2023ESHRE/ASRM International Guideline — AMH-inclusive
Rotterdam Criteria
2023 update: AMH ≥3.4 ng/mL on a validated assay now qualifies as PCOM in adults. Diagnosis requires exclusion of thyroid dysfunction, hyperprolactinemia, and NCCAH first.
0 of 3 criteria selected
Awaiting Criteria
Select at least one of the three Rotterdam criteria above to see the diagnostic result.
Guidelines & Evidence
Verified
Last Review: 2026-07-17
When to Use
When to Use
Reproductive-age patient with oligomenorrhea (cycles >35 days) or amenorrhea (>3 missed cycles).
Clinical signs of hyperandrogenism: hirsutism (mFG ≥4–6), acne, androgenetic alopecia.
Biochemical hyperandrogenism: elevated total or free testosterone on a morning fasting sample.
Infertility evaluation for suspected anovulatory cause.
Polycystic ovarian morphology found incidentally on pelvic ultrasound or high AMH in the context of reproductive symptoms.
Screening for metabolic comorbidities in at-risk women (obesity, insulin resistance, family history of T2DM).
Core Diagnostic Rule
PCOS requires ≥2 of 3 Rotterdam criteria PLUS exclusion of other causes. This is not a standalone positive test — differential diagnosis must be completed before applying the label.
Who Should NOT Be Diagnosed by Rotterdam Alone
Adolescents within 2 years of menarche — physiological anovulation is normal; do not use ultrasound criterion at all.
Patients with confirmed non-classic congenital adrenal hyperplasia (NCCAH) — serum 17-OHP must be checked first.
Known hyperprolactinemia — prolactin elevation independently causes oligo-anovulation mimicking PCOS.
Active thyroid dysfunction — TSH must be normal before attributing menstrual irregularity to PCOS.
Androgen-secreting tumor suspected (rapid-onset virilisation, testosterone >150 ng/dL) — requires imaging before PCOS diagnosis.
Postmenopausal women — Rotterdam criteria are not applicable.
Mandatory Exclusionary Workup
| TSH | Exclude hypo/hyperthyroidism |
| Prolactin (fasting) | Exclude hyperprolactinemia; draw before pelvic exam |
| 17-OHP (follicular phase) | Exclude non-classic CAH; >200 ng/dL requires ACTH stim test |
| Total testosterone | Tumor screen if >150–200 ng/dL or rapidly progressive virilisation |
| DHEA-S | Adrenal androgen source (elevated in adrenal tumors, mild elevation in PCOS) |
How it Works
Criterion 1 — Oligo/Anovulation
Defined as cycles >35 days or <9 menstrual periods per year. In women with regular cycles, a luteal-phase progesterone <3 ng/mL confirms anovulation. Primary amenorrhea (no period by age 15) is a separate diagnosis. Quantify cycle history for at least the preceding 12 months — transient irregularity post-pill use or after significant weight change may confound.
Criterion 2 — Hyperandrogenism
01
Clinical: Modified Ferriman-Gallwey (mFG) score ≥4–6 for hirsutism (threshold varies by ethnicity — East Asian women score lower for equivalent androgen burden). Acne and androgenetic alopecia are supportive but less specific.
02
Biochemical: Elevated free testosterone (calculated or by equilibrium dialysis) is most sensitive. Total testosterone > upper limit of normal for the assay. SHBG is often low in PCOS, amplifying free androgen fraction. Androstenedione may be added in borderline cases.
03
Biochemical testing is most accurate when drawn fasting, in the morning, and in the follicular phase (days 2–5). Avoid testing while on hormonal contraception — it suppresses androgens and produces false negatives.
Criterion 3 — Polycystic Ovarian Morphology (PCOM)
01
2003 original threshold: ≥12 follicles (2–9 mm) in either ovary OR ovarian volume >10 mL (excluding dominant follicle or corpus luteum).
02
2018/2023 update (high-frequency transvaginal probe ≥8 MHz): ≥20 follicles per ovary OR ovarian volume >10 mL in either ovary. FNPO ≥20 replaces the 12-follicle rule when modern equipment is used.
03
AMH as PCOM surrogate (2023 guideline): AMH ≥3.4 ng/mL (25 pmol/L) on Beckman Coulter Access 2 assay (age-specific thresholds apply). AMH is reproducible, cycle-independent, and preferred when transvaginal ultrasound is not acceptable to the patient.
04
Do not use transabdominal ultrasound for PCOM diagnosis — resolution is insufficient for accurate follicle counting.
