AMH and egg count: from a test result to a cycle plan

AMH and antral follicle count predict how many eggs one stimulation cycle will yield — not whether those eggs become a baby. Use them to estimate eggs per cycle and how many cycles you need; use your age and your banked mature (MII) egg count to estimate live-birth odds. Units matter: 1 ng/mL equals 7.14 pmol/L.

What AMH actually measures, and in which units

Anti-Müllerian hormone is made by the granulosa cells of small growing follicles. Because its production does not depend on the pituitary hormones that drive a menstrual cycle, AMH stays relatively consistent within and between cycles, which is why it largely replaced day-3 FSH as the standard reserve marker (American Society for Reproductive Medicine, Testing and interpreting measures of ovarian reserve, 2020). One practical caveat from the same document: AMH can read lower in women currently on hormonal contraception, so a result taken on the pill should be interpreted with caution and, if it changes your plan, repeated off it.

Two units are in circulation and they are not interchangeable on sight. US labs usually report ng/mL; UK, European and Australian labs usually report pmol/L. The conversion is ng/mL × 7.14 = pmol/L. So 1.0 ng/mL is 7.14 pmol/L, and a UK result of 10 pmol/L is about 1.4 ng/mL. Every number below is given in both, and every AMH figure you collect should be written down with its unit and the lab that produced it.

AMH has a partner test. Antral follicle count is the sum of antral follicles across both ovaries on transvaginal ultrasound in the early follicular phase, counting follicles 2–10 mm across (ASRM 2020). Order both. They are usually correlated but, by ASRM's own summary of the evidence, AMH and AFC disagree in up to 30% of cases — and there is no agreed rule for what to do when they do.

Is there a normal AMH level by age?

Not in the way most people want. There is no single published reference range you can hold your result against, for two separate reasons, and both matter more than the age question itself.

First, the assays disagree. Different manufacturers' platforms return different values for the same blood. Xu and colleagues built a cross-platform conversion tool in PeerJ (2023) precisely because, in their words, a result from one hospital's AMH kit "cannot be interpreted by other hospitals utilizing a different AMH kit" — and they note that even the proposed international standard has been valued differently across assays, from 282 to 1,157 ng per ampoule. Comparing a Roche result to a Beckman result, or to a number a friend quotes, is not a comparison.

Second, age acts on AMH as a slope, not a band. In the largest US dataset of its kind, 17,120 women tested at fertility centers, median AMH fell by roughly 0.2 ng/mL per year through age 35 and then by about 0.1 ng/mL per year after that (Seifer, Baker and Leader, Fertility and Sterility, 2011). Your result sits on a continuum, and what a given value implies at 31 is not what it implies at 41.

What does exist is a set of published action thresholds — values at which named bodies expect a particular stimulation response.

AMH bands, expected stimulation response, and what to plan for

Note what every row of that table is about: how many eggs come out. None of these bodies attaches a live-birth probability to an AMH value, because none of the underlying evidence supports one.

Why AMH does not predict whether you will have a baby

This is the part clinic marketing tends to blur. ASRM's 2020 committee opinion states the position in one sentence: markers of ovarian reserve "can be useful as predictors of oocyte yield following controlled ovarian stimulation and oocyte retrieval. However, they are poor predictors of reproductive potential independently from age."

The evidence behind that is specific. In the Time to Conceive cohort of 750 women aged 30–44 with no known infertility risk factors, those with low AMH (below 0.7 ng/mL) had similar cumulative pregnancy rates at 6 and 12 cycles of trying as women with normal values. In the EAGER trial's 1,202 women, AMH below 1 ng/mL produced similar cumulative pregnancy rates to values of 1.0–3.5 ng/mL. ASRM summarizes both, and adds that AMH is "only weakly predictive of pregnancy at best" for qualitative outcomes.

The American College of Obstetricians and Gynecologists reached the same conclusion for the general population in Committee Opinion 773 (2019): a single AMH measurement in women with presumed fertility "does not appear to be useful in predicting time to pregnancy and should not be used for counseling patients in this regard."

What does predict live birth from frozen eggs is a different pair of variables. In NYU's fifteen-year autologous thaw series of 543 patients, "age at cryopreservation and the number of M2s thawed were predictive" of live birth (Cascante et al., Fertility and Sterility, 2022) — age when the eggs were banked, and how many mature eggs there were. Those two inputs are exactly what our egg freezing success rate calculator asks for, and AMH is deliberately absent from it. The methodology page shows why: the published model runs on age and mature egg count, and adding a marker the evidence does not support would make the output look more personal without making it more true.

Where this is contested: the yield claim itself is not universal. In a London cohort of 373 women doing elective egg freezing, AMH and AFC predicted oocyte yield on univariate analysis but lost significance once age, BMI, estradiol and follicle count at trigger were controlled for (Kasaven et al., Archives of Gynecology and Obstetrics, 2022). The authors' own explanation is that elective freezers are young and healthy, unlike the diminished-reserve populations most AMH research is built on. Read AMH as a useful planning input with a wide error bar, not as a measurement.

From an AMH result to a cycle plan, in four steps

The order matters, because each step feeds the next and only the last one is about odds.

1. Get AMH and AFC together, with units and lab recorded. If they disagree, say so out loud and ask which one your clinic is dosing from.
2. Convert to expected mature (MII) eggs per cycle. Not all retrieved eggs are mature, and only mature ones get vitrified. In the London cohort above, at a median age of 38.3 and a median AMH of 9.9 pmol/L (about 1.4 ng/mL), the median cycle produced 8 eggs retrieved and 6 mature. Ask your clinic for its own predicted number in MII, not in "eggs."
3. Set the egg target for your age. On the Goldman 2017 model behind this site, a 75% modeled chance of at least one live birth needs about 12 mature eggs at 35, 20 at 37, 24 at 38 and 38 at 40. The full table is on how many eggs to freeze.
4. Divide. Target ÷ expected MII per cycle = cycles. Six mature eggs per cycle against a target of 24 at 38 is four cycles, not one — and knowing that before you start is the entire point of testing AMH. Then, once eggs are actually in the tank, put the real MII count into the calculator.

That arithmetic is often uncomfortable, and it should be done before you pay for cycle one rather than after. A target you cannot reach in two or three retrievals is information, not a verdict: it may point toward freezing sooner, budgeting for more cycles, or discussing donor eggs.

Where AMH genuinely changes the plan

None of the above makes the test pointless. AMH earns its place in four decisions: the starting gonadotropin dose, the trigger and monitoring plan, whether to budget for multiple cycles up front, and how urgently to act. A high result is the one that carries a safety implication rather than a scheduling one — hyperstimulation risk sits at the top of the range, which is also why egg freezing with PCOS is a protocol conversation rather than a reassurance about odds.

Two questions are worth asking verbatim at your consultation: how many mature eggs do you expect from me per cycle, and what is that estimate based on? and what is your lab's thaw survival rate for eggs frozen at my age? More of them are on our clinic questions list.

The bottom line on AMH and egg count

AMH is a supply forecast. It tells you, roughly and with a real error bar, how many eggs a stimulation cycle is likely to produce, which tells you how many cycles you need and what this will cost. It does not tell you whether you can conceive naturally, and it does not tell you what your frozen eggs will do years from now. For that second question the honest inputs are your age on the day of retrieval and the number of mature eggs in the tank — so run the forecast first, bank the eggs, then run the odds.