High frequency of discordance between antimullerian hormone and follicle stimulating hormone levels in serum from estradiol confirmed days 2 to 4 of the menstrual cycle from 5354 women in US…

Frequent discordance was observed between antimullerian hormone and follicle-stimulating hormone, which increased with age in serum from 5,354 women in estradiol-confirmed cycle days 2 to 4.

Benjamin Leader, M.D., Ph.D., Aparna Hegde, M.D., Quentin Baca, Ph.D., Kimberly Stone, B.A., Benjamin Lannon, M.D., David B. Seifer, M.D., Frank Broekmans, M.D., Ph.D., Valerie L. Baker, M.D.

Volume 98, Issue 4, Pages 1037-1042, October 2012


To determine the frequency of discordance between AMH (ng/ml) and FSH (IU/L), using cut points defined by response to controlled ovarian stimulation, in the same serum sample drawn on estradiol-confirmed, menstrual cycle days 2 to 4.

Retrospective analysis.

Fertility centers in 30 US states; single reference laboratory, uniform testing protocols.

5,354 women, 20-45 years of age.


Main outcome measures:
Frequency of discordance between serum AMH and FSH values.

Of the 5,354 women tested, 1 in 5 had discordant AMH and FSH values defined as AMH<0.8 (concerning) with FSH<10 (reassuring) or AMH≥0.8 (reassuring) with FSH≥10 (concerning). Of women with reassuring FSH (n=4,469), concerning AMH values were found in 1 in 5 women in a highly age dependent fashion, ranging from 1 in 11 women under 35 years of age to 1 in 3 women above 40 years of age. On the other hand, of the women with reassuring AMH (n=3,742), 1 in 18 had concerning FSH, a frequency which did not differ significantly by age. Conclusions:
Clinical discordance in serum AMH and FSH values was frequent and age dependent using common clinical cut points, a large patient population, one reference laboratory, and uniform testing methodology. This conclusion is generalizable to women undergoing fertility evaluation, although AMH testing has not been standardized among laboratories, and the cut points presented are specific to the laboratory in this study.

  • Alberto Pacheco Castro

    It is a very interesting article, with a large number of patients included.
    But I agree with previous comments concerning the absence of some
    important data (number of AFC and MII oocyte retrieved). I
    work at the lab, and in my opinion AMH values are more consistent
    than FSH ones (at least, at increased age, similar to your data shown in Fig 2) and have much lower variability
    between different cycles, as you mention at discussion. Thank you

  • I agree with the previous commentaries both in congratulating the authors and the opinion that to have clinical data would be of great interest. Anyway, due to the lack of accuracy of both FSH and AMH in some cases and the every time more demonstrated good correlation between ovarian response and the AFC and woman age, the clinical evaluation is more important than the biochemical one in order to prognoses ovarian response. Thank you

  • This is an interesting article, well-designed with a large n. In order to do this study the authors needed to use cut-offs for AMH and FSH that divided patients into “concerning” and “reassuring”. This sharp distinction goes against the notion that ovarian reserve represents a spectrum, and gradually declines; as opposed to suddenly going from reassuring to concerning. It also may over-represent discrepancy if many patients have AMH and FSH measurements near the cutoffs for the tests. Nevertheless it is an interesting study on a clinically relevant topic that may aid in patient counseling.

  • laurenwroth

    I agree with the previous poster. It is a nice study in that it had a large population of all-comers to multiple infertility centers but all samples run with the same assay. The good study design certainly makes the findings very believable. However, having no clinical outcomes data makes it difficult to know what the clinical significance of an FSH/AMH discordance is.

  • Micah Hill

    This article is interesting in giving insight to the frequency of AMH-FSH discordance. Unfortunately, without any outcome data on the cycles, it doesn’t provide any guidance in what to do when such discordance is encountered.

Translate »