The Four PCOS Phenotypes
| Phenotype A (Classic Full) | Hyperandrogenism + Anovulation + PCOM — highest metabolic risk |
| Phenotype B (Classic Non-PCO) | Hyperandrogenism + Anovulation, no PCOM — similar risk to A |
| Phenotype C (Ovulatory) | Hyperandrogenism + PCOM, regular cycles — milder metabolic risk |
| Phenotype D (Non-androgenic) | Anovulation + PCOM, no hyperandrogenism — controversial, mildest risk |
Insulin Resistance — Not a Diagnostic Criterion, But Central
Insulin resistance is present in 65–80% of PCOS patients regardless of BMI. Hyperinsulinaemia drives ovarian androgen synthesis (LH-mediated) and suppresses SHBG. It is the dominant modifiable pathophysiological driver — but is absent from the Rotterdam criteria because no validated clinical cutoff exists. HOMA-IR >2.5–3.0 is used in practice.
Clinical Pearls
Adolescent Diagnostic Trap
In girls within 2 years of menarche, anovulatory cycles are physiologically normal. The 2023 international guideline requires BOTH hyperandrogenism AND anovulation to diagnose PCOS in adolescents — PCOM by ultrasound should not be used. A provisional diagnosis can be applied and confirmed at 8 years post-menarche.
Normal-Weight PCOS
Up to 30% of PCOS patients have normal BMI. Lean women with PCOS have the same prevalence of insulin resistance as obese women when adjusted for body fat distribution, though often at lower absolute HOMA-IR values. Do not reassure lean patients that metabolic workup is unnecessary — visceral adiposity is not captured by BMI.
AMH: The Emerging Fourth Criterion
AMH correlates with antral follicle count and is 2–4x elevated in PCOS. The 2023 guideline formally accepts AMH as an alternative to ultrasound for PCOM in adult women on validated assays. AMH is not confounded by the pill (unlike androgens), does not vary with cycle day, and is measurable in patients declining transvaginal ultrasound. Age-specific thresholds are required — AMH declines steeply with age.
Long-Term Risks Often Missed at Diagnosis
Endometrial hyperplasia and cancer — chronic anovulation causes unopposed estrogen stimulation of the endometrium; risk reduced by inducing regular withdrawal bleeds.
Type 2 diabetes — 3–10× increased lifetime risk vs. controls; screen with 2-hour 75 g OGTT (preferred over HbA1c, which underestimates risk in PCOS).
Gestational diabetes — significantly elevated risk even in women with confirmed ovulation on treatment.
Obstructive sleep apnoea — 5–30× higher prevalence in PCOS, often underdiagnosed; inquire about daytime somnolence and snoring.
Depression and anxiety — 2–3× higher prevalence; the diagnostic delay (average 2 years) compounds psychological burden.
Non-alcoholic fatty liver disease (NAFLD/MASLD) — independent of BMI; screen with ALT and liver ultrasound in high-risk patients.
Diagnosing PCOS in Women on Hormonal Contraception
Combined oral contraceptives normalise cycles, suppress androgens, and may reduce PCOM. It is not possible to apply Rotterdam criteria reliably while a patient is on hormonal contraception. The 2023 guideline recommends waiting 3 months after discontinuation before formal assessment — or applying clinical judgement based on pre-pill history and other available features.
Letrozole vs. Clomiphene — 2023 Update
Letrozole is now first-line for ovulation induction in PCOS (Legro et al., NEJM 2014; confirmed in 2023 guideline). Live birth rate 27.5% vs 19.1% with clomiphene (p=0.007). Mechanism: aromatase inhibition reduces estrogen-mediated GnRH suppression, promoting more physiological FSH rise. Clomiphene retains second-line status when letrozole is unavailable.
Next Steps
Step-by-Step Management by Presenting Concern
01
All patients: Lifestyle modification (even 5% weight loss restores ovulation in 55–100% of overweight women). 150 min/week moderate activity minimum. Mediterranean-pattern diet. Address OSA if suspected.
02
Metabolic screening (all): 2-hour 75 g OGTT, fasting lipid panel, blood pressure, waist circumference. Repeat OGTT every 3 years if normal.
03
Irregular cycles / endometrial protection: COCP first-line (any low-androgenic progestin). If COCP contraindicated or declined: cyclic progestogen for ≥12 days every 3 months to induce withdrawal bleed. Metformin as adjunct in metabolic phenotype.
04
Hirsutism / acne: COCP minimum 6 months before assessing response. Add spironolactone 50–200 mg/day for inadequate response (use reliable contraception — teratogenic). Finasteride second-line. Topical eflornithine for facial hirsutism.
05
Infertility — first-line: Letrozole 2.5–7.5 mg days 3–7 of cycle. Add timed intercourse or IUI. Reassess at 6 cycles.
06
Infertility — if letrozole fails: Ovarian drilling (laparoscopic) or low-dose FSH if drilling declined. IVF if bilateral tubal factor or male factor coexists.
