Recommendations from the international evidence‐based guideline for the assessment and management of polycystic ovary syndrome

This is not a single experiment but a clinical practice guideline. The authors tested the following clinical questions through systematic evidence reviews:

**Diagnostic criteria:** What are the most accurate and practical criteria for diagnosing PCOS across different age groups (adolescents vs adults)?

**Assessment:** What tests (hormonal, metabolic, ultrasound) are necessary and which are unnecessary?

**Lifestyle interventions:** What is the evidence for diet, exercise, and weight management in improving PCOS outcomes?

**Pharmacological treatments:** What is the evidence for metformin, oral contraceptives, anti-androgens, and fertility medications (clomiphene, letrozole, gonadotropins)?

**Psychological interventions:** What is the evidence for screening and managing depression, anxiety, and quality of life?

**Fertility management:** What is the safest and most effective approach to ovulation induction and assisted reproduction?

The comparators varied by question but included placebo, no treatment, alternative drugs, different doses, and different lifestyle approaches. Outcome measures included menstrual regularity, ovulation rates, pregnancy rates, hirsutism scores (modified Ferriman-Gallwey scale), acne, metabolic markers (fasting glucose, insulin, lipids), weight/BMI, and psychological measures (depression, anxiety, quality of life).

The guideline synthesises evidence from hundreds of studies involving tens of thousands of women with PCOS across multiple continents. Specific populations varied by question:

**Diagnosis studies:** Included women aged 15–45 years from general gynaecology and endocrinology clinics, with some adolescent-specific data (ages 12–18).

**Lifestyle studies:** Included overweight/obese women with PCOS (BMI >25 kg/m²), typically aged 18–40, from outpatient settings in Australia, Europe, North America, and Asia.

**Pharmacological studies:** Included women with PCOS seeking fertility treatment (ages 20–40) and women with PCOS seeking symptom management (hirsutism, acne, menstrual irregularity).

**Psychological studies:** Included women with PCOS of all ages, with some studies specifically recruiting from dermatology or fertility clinics.

The guideline development process itself involved 37 professional societies and organisations covering 71 countries, with multidisciplinary experts from paediatrics, endocrinology, gynaecology, primary care, reproductive endocrinology, obstetrics, psychiatry, psychology, dietetics, exercise physiology, and public health, alongside consumer representatives.

The guideline used standardised outcome measures across the included studies:

**Diagnosis:** Modified Rotterdam criteria (presence of 2 of 3: oligo/anovulation, clinical/biochemical hyperandrogenism, polycystic ovaries on ultrasound). For adolescents, specific criteria requiring both oligo/anovulation AND hyperandrogenism.

**Hirsutism:** Modified Ferriman-Gallwey (mFG) score (0–36 scale, higher = more hair growth, with ethnic-specific cut-offs; ≥4–6 considered abnormal depending on ethnicity).

**Menstrual regularity:** Self-reported cycle length (normal 21–35 days), number of menses per year.

**Metabolic markers:** Fasting glucose, fasting insulin, HOMA-IR (homeostatic model assessment of insulin resistance), oral glucose tolerance test (OGTT), lipid profile (total cholesterol, LDL, HDL, triglycerides).

**Ovulation:** Serum progesterone (≥3 ng/mL in mid-luteal phase), urinary ovulation predictor kits, ultrasound monitoring of follicular development.

**Pregnancy outcomes:** Clinical pregnancy rate, live birth rate, miscarriage rate.

**Psychological outcomes:** Depression screening (PHQ-9, 0–27 scale), anxiety screening (GAD-7, 0–21 scale), quality of life (PCOSQ, a 26-item disease-specific questionnaire).

**Ultrasound:** Transvaginal ultrasound for follicle number per ovary (FNPO) and ovarian volume. Updated criteria: ≥20 follicles per ovary or ovarian volume ≥10 mL (using modern high-frequency transducers).

**Study design:** This is an international clinical practice guideline developed using the AGREE II framework (Appraisal of Guidelines for Research and Evaluation) and the GRADE framework (Grading of Recommendations, Assessment, Development, and Evaluation). It is not a primary study but a synthesis of existing evidence.

**Process:** The guideline development involved:

20 face-to-face meetings over 15 months

60 prioritised clinical questions

40 systematic reviews and 20 narrative reviews

5 guideline development groups (GDGs) covering: diagnosis, assessment, lifestyle, pharmacological, and fertility

Consumer and translation committees with direct involvement of women with PCOS

International feedback and peer review before final approval

**Evidence grading:** Each recommendation was assigned a strength (strong or conditional/weak) based on:

Quality of evidence (high, moderate, low, very low)

Balance of benefits vs harms

Patient values and preferences

Resource implications

Feasibility and acceptability

**What this design can and cannot prove:**

**Can prove:** The guideline provides a structured, transparent synthesis of the best available evidence. It identifies where evidence is strong (e.g., letrozole is superior to clomiphene for ovulation induction) and where evidence is weak (e.g., optimal dietary composition for PCOS).

