Of the three findings that can add up to a PCOS diagnosis, the third one used to mean a transvaginal ultrasound. Since 2023 it does not have to. A blood test can stand in its place — for some people, and the people it excludes are the ones most likely to be offered it.
That exclusion is the whole reason this page exists. The change itself is simple enough to state in a sentence, and the sentence is almost always quoted without the condition attached to it.
What changed in 2023
The 2023 international guideline added serum AMH as an alternative to ultrasonography for assessing polycystic ovarian morphology — the third of the three Rotterdam criteria.
In practice that means an adult being assessed for PCOS can have the third criterion addressed by a blood draw rather than a scan. For anyone who has found that examination distressing, or waited months for an appointment with a sonographer, that is a material change rather than a technicality.
What did not change is everything around it. Rotterdam still requires two of three: irregular or absent ovulation, hyperandrogenism, and polycystic ovarian morphology. AMH replaced one route to the third finding. It did not become a test for PCOS.
The condition nobody quotes
Neither AMH nor ultrasound is recommended for diagnosis within eight years of a first period, and AMH specifically is not recommended in adolescents at all.
The reason is specificity. AMH is commonly raised in healthy young women whose cycles are still settling, so in that group a high figure does not distinguish PCOS from ordinary puberty — it mostly identifies being young. A test that returns “abnormal” for a large share of people who do not have the condition is not a diagnostic test, it is a source of false positives, and the guideline treats it as one.
Which produces a situation worth knowing about before an appointment: the group most likely to be searching for a quick blood test instead of a scan is the group the guideline excludes from using it. A raised AMH at nineteen, four years after a first period, is not the finding it would be at thirty.
Who the rule applies to
| Situation | Ultrasound | AMH |
|---|---|---|
| Adult, more than 8 years after a first period | Accepted for the third criterion | Accepted as an alternative |
| Adult, fewer than 8 years after a first period | Not recommended for diagnosis | Not recommended for diagnosis |
| Adolescent | Not recommended for diagnosis | Not recommended — poor specificity |
For adolescents the guideline does not simply remove a test and leave the criteria otherwise intact. It requires both of the other two findings — menstrual irregularity and hyperandrogenism — rather than two of three, and places anyone with only one of them in an “at risk” category for review over time rather than diagnosing them.
There is no universal AMH threshold either
If you have a result and want to know what counts as raised, the answer has the same shape as it does for testosterone: it depends on who measured it.
AMH assays are not harmonised across platforms. The same serum run on two different analysers can return materially different numbers, and the cut-off applied is the one the reporting laboratory has validated for its own method. Work is ongoing to standardise this; it is not finished.
So a threshold quoted in an article, a forum, or a comparison with someone else’s result is being applied to a measurement it was not built for. The figure beside your own result, on your own report, is the one that was.
What most articles get wrong
Three errors travel together on this topic, and each one is a step further from what the guideline says.
They present AMH as a PCOS test. It addresses one criterion out of three, and a diagnosis needs two. A raised AMH alone is one finding, and the guideline is explicit that conditions mimicking PCOS must be excluded before any of this counts.
They quote a number. Usually a single threshold, with no assay named. See above.
They drop the age condition. This is the one that does actual harm, because it is the difference between a useful test and a test that labels healthy teenagers. The eight-year rule is in the guideline for a reason that is about the biology of the years after a first period, not about caution for its own sake.
What to do with this at an appointment
If a scan has been proposed and you would rather not have one, asking whether AMH is an option is a reasonable question with a published answer behind it — provided you are an adult more than eight years past your first period.
If you already have an AMH result, the useful question is not what the number is but what it is being used for. As the third criterion it is one finding. As a fertility marker it is answering something else entirely, and the two get conflated constantly because the same test appears in both conversations.
And if you are being assessed in your teens, the guideline is on the side of waiting. Both other findings are required, “at risk” is a recognised answer rather than a non-answer, and a label applied at sixteen on a test that cannot separate PCOS from puberty is harder to remove than it was to apply.
That interval — the years of review between “at risk” and an answer — is the one place a record genuinely changes the outcome. Cycle dates, and the symptoms beside them, are what turn a wait into evidence rather than a gap.