07
Insulin resistance / T2DM risk: Metformin 1500–2000 mg/day (titrate from 500 mg to reduce GI side effects). Does not reliably restore ovulation as monotherapy — do not use as sole fertility treatment.
08
Pregnancy: Screen for GDM at 24–28 weeks. Monitor for hypertensive disorders. Higher ovarian hyperstimulation risk with ART.
Metformin — When to Use
| Metabolic phenotype (high HOMA-IR) | First-line with lifestyle; reduces T2DM progression by 40% |
| Menstrual regulation | Adjunct to COCP; not first-line alone |
| Ovulation induction | Inferior to letrozole; combination may add modest benefit |
| Letrozole + Metformin in ART | Reduces OHSS risk; recommended in high-responder phenotype |
Monitoring Schedule
Annual: blood pressure, waist circumference, fasting glucose/HbA1c, lipid panel.
Every 3 years: 75 g OGTT if initially normal.
Ultrasound for endometrial thickness if >6 months amenorrhea without progestogen cover.
Reassess fertility goals at each visit — plan changes management pathway.
The Evidence
Original Rotterdam Consensus
Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome.
Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. • Fertility and Sterility.. 2004;81(1):19–25. Established the three-criterion framework by expanding the 1990 NIH criteria to include polycystic ovarian morphology. Enabled recognition of four distinct PCOS phenotypes.
View Source2023 International Evidence-Based Guideline
Recommendations from the 2023 international evidence-based guideline for the assessment and management of polycystic ovary syndrome.
Teede HJ et al. • Fertility and Sterility.. 2023;120(4):767–793. Current global standard. Key updates: AMH as PCOM surrogate, updated ultrasound thresholds (FNPO ≥20), letrozole as first-line fertility treatment, mental health and lifestyle integration, revised adolescent criteria.
View SourceLetrozole vs. Clomiphene for Infertility
Letrozole versus clomiphene for infertility in the polycystic ovary syndrome.
Legro RS et al. • New England Journal of Medicine.. 2014;371(2):119–129. n = 750. Live birth rate: letrozole 27.5% vs clomiphene 19.1% (p=0.007). Established letrozole as superior first-line ovulation induction agent in PCOS.
View SourceAMH as Diagnostic Criterion
The physiology and clinical utility of anti-Müllerian hormone in women.
Dewailly D et al. • Human Reproduction Update.. 2014;20(3):370–385. Established the biological basis for AMH as PCOM surrogate; demonstrated correlation with antral follicle count and ovarian reserve in PCOS.
View SourceKey References
Origins & History
From NIH 1990 to Rotterdam 2003
The 1990 NIH criteria required both clinical hyperandrogenism and chronic anovulation — making PCOS essentially an androgenic anovulation syndrome. This excluded a significant population of women with polycystic ovaries and anovulation without overt androgen excess (Phenotype D), and women with regular cycles but polycystic morphology and hyperandrogenism (Phenotype C). The Rotterdam workshop, held in the Netherlands in 2003 and jointly convened by ESHRE and ASRM, added ultrasound morphology as the third criterion, expanding the diagnostic population and capturing the full spectrum of PCOS presentations.
Historical Timeline
01
1935: Stein and Leventhal describe bilateral polycystic ovaries with amenorrhea, obesity, and hirsutism — the original Stein-Leventhal syndrome.
02
1980s: LH hypersecretion and insulin resistance identified as central pathophysiological mechanisms.
03
1990: NIH criteria published — requires hyperandrogenism + anovulation only. Excludes PCO morphology as standalone criterion.
04
2003: Rotterdam ESHRE/ASRM workshop — expands to three criteria, enables four phenotypes, ~60% increase in diagnosed population vs NIH criteria.
05
2006: AES (Androgen Excess Society) position: hyperandrogenism required for PCOS diagnosis — partially contradicts Rotterdam by excluding Phenotype D.
06
2012: NIH Evidence-Based Methodology Workshop endorses Rotterdam over NIH 1990 — current standard in US practice.
07
2018: Update to follicle count threshold (≥20 per ovary with modern probes ≥8 MHz).
08
2023: International guideline formalises AMH as PCOM surrogate, updates adolescent criteria, confirms letrozole as first-line fertility treatment.
Why PCOS Is Still Controversial
The Rotterdam criteria have been criticised for medicalising normal physiological variation — PCO morphology is present in ~20% of reproductively normal women. Phenotype D (anovulation + PCOM without hyperandrogenism) in particular lacks the metabolic risk profile of Phenotypes A and B, raising questions about whether it represents a distinct syndrome or a normal variant. The 2023 guideline does not resolve this debate but strengthens the evidence base for management across all phenotypes.
Last Comprehensive Review: 2026-07-17