**Cannot prove:** The guideline cannot establish causality for any single intervention because it relies on the underlying studies, many of which are observational or small RCTs. The low-to-moderate quality of most PCOS evidence means many recommendations are based on expert consensus rather than high-certainty data. The guideline also cannot account for individual variability — what works for the average woman in a trial may not work for you.

**Major methodological weaknesses:**

Overall evidence quality is low to moderate, meaning many recommendations are GRADE "conditional" rather than "strong"

Heterogeneity across studies in diagnostic criteria, outcome measures, and populations

Limited long-term follow-up data (most lifestyle studies are 6–12 months)

Underrepresentation of certain ethnic groups and adolescents

Potential publication bias (positive results more likely to be published)

Industry funding for some pharmacological studies (though guideline panels attempted to manage conflicts of interest)

**Diagnosis:**

The Rotterdam criteria (2 of 3 features) remain the international standard for adults, but with refined thresholds: ≥20 follicles per ovary (not 12) using modern ultrasound, and ethnic-specific mFG cut-offs for hirsutism.

For adolescents, diagnosis requires both oligo/anovulation AND hyperandrogenism (not just 2 of 3), because polycystic ovaries on ultrasound are common in normal adolescents.

Anti-Müllerian hormone (AMH) is not yet recommended as a diagnostic substitute for ultrasound (evidence insufficient).

**Assessment:**

All women with PCOS should be screened for depression and anxiety using validated tools (PHQ-9, GAD-7) — prevalence of depression is 3–4 times higher than in women without PCOS.

All women should have a 2-hour oral glucose tolerance test (OGTT) at diagnosis, then every 1–3 years depending on risk factors. Fasting glucose alone misses ~30% of impaired glucose tolerance.

Cardiovascular risk assessment (blood pressure, lipids) should be done at diagnosis and periodically thereafter.

**Lifestyle interventions:**

Weight loss of 5–10% of body weight improves menstrual regularity, ovulation, hirsutism, and metabolic markers. This is the most consistent finding across studies.

No specific diet (low-carb, low-GI, Mediterranean, etc.) is superior to any other for weight loss or metabolic outcomes when calories are matched. Adherence matters more than composition.

Exercise alone (without weight loss) improves metabolic health but has minimal effect on menstrual regularity or hirsutism.

Combined lifestyle (diet + exercise + behavioural support) is more effective than either alone.

**Pharmacological treatments:**

**Metformin:** Improves ovulation rates (odds ratio ~1.5–2.0 vs placebo) and metabolic markers, but effect on hirsutism is modest. First-line for metabolic management, second-line for fertility (after letrozole/clomiphene).

**Oral contraceptives:** First-line for menstrual regulation and hirsutism/acne management. All formulations appear similarly effective; no evidence that one progestin is superior.

**Anti-androgens (spironolactone, finasteride):** Effective for hirsutism (30–60% reduction in mFG score over 6–12 months), but require adequate contraception due to teratogenicity.

**Letrozole vs clomiphene for ovulation induction:** Letrozole is superior for live birth rate (odds ratio ~1.4–1.6) and has lower risk of multiple pregnancy. Letrozole is now first-line.

**Metformin + clomiphene:** More effective than clomiphene alone for ovulation and pregnancy in women with BMI >30 kg/m².

**Fertility:**

Letrozole is first-line for ovulation induction (live birth rate ~27% vs ~19% for clomiphene over up to 6 cycles).

Gonadotropins are second-line (after letrozole ± clomiphene), with higher pregnancy rates but also higher multiple pregnancy rates and cost.

In vitro fertilisation (IVF) is third-line, with similar success rates to women without PCOS but higher risk of ovarian hyperstimulation syndrome (OHSS).

**Psychological:**

Screening for depression and anxiety is recommended at diagnosis and periodically thereafter.

Cognitive behavioural therapy (CBT) and peer support groups improve quality of life and psychological distress, but evidence is limited to small studies.

