Diagnosis and Treatment of Polycystic Ovary Syndrome: An Endocrine Society Clinical Practice Guideline

This is not a single experiment but a systematic review and clinical practice guideline. The Task Force evaluated the entire body of published evidence on PCOS diagnosis and treatment to answer several clinical questions:

**Diagnostic criteria:** Which combination of symptoms, blood tests, and ultrasound findings best identifies PCOS while excluding other conditions?

**Associated health risks:** What comorbidities (diabetes, endometrial cancer, depression, sleep apnea, cardiovascular disease) should women with PCOS be screened for?

**Treatment efficacy:** How do hormonal contraceptives, metformin, clomiphene, lifestyle interventions, and other drugs compare for managing different PCOS symptoms?

The guideline compared multiple diagnostic approaches (NIH criteria requiring both androgen excess AND ovulatory dysfunction vs. Rotterdam criteria requiring two of three features vs. Androgen Excess Society criteria requiring hyperandrogenism plus one other feature). For treatment, they compared hormonal contraceptives vs. no treatment, metformin vs. placebo, clomiphene vs. metformin for infertility, and lifestyle intervention vs. standard care.

Outcome measures included: menstrual regularity, hirsutism scores (modified Ferriman-Gallwey scale), acne severity, pregnancy rates, ovulation rates, fasting glucose and insulin, hemoglobin A1c, lipid profiles, blood pressure, and body mass index.

The guideline synthesises evidence from dozens of studies encompassing thousands of women with PCOS across multiple populations. Specific studies cited include:

Women aged 18–45 diagnosed with PCOS by various criteria

Adolescent girls aged 12–18 with hyperandrogenism and oligomenorrhea

Perimenopausal and postmenopausal women with documented history of PCOS

Overweight and obese women (BMI >25 kg/m²) with PCOS

Women seeking fertility treatment for anovulatory infertility

The guideline explicitly notes that most studies were conducted in tertiary care referral centres, meaning the populations studied may have more severe phenotypes than women with PCOS in the general community. Studies were predominantly from North America and Europe, with limited data from Asian, African, and Latin American populations.

The guideline used standardised clinical and laboratory measures:

**Hyperandrogenism:** Modified Ferriman-Gallwey score (0–36 scale, ≥8 indicates hirsutism in most populations), serum total testosterone, free testosterone, androstenedione, DHEAS

**Ovulatory dysfunction:** Menstrual history (cycles <21 days or >35 days indicate oligo-ovulation), midluteal serum progesterone (>3 ng/mL confirms ovulation)

**Polycystic ovary morphology:** Transvaginal ultrasound showing ≥12 follicles 2–9 mm in diameter and/or ovarian volume >10 mL in either ovary

**Metabolic status:** Oral glucose tolerance test (75g glucose, fasting and 2-hour glucose), hemoglobin A1c, fasting insulin, HOMA-IR (insulin resistance index)

**Cardiovascular risk:** Blood pressure, lipid panel (total cholesterol, LDL, HDL, triglycerides), waist circumference

**Treatment response:** Menstrual bleeding patterns (number of cycles per year), hirsutism scores, acne lesion counts, ovulation rates (by progesterone or ultrasound), pregnancy rates

**Study design:** This is a clinical practice guideline developed using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system. The Task Force conducted two systematic reviews of the literature to summarise supporting evidence. Recommendations are graded as strong (1) or weak (2) based on the balance of benefits vs. harms, and the quality of evidence is rated as high (⊕⊕⊕⊕), moderate (⊕⊕⊕○), low (⊕⊕○○), or very low (⊕○○○).

**Consensus process:** One in-person group meeting, multiple conference calls, and email communications. Drafts were reviewed by committees and members of The Endocrine Society and the European Society of Endocrinology.

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

**Can prove:** The guideline provides expert consensus on best clinical practice based on systematic evaluation of available evidence. Strong recommendations (Grade 1) are supported by moderate-to-high quality evidence showing clear benefit over harm.

**Cannot prove:** This is not a randomised trial. The guideline cannot establish causality for any individual treatment recommendation with the same certainty as a well-powered RCT. Weak recommendations (Grade 2) reflect uncertainty about the balance of benefits and harms, often due to low-quality evidence or closely balanced trade-offs. The guideline also cannot account for individual patient variability—it provides population-level recommendations, not personalised treatment plans.

