 after LH peak : 25.4 ± 4.3. These variations were significant (p<0.05). There were no correlations between AMH levels and FSH, LH, inhibin A, or progesterone levels. Interestingly AMH levels were negatively correlated with estradiol levels : r = 0.27, p < 0 .02. As expected since both hormones are indexes of luteinization, inhibin A and progesterone levels were strongly correlated : r = 0.62, p <0.001. In 21 anovulatory cycles observed in women treated with the progesterone receptor antagonist, AMH levels exhibited a quite different pattern : no significant changes from D3 to D14 of cycle, significantly higher levels in the second phase of the cycle (D18-D28) : 28.2 ± 2.2 relative to the first phase (D3-D14) : 24.2 ± 1.9, p<0.05. Again no correlations were found between AMH and gonadotropin or ovarian hormone levels. In addition AMH and estradiol levels were not correlated in these anovulatory cycles. It should be noted however that estradiol levels in the second phase of the anovulatory cycles were lower than in luteal phase Thanks to this new AMH assay, we report for the first time a diphasic pattern of AMH levels throughout the menstrual cycle, contrasting with previous reports of unsignificant changes during the cycle. In addition the inverse relationship of estradiol and AMH in ovulatory cycles suggest a possible role of estradiol in the regulation of AMH secretion by the ovary.