Rotterdam Criteria (2003)

At least 2 of the 3 criteria must be present — oligo-anovulation, clinical or biochemical hyperandrogenism, and polycystic ovarian morphology (PCOM) on ultrasound — and only after other conditions with similar symptoms have been excluded. This "2 of 3" rule is what distinguishes the Rotterdam criteria (2003) from the stricter NIH 1990 definition, which required both irregular cycles and hyperandrogenism.

Because any 2 of the 3 criteria are enough, four combinations are possible. Phenotype A ("full") has all three: hyperandrogenism, ovulatory dysfunction and polycystic ovarian morphology. Phenotype B has hyperandrogenism plus ovulatory dysfunction, without polycystic morphology. Phenotype C ("ovulatory") has hyperandrogenism plus polycystic morphology with broadly regular cycles. Phenotype D ("non-hyperandrogenic") has ovulatory dysfunction plus polycystic morphology without signs of excess androgens. Phenotypes A and B carry the highest metabolic risk, which is why the phenotype influences monitoring and treatment priorities.

Yes. The 2023 International Evidence-Based Guideline (Monash/ESHRE/ASRM) kept the Rotterdam criteria as the diagnostic standard, updating the ultrasound threshold to at least 20 follicles of 2–9 mm per ovary and allowing AMH as an alternative to ultrasound in adults. The May 2026 consensus that renamed PCOS to PMOS changed the name only — the "2 of 3" rule, the exclusion work-up and the four phenotypes are unchanged.

No. Because only 2 of the 3 criteria are required, a diagnosis can be made on irregular cycles plus hyperandrogenism alone — ultrasound adds nothing in that case and the 2023 international guideline says it is then not necessary. Ultrasound (or AMH, accepted as an alternative in adults since 2023) becomes decisive when only one of the other two criteria is present. In adolescents, the guideline advises against using ovarian morphology at all within 8 years of the first period, because multifollicular ovaries are common and normal at that age.

They differ in how much weight they give to hyperandrogenism. The NIH 1990 criteria are the narrowest: they require both ovulatory dysfunction and hyperandrogenism, and ignore ovarian morphology entirely — which corresponds to Rotterdam phenotypes A and B only. The Rotterdam 2003 criteria are the broadest: any 2 of the 3 features are enough, which adds the ovulatory phenotype C and the non-hyperandrogenic phenotype D. The AE-PCOS Society 2006 criteria sit in between: they treat hyperandrogenism as mandatory, plus either ovulatory dysfunction or polycystic morphology, so they include phenotypes A, B and C but exclude D. This is why prevalence figures vary so much between studies — a cohort defined by NIH criteria is substantially smaller than the same population defined by Rotterdam. The 2023 International Evidence-Based Guideline, and the 2026 consensus that renamed PCOS to PMOS, both use Rotterdam.

Yes. Phenotype C (the "ovulatory" phenotype) is defined by hyperandrogenism plus polycystic ovarian morphology while cycles remain regular. Because 2 of the 3 criteria are enough, ovulatory PCOS / PMOS is possible — which is one reason a diagnosis should never be ruled out on cycle regularity alone.