**Major methodological weaknesses acknowledged by the authors:**

Most PCOS treatment studies are short-term (3–6 months) with small sample sizes

Few studies compare different treatments head-to-head

Diagnostic criteria vary across studies, making meta-analysis difficult

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

Minimal evidence in adolescents and postmenopausal women

Lack of standardised androgen assays across laboratories

Industry funding of many pharmaceutical trials

**Diagnosis:**

The Rotterdam criteria (two of three: androgen excess, ovulatory dysfunction, polycystic ovaries) are recommended for diagnosis in adults (weak recommendation, moderate-quality evidence)

In adolescents, diagnosis requires both clinical/biochemical hyperandrogenism AND persistent oligomenorrhea—polycystic ovaries alone are insufficient because they can occur during normal puberty (weak recommendation, very low-quality evidence)

In perimenopausal/menopausal women, diagnosis is based on well-documented history of oligomenorrhea and hyperandrogenism during reproductive years (weak recommendation, very low-quality evidence)

**Associated morbidity:**

Women with PCOS have a 2- to 4-fold increased risk of impaired glucose tolerance and type 2 diabetes compared to women without PCOS

Risk of gestational diabetes is increased (odds ratio approximately 2.5–3.0)

Prevalence of depression is 2- to 3-fold higher than in women without PCOS

Obstructive sleep apnea prevalence is 5- to 10-fold higher in women with PCOS compared to weight-matched controls

Endometrial cancer risk is increased approximately 2.7-fold (odds ratio 2.7, 95% CI 1.3–5.6)

**Treatment:**

**Hormonal contraceptives (first-line for menstrual abnormalities and hirsutism/acne):** Reduce free testosterone by 40–60%, improve hirsutism scores by 30–50% over 6–12 months, regulate menstrual cycles in 80–90% of women (strong recommendation, moderate-quality evidence)

**Clomiphene (first-line for infertility):** Ovulation induction success rate 60–80% per cycle, pregnancy rate 15–20% per cycle, live birth rate approximately 20–25% over 6 cycles (strong recommendation, moderate-quality evidence)

**Metformin:** Improves menstrual regularity (50–60% of women resume regular cycles), reduces fasting insulin by 20–30%, improves glucose tolerance. However, no significant benefit for hirsutism or acne compared to placebo. For infertility, metformin alone is inferior to clomiphene (ovulation rate 30–40% vs. 60–80%) (strong recommendation for metabolic benefits, weak recommendation against use for hirsutism/acne/infertility alone)

**Lifestyle intervention (diet + exercise):** 5–10% weight loss improves menstrual regularity in 30–50% of overweight/obese women, reduces free testosterone by 15–25%, improves insulin sensitivity. However, the effect on PCOS status independent of weight loss is uncertain (weak recommendation, low-quality evidence)

**Thiazolidinediones (e.g., pioglitazone):** Improve ovulation and metabolic parameters but associated with weight gain, fluid retention, and potential cardiovascular risk—unfavourable risk-benefit ratio overall (recommendation against routine use)

**Statins:** Limited evidence; may improve lipid profiles and reduce testosterone, but long-term safety and efficacy in PCOS are unknown (recommendation for further study)

To put these numbers in plain English:

**Hormonal contraceptives** reduce free testosterone by roughly half—equivalent to dropping from the top 5% of testosterone levels in women to the middle of the normal range. Hirsutism improvement is noticeable but gradual: about a 1-point drop on the Ferriman-Gallwey scale every 3–4 months, meaning a woman with a score of 15 might reach a score of 10 after 12–18 months of treatment.

**Clomiphene** gives a woman with PCOS roughly a 1 in 5 chance of pregnancy each cycle—compared to essentially zero without treatment if she is not ovulating. Over 6 cycles, about 1 in 4 women will have a live birth.

**Metformin** restores regular menstrual cycles in about half of women who take it. For a woman having 2–4 periods per year, this means going to 9–12 periods per year. The effect on insulin resistance is roughly equivalent to losing 5–7% of body weight through diet alone.

**Weight loss of 5–10%** (e.g., 10–20 pounds for a 200-pound woman) restores ovulation in about 1 in 3 overweight women with PCOS. This is roughly the same effect as taking metformin, but without the gastrointestinal side effects.

**What the authors acknowledge:**

The quality of evidence for many recommendations is low or very low, particularly for adolescents and postmenopausal women

Most treatment studies are short-term (3–6 months), so long-term safety and efficacy data are lacking

Diagnostic criteria have changed over time, making it difficult to compare studies from different eras

Androgen assays are not standardised across laboratories, making it hard to define universal cut-off values

The guideline does not address individual patient preferences, which may influence treatment choices

**What a critical reader would note:**

**Industry funding:** Many of the studies cited were funded by pharmaceutical companies that manufacture hormonal contraceptives, metformin, or fertility drugs. The guideline does not disclose the funding sources of individual studies.

**Population limits:** Most evidence comes from women seen in tertiary care centres, who likely have more severe PCOS than women in the general population. Findings may not apply to women with mild PCOS who are not seeking medical care.