**Lifestyle (weight loss):**

A 5–10% weight loss (e.g., 5–10 kg for a 100 kg woman) leads to:

- 50–80% improvement in menstrual regularity (many women resume ovulation)

- 20–30% reduction in hirsutism scores (mFG) over 6–12 months

- 10–20% improvement in fasting insulin and HOMA-IR

- These effects are roughly equivalent to what you'd get from starting metformin, but without the side effects

**Letrozole vs clomiphene for fertility:**

Live birth rate per cycle: ~10–12% with letrozole vs ~7–8% with clomiphene

Over 6 cycles: ~27% cumulative live birth with letrozole vs ~19% with clomiphene

Multiple pregnancy rate: ~3% with letrozole vs ~7% with clomiphene

This means for every 12–13 women treated with letrozole instead of clomiphene, one additional live birth occurs

**Metformin:**

Ovulation rate: ~30–50% of women ovulate on metformin alone (vs ~15–25% on placebo)

Hirsutism reduction: ~10–20% reduction in mFG score over 6–12 months (modest)

Weight loss: ~2–4 kg average (modest, not a weight loss drug)

These effects are meaningful but not transformative for most women

**Oral contraceptives for hirsutism:**

~30–50% reduction in mFG score over 6–12 months

Effect is comparable to anti-androgens but takes longer to see full benefit

For context: a woman with an mFG score of 12 (moderate hirsutism) might drop to 6–8 after 1 year

**What the authors acknowledge:**

Overall evidence quality is low to moderate — this is the single biggest limitation

Regional health system variation limits generalisability

Need for more research on diagnostic thresholds in different ethnic groups

Limited data on long-term outcomes (cardiovascular, cancer) in PCOS

Lack of high-quality studies on optimal dietary composition

Underrepresentation of adolescents and perimenopausal women in research

**What a critical reader would note:**

Many recommendations are based on expert consensus (59 of 166) rather than high-quality evidence — these are opinions, not facts

The guideline does not address individual variability in treatment response — what works for the average woman may not work for you

Lifestyle recommendations are generic ("eat healthy, exercise") without specific, testable protocols

The guideline does not provide effect sizes for many recommendations, making it hard to know what "meaningful improvement" looks like

Potential conflicts of interest: many guideline panel members had ties to pharmaceutical companies (though these were declared and managed)

The guideline is now 6+ years old (2018) — some recommendations may be outdated, especially regarding newer medications (e.g., GLP-1 agonists for weight management in PCOS)

No discussion of complementary/alternative treatments (inositol, berberine, spearmint tea, etc.) despite widespread use by women with PCOS

The guideline focuses on clinical outcomes (pregnancy, menses, hirsutism) but does not address patient-reported outcomes like fatigue, brain fog, or body image

For someone running their own n=1 experiment:

### What to test (specific intervention and dose)

**Lifestyle (most evidence-based):**

**Test:** A structured weight loss program targeting 5–10% of body weight over 3–6 months

**Dose:** 500–750 kcal/day deficit, 150+ minutes/week of moderate-to-vigorous exercise, behavioural support (tracking, coaching)

**Alternative test:** A low-glycaemic-index diet vs your usual diet (keep calories matched) — test whether glycaemic control improves independent of weight loss

**Metformin:**

**Test:** Metformin 500–2000 mg/day (start low, titrate up over 4–6 weeks to minimise GI side effects)

**Duration:** Minimum 3 months to see effects on ovulation, 6 months for hirsutism

**Note:** Requires prescription; side effects (nausea, diarrhoea) are common in ~20–30% of users

**Letrozole (for fertility):**

**Test:** Letrozole 2.5–7.5 mg/day on days 3–7 of cycle

**Duration:** Up to 6 cycles

**Note:** Requires prescription and monitoring; not for self-experimentation without medical supervision

**Inositol (not in guideline but commonly used):**

**Test:** Myo-inositol 4 g/day + D-chiro-inositol 400 mg/day (40:1 ratio)

**Duration:** Minimum 3 months

**Note:** Some evidence for improved ovulation and metabolic markers, but quality is lower than for metformin

### Minimum meaningful duration

**Menstrual regularity:** 3–6 months (need at least 3 cycles to assess pattern)

**Hirsutism:** 6–12 months (hair growth cycles are slow; you won't see changes before 3 months)

**Weight loss:** 3–6 months (5–10% loss is realistic)

**Metabolic markers (glucose, insulin, lipids):** 3 months (fasting blood tests)

**Fertility (ovulation induction):** 3–6 cycles (each cycle is ~28 days)

**Psychological outcomes (mood, quality of life):** 4–8 weeks for initial changes, 3 months for stable improvement

### What to measure (specific metrics)

**Primary outcomes (choose 1–2 that matter most to you):**

**Menstrual regularity:** Track cycle length (days between first day of bleeding), number of menses per 90 days. A positive result: