AMH Predicts Your Response to IVF Drugs, Not Your Fertility

No, and the guidance is unusually blunt about it. NICE's fertility guideline NG257 carries a 2026 recommendation, 1.18.3, that reads: do not use anti-Mullerian hormone measurement as a predictor of clinical pregnancy through spontaneous conception.

The cohort evidence points the same way. Steiner's prospective time-to-pregnancy study recruited 981 women aged 30 to 44 with no history of infertility who had been trying to conceive for three months or less, and analysed the 750 who gave blood and urine samples. Of 750 analysed, the 84 with an AMH below 0.7 ng/mL conceived by 12 cycles at 84% (95% CI 70-91), against 75% (95% CI 70-79) among the 579 with normal values. The difference was not statistically significant.

Two features of that study are worth holding onto. The women were recruited from the community rather than from a fertility clinic, and they were followed from the start of trying — which is precisely the group most likely to be handed an AMH result and a bleak reading of it. The authors' own conclusion was that AMH should not be used to predict natural fertility, to exclude patients from assisted reproduction, or to predict the age of menopause.

One: how many eggs a stimulation cycle is likely to yield. AMH is made by the small, early-stage follicles, so the level in your blood tracks roughly how many of those you have. It is a headcount, and the distinction between counting eggs and describing them turns out to matter more than almost anything else on the report.

Because it reflects a standing pool rather than a monthly event, AMH can be drawn at any point in the cycle. That convenience is part of why it became popular, and a poor reason to read more into the result.

Two practical traps on the report itself:

• Units — AMH is reported in ng/mL or in pmol/L, and Indian laboratories use both. Multiply ng/mL by roughly 7.14 to get pmol/L. A result of "2" means entirely different things depending on which unit is printed, and comparing your figure to a range quoted in the other unit is a common way to frighten yourself for no reason.
• Assays — platforms differ between laboratories. If you are tracking AMH over time, use the same lab each time, or the trend is not measuring what you think it is.

AMH is read alongside your antral follicle count and your age, not on its own. Our page on what diminished ovarian reserve does and does not tell you covers how the three are weighed together, and our AMH interpreter places a specific value in the expected range for your age.

Because it does one job well, and that job matters once treatment is on the table. Broer's individual-patient-data meta-analysis pooled 28 databases covering 5,705 women having IVF and compared how well each test predicted a poor response to stimulation. Age alone reached an area under the curve of 0.61; AMH reached 0.78 and antral follicle count 0.76. Combining AMH and follicle count did not improve on either alone.

The half that rarely gets quoted is the other outcome. For predicting an ongoing pregnancy after IVF, age was the best single predictor at an AUC of 0.57 — weak — and none of the ovarian reserve tests added anything to it. The authors concluded that the clinical usefulness of these tests before IVF is limited to predicting ovarian response.

NICE draws the same line in its 2026 recommendations: AMH or antral follicle count should be used as predictors of ovarian response to inform decisions and counselling, while day-3 FSH should not be used for that purpose at all. The ASRM's 2020 committee opinion puts it as plainly as a guideline can: these markers are useful predictors of oocyte yield after controlled ovarian stimulation, but poor predictors of reproductive potential independently from age.

A very high AMH earns its place too, because it flags a risk of over-response and ovarian hyperstimulation, which changes protocol and trigger choices. Either way the prediction is about behaviour under drugs. For how AMH sits against the older day-3 panel, see what a day-3 hormone panel shows.

It counts follicles; it does not describe what is inside them. A retrospective analysis of 492 IVF and ICSI cycles in women aged 37 and over, split by AMH level, found similar fertilisation rates, similar blastocyst formation, and similar implantation, miscarriage and live-birth rates between the high and low AMH groups.

Read carefully, though, because the paper is more interesting than its title. Clinical pregnancy was lower in the low-AMH group — but the authors attribute that to there being fewer top-grade embryos available to transfer, not to the eggs being worse. Once they compared outcomes per top-grade embryo transferred, the groups looked alike. Fewer eggs, not worse eggs, and the practical consequence is about how many cycles may be needed rather than about whether they can work.

Quality tracks with age, and the two can point opposite ways. A woman of 31 with a low AMH generally has better egg quality than a woman of 42 with a high one, because the variable driving quality is not the one being measured. That is why a low AMH at 31 is a different situation from the same number at 41, even though the report looks identical.

Not on its own. The 2023 international evidence-based PCOS guideline is explicit: recommendation 1.5.3 states that serum AMH should not be used as a single test for the diagnosis of PCOS.

What changed in 2023 is narrower than it is often reported to be. AMH can now be used in adults to define polycystic ovarian morphology — the ultrasound criterion — in place of the scan. It substitutes for one of three diagnostic features; it does not become the diagnosis. The guideline also advises using either AMH or ultrasound but not both, specifically to limit overdiagnosis, and notes that where a woman already has irregular cycles and hyperandrogenism, an AMH level is not needed to diagnose PCOS at all.

So a high AMH with regular cycles and no other features is not a PCOS diagnosis, and it is not a reason to start treatment for one. Our page on PCOS and fertility sets out what the diagnosis actually requires.

Treat it as information about likely response to treatment, not as a verdict on whether you can conceive. Concretely:

• Ask for it to be read alongside your antral follicle count and your age. A single figure on its own supports considerably less than the three together.
• If it is being repeated, use the same laboratory, because assay differences can masquerade as a falling trend.
• Ask what it changes about the plan — the expected number of eggs from a cycle, and therefore whether one retrieval is likely to be enough. Those are real planning consequences.
• Ask what it does not change. It is not grounds for excluding you from treatment, and on the cohort evidence above it does not predict your chance of conceiving without treatment.
• If urgency is being argued from the AMH figure alone, that is worth questioning. Age is the variable that genuinely rewards acting sooner.

It is worth saying plainly that a low AMH lands hard. Being handed a number that sounds like a countdown, often without context and sometimes without anyone to interpret it, is frightening, and the fact that the number answers a narrower question than it appears to does not make that reaction unreasonable. The honest response is not to soften the figure but to be precise about what it measures.

Specifically:

• That you cannot conceive naturally. NICE recommends against using AMH for that prediction, and the Steiner cohort found no significant difference in conception by 12 cycles.
• That your eggs are chromosomally abnormal. AMH counts follicles and says nothing about the chromosomes inside them.
• When you will reach menopause. The Steiner authors specifically advise against using AMH to predict the age of menopause.
• That an ongoing pregnancy after IVF is less likely. In Broer's pooled data no ovarian reserve test added anything to age for that outcome, and age itself was a weak predictor.
• That you have PCOS, if the result is high. The 2023 guideline rules out AMH as a single diagnostic test.
• That anything can raise it. No intervention retrieved for this page has been shown to increase the underlying follicle pool. Some supplements alter the measured value, which is a change in a number rather than in your ovaries.

None of this makes AMH a bad test. It makes it a specific test, useful for a specific purpose, routinely asked to answer a question it was never able to answer.