**Lack of blinding:** Many lifestyle intervention studies cannot be blinded, introducing potential bias. Women who lose weight may also change other behaviours that affect outcomes.

**Short duration for chronic condition:** PCOS is a lifelong condition, yet most treatment studies last only 3–6 months. We don't know if benefits persist for years or if side effects emerge with long-term use.

**No head-to-head comparisons of different hormonal contraceptives:** The guideline recommends "hormonal contraceptives" as a class, but different pills, patches, and rings may have different effects on androgens and metabolic parameters.

**Limited data on combination therapy:** Few studies have examined whether combining treatments (e.g., metformin + clomiphene, or hormonal contraceptives + lifestyle) is better than either alone.

**No cost-effectiveness analysis:** The guideline does not consider the relative costs of different treatment approaches, which may be important for patients making real-world decisions.

For someone running their own n=1 experiment—whether you have a PCOS diagnosis or suspect you might—here is how to apply this guideline to test what works for you:

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

**Option A: Lifestyle intervention**

Test: Hypocaloric diet (500–750 calorie deficit per day) combined with 150–300 minutes of moderate-intensity exercise per week

Target: 5–10% weight loss if overweight/obese (BMI >25)

Duration: Minimum 12 weeks to see changes in menstrual regularity; 6 months for changes in hirsutism

**Option B: Metformin**

Test: Metformin extended-release 1500–2000 mg daily (start at 500 mg and increase gradually over 4 weeks to minimise gastrointestinal side effects)

Duration: Minimum 3 months for menstrual effects; 6 months for metabolic effects

**Option C: Inositol supplementation** (not covered in this guideline but commonly used)

Test: Myo-inositol 4 g daily plus D-chiro-inositol in 40:1 ratio

Duration: Minimum 3 months

### Minimum meaningful duration

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

**Hirsutism/acne:** 6 months minimum (hair growth cycles are 3–6 months; acne improvement takes 2–3 cycles)

**Fertility/ovulation:** 3–6 cycles of treatment

**Metabolic parameters (insulin, glucose):** 3 months

**Weight loss:** 12 weeks minimum to see 5% weight loss

### What to measure (specific metrics)

**Primary outcomes (track weekly):**

Menstrual cycle length (days between first day of bleeding)

Number of bleeding days

Subjective ovulation signs (cervical mucus changes, mittelschmerz, breast tenderness)

Basal body temperature (BBT) daily to confirm ovulation

**Secondary outcomes (track monthly):**

Hirsutism: Modified Ferriman-Gallwey self-score (rate 9 body areas 0–4, total 0–36)

Acne: Count inflammatory lesions (papules, pustules, nodules) on face, chest, back

Weight and waist circumference (weekly, same time of day)

Fasting glucose and insulin (baseline and at 3 months)

Free testosterone (baseline and at 3–6 months, same lab, same time of day, early follicular phase if menstruating)

**For fertility-focused experiments:**

Ovulation predictor kits (LH surge) daily from cycle day 10

Midluteal progesterone (cycle day 21 if 28-day cycle, or 7 days after LH surge)

Pregnancy test if period is >5 days late

### Key confounds to control for

**Timing of blood tests:** Testosterone and insulin vary with menstrual cycle phase. Always test in early follicular phase (cycle days 2–5) and at the same time of day (morning, fasting).

**Weight changes:** Even 2–3 kg weight change can affect PCOS symptoms. Track weight weekly and account for it in your analysis.

**Stress:** Cortisol increases androgens. Track perceived stress (e.g., 1–10 scale daily) as a potential confound.

**Sleep:** Poor sleep increases insulin resistance. Track sleep duration and quality.

**Other medications:** Birth control pills, anti-androgens (spironolactone), antidepressants, and supplements can all affect PCOS symptoms. Do not change multiple variables at once.

**Seasonal variation:** Some women experience worse PCOS symptoms in winter. Run experiments for at least 3 months to account for seasonal effects.

### What a positive result would look like

**For menstrual regularity:**

Cycle length stabilises to 25–35 days consistently for 3 consecutive cycles

BBT shows biphasic pattern (temperature rise of 0.3–0.5°F after ovulation sustained for 12–16 days)

Midluteal progesterone >3 ng/mL

**For hirsutism:**

Ferriman-Gallwey score decreases by ≥3 points over 6 months (this is clinically meaningful)

Reduced frequency of shaving/waxing (e.g., from every 2 days to every 4–5 days)

Hair becomes finer and lighter in colour

**For metabolic health:**

Fasting insulin decreases by ≥20% from baseline

HOMA-IR decreases by ≥25%

-