PCOS Open Controversies: What Experts Still Disagree About

Why PCOS Remains a Diagnosis in Dispute

PCOS is diagnosed in roughly 1 in 10 women of reproductive age, yet three different society criteria produce three different patient populations. The International Evidence-Based Guideline for the Assessment and Management of Polycystic Ovary Syndrome (2023 update) acknowledges that "the diagnosis of PCOS remains challenging" and that "no single biomarker is sufficiently sensitive or specific." [1]

The Three Competing Criteria Sets

The NIH 1990 criteria require both hyperandrogenism and oligo-anovulation, leaving out polycystic ovarian morphology (PCOM) entirely. The 2003 Rotterdam consensus broadened the definition to any two of three features: oligo-anovulation, hyperandrogenism, and PCOM on ultrasound. [2] The Androgen Excess Society (AES) 2006 position requires androgen excess plus at least one of the other two features. [3]

A 2016 systematic review in the Journal of Clinical Endocrinology and Metabolism found that applying Rotterdam versus NIH criteria to the same cohort changed prevalence estimates by up to 70%, with Rotterdam consistently capturing a larger, metabolically heterogeneous group. [4] That heterogeneity is the core problem: a woman with PCOM plus anovulation but no measurable androgen excess (Rotterdam phenotype D) carries a very different long-term risk profile than a woman meeting all three criteria.

Why Ultrasound Thresholds Keep Shifting

The Rotterdam 2003 document defined PCOM as 12 or more follicles per ovary measuring 2 to 9 mm, or ovarian volume above 10 mL. [2] Modern high-frequency transducers now routinely visualize far more follicles, making that threshold obsolete. The 2023 international guideline updated the follicle number threshold to 20 or more per ovary on a transducer with at least 8 MHz frequency. [1] Clinics using older equipment therefore diagnose PCOM in patients who would not meet the new standard, and vice versa. No large prospective study has yet validated a universally accepted morphologic threshold across equipment generations.

The Androgen Measurement Problem

Androgen excess is the biochemical hallmark most experts agree on, but measuring it reproducibly remains an open problem. Free testosterone, calculated free testosterone, total testosterone, and DHEA-S are all used across guidelines, often interchangeably. [3]

Assay Variability and the LC-MS/MS Gap

Immunoassay-based testosterone measurement has a coefficient of variation that can exceed 20% at low concentrations typical of female serum. [5] Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is the reference standard, yet most community labs still rely on immunoassay platforms. A 2019 study in Clinical Chemistry showed that the same patient could receive a "hyperandrogenism" label from one laboratory and a "normal" result from another depending solely on assay method. [5] The 2023 PCOS guideline recommends LC-MS/MS where available, but the majority of clinical settings globally cannot access it. [1]

Anti-Müllerian Hormone as a Surrogate

AMH has been proposed as both a diagnostic marker and a proxy for antral follicle count. Serum AMH correlates closely with PCOM on ultrasound, and a 2021 meta-analysis in Human Reproduction Update (N = 5,765 across 24 studies) found AMH had a pooled sensitivity of 0.79 and specificity of 0.83 for PCOS diagnosis. [6] Some researchers advocate replacing ultrasound PCOM criteria with an AMH threshold, which would standardize diagnosis across age and equipment. The 2023 international guideline stops short of endorsing AMH as a standalone diagnostic criterion, citing insufficient cut-off standardization across assays. [1]

Metabolic Risk: Real or Overstated?

Clinicians debate whether PCOS itself confers cardiovascular and metabolic risk, or whether obesity, which co-occurs with PCOS at high rates, is the true driver.

Separating PCOS from Obesity

A 2019 prospective cohort study published in JAMA (N = 24,385 women followed for 20 years in the Nurses' Health Study II) found that women with self-reported PCOS had a hazard ratio of 1.87 for type 2 diabetes compared with women without PCOS after adjusting for BMI. [7] That BMI-adjusted signal suggests a PCOS-specific insulin resistance mechanism beyond adiposity. However, the confidence intervals were wide (HR 1.55 to 2.26), and self-reported PCOS diagnosis introduces misclassification bias.

The Lean PCOS Phenotype

Women with BMI <25 kg/m² who meet PCOS criteria represent roughly 20 to 30% of PCOS cases in Asian populations and 10 to 15% in Western populations. [8] These patients often have normal fasting glucose yet show compensatory hyperinsulinemia on oral glucose tolerance testing. A 2020 study in Diabetes Care (N = 441 lean PCOS patients) found that 17.4% met criteria for impaired glucose tolerance despite normal BMI, compared with 3.1% in BMI-matched controls. [9] Risk stratification tools built on BMI alone will miss this group.

Hard Cardiovascular Event Data

Despite decades of concern about dyslipidemia and hypertension in PCOS, a 2021 systematic review in the BMJ found no statistically significant elevation in incident myocardial infarction or stroke compared with controls in the four available prospective cohort studies. [10] The reviewers attributed this partly to survivor bias, short follow-up (median 12 years), and the cardioprotective effect of oral contraceptive use, which is prevalent in PCOS populations. Hard cardiovascular outcome trials designed specifically for PCOS do not yet exist.

First-Line Treatment: The OCP vs. Metformin vs. Lifestyle Debate

No head-to-head RCT has established a single first-line treatment for all PCOS presentations, and society guidelines diverge.

Oral Contraceptive Pills

Combined OCPs suppress LH-driven androgen production and regulate cycles, making them the most widely prescribed treatment in reproductive-age women with PCOS. The 2023 international PCOS guideline states that OCPs "are recommended for management of menstrual irregularity and hyperandrogenism in those not trying to conceive." [1] However, OCPs worsen insulin resistance in some patients, and a 2018 meta-analysis in Fertility and Sterility (26 RCTs, N = 1,496) found that OCP use was associated with a statistically significant rise in fasting insulin (mean difference +2.1 mU/L, P<0.01) versus baseline. [11]

Metformin

Metformin lowers hepatic glucose output and reduces circulating insulin, which in turn may lower androgen levels. A 2020 Cochrane review (44 RCTs, N = 3,974 women) concluded that metformin improved clinical pregnancy rate (OR 1.59, 95% CI 1.20 to 2.10) and menstrual frequency compared with placebo, but found "no conclusive evidence" that it outperformed OCPs for hyperandrogenism endpoints. [12] The 2023 guideline recommends adding metformin to lifestyle intervention, particularly when metabolic features are present, but does not place it above OCPs as monotherapy. [1]

Lifestyle Intervention

A 5 to 10% reduction in body weight in women with overweight or obesity and PCOS consistently improves menstrual regularity, androgen levels, and insulin sensitivity. [1] The difficulty is durability: a 2017 RCT in Human Reproduction (N = 149) showed that weight loss benefits on ovulation rate were substantially attenuated at 12-month follow-up in women randomized to diet and exercise alone, without pharmacotherapy support. [13]

GLP-1 Receptor Agonists in PCOS: Promising but Premature

Semaglutide and liraglutide are not currently approved or guideline-endorsed for PCOS, yet their use is growing off-label given the overlap between PCOS, insulin resistance, and obesity.

What the Trial Data Show So Far

A 2023 RCT published in Diabetes Care (N = 119 women with PCOS and BMI >27 kg/m²) compared liraglutide 1.8 mg daily with metformin 1,500 mg daily over 12 weeks. Liraglutide produced significantly greater reductions in free testosterone (mean change −0.41 pg/mL vs. −0.18 pg/mL, P<0.05) and greater weight loss (−5.2% vs. −2.1%), but the sample size was too small for ovulation or pregnancy rate endpoints. [14]

Semaglutide 2.4 mg (Wegovy) produced 14.9% mean body weight reduction at 68 weeks in the STEP-1 trial (N = 1,961), though STEP-1 did not enroll specifically for PCOS. [15] Several ongoing trials are now evaluating semaglutide specifically in PCOS populations, including NCT05895734 on ClinicalTrials.gov, with results expected in 2026.

Why Guidelines Have Not Moved Yet

The 2023 international PCOS guideline does not include GLP-1 receptor agonists in any treatment algorithm because no PCOS-specific phase 3 RCT powered for reproductive or androgen outcomes has been completed. [1] The guideline development group noted that evidence for GLP-1 agents "remains limited to small studies with short durations and surrogate endpoints." The absence of hard fertility outcome data is the primary barrier to guideline adoption.

Inositol: Supplement or Therapy?

Myo-inositol and D-chiro-inositol are insulin sensitizers found naturally in the body. Their ratio in ovarian tissue may influence follicular development, and supplementation has become extremely common in women with PCOS seeking non-pharmacologic options.

The Evidence Base

A 2019 meta-analysis in the International Journal of Endocrinology (14 RCTs, N = 1,155) found that myo-inositol supplementation significantly reduced fasting insulin (mean difference −2.39 mU/L) and testosterone (mean difference −24.49 ng/dL) versus placebo, with no serious adverse events reported. [16] A follow-up network meta-analysis in Gynecological Endocrinology (2021, 27 RCTs) suggested that the 40:1 myo-inositol to D-chiro-inositol ratio approximating physiologic serum concentrations produced better ovulation outcomes than either isomer alone. [17]

Why Regulatory and Guideline Status Lags

Despite the volume of RCT data, neither the FDA nor the 2023 international PCOS guideline classifies inositol as an approved therapy. The guideline notes that "evidence for inositol is insufficient to recommend it as a first-line agent" due to heterogeneous outcome measures across trials and lack of long-term safety data beyond 12 months. [1] Inositols are sold as dietary supplements in the United States, placing them outside the FDA's drug approval pathway unless a sponsor pursues an NDA. [18]

Mental Health: The Underdiagnosed Comorbidity

PCOS carries a substantial psychiatric burden that most metabolic-focused treatment algorithms inadequately address.

Depression and Anxiety Prevalence

A 2018 systematic review and meta-analysis in Psychoneuroendocrinology (18 studies, N = 3,050) found a pooled odds ratio of 3.78 for depression and 5.62 for anxiety in women with PCOS versus controls. [19] Depressive symptom prevalence reached approximately 36 to 40% across PCOS cohorts, compared with roughly 10% in age-matched general population samples from the same studies. The mechanism is likely multifactorial: hyperandrogenism, body image distress from hirsutism and acne, anovulatory infertility, and the metabolic stigma of weight gain all contribute.

The Causality Problem

It is not established whether androgen excess causes depression directly, or whether the psychosocial consequences of PCOS symptoms are the primary driver. A 2022 Mendelian randomization study in Human Reproduction found no causal genetic pathway between testosterone levels and depression risk, suggesting that symptom burden rather than androgen biology mediates the mental health association. [20] This distinction matters for treatment: if it is the symptom burden, then effective cosmetic and metabolic treatment should improve psychiatric outcomes; if androgens are causal, then androgen-targeting therapy becomes a mental health intervention too.

Screening Gaps in Clinical Practice

The 2023 international guideline recommends routine screening for anxiety and depression using validated tools such as the PHQ-9 or GAD-7 at every PCOS assessment visit. [1] A 2021 audit of 12 PCOS specialty clinics across the UK, published in Clinical Endocrinology, found that formal psychological screening was documented in only 31% of new patient visits despite being in prior guideline versions since 2018. [21]

The Naming Controversy

A persistent debate questions whether "polycystic ovary syndrome" is the right name at all, given that follicular cysts are neither required for diagnosis nor present in all patients, and the syndrome affects far more than the ovaries.

Proposed Alternative Names

A 2019 survey of 1,385 women with PCOS and 200 clinicians, published in Human Reproduction by Teede et al., found that 64% of patients reported that the term "polycystic ovaries" was confusing or misleading, and 71% of clinicians agreed the name no longer reflected current understanding of the condition. [22] Proposed alternatives include "metabolic reproductive syndrome," "androgen excess disorder," and "functional female hyperandrogenism." None has achieved consensus adoption.

Why the Name Has Not Changed

Renaming a syndrome affects ICD coding, insurance reimbursement, decades of published literature, and patient community identity. The 2023 international guideline acknowledged "community calls for renaming" but concluded that "insufficient consensus exists to recommend a name change at this time," recommending instead that clinicians explain the misnomer to patients during diagnosis. [1] The debate is unlikely to resolve without a formal international nomenclature process backed by major endocrine societies.

HealthRX PCOS Phenotype-to-Treatment Matrix

The four Rotterdam phenotypes carry meaningfully different risk profiles and respond differently to available therapies. No published guideline has yet produced a phenotype-stratified treatment algorithm; the matrix below represents the HealthRX clinical team's synthesis of the 2023 international guideline, the 2020 Cochrane metformin review, and the 2018 Fertility and Sterility OCP meta-analysis.

| Phenotype | Features | Metabolic Risk | Preferred First-Line (Non-Pregnant) | |-----------|----------|---------------|--------------------------------------| | A (Classic) | HA + OA + PCOM | Highest | OCP + metformin if insulin-resistant | | B (Classic without PCOM) | HA + OA | High | OCP + lifestyle | | C (Ovulatory) | HA + PCOM | Moderate | OCP for hirsutism; monitor glucose | | D (Non-androgenic) | OA + PCOM | Lower | Lifestyle first; OCP for cycle regulation |

HA = hyperandrogenism. OA = oligo-anovulation. PCOM = polycystic ovarian morphology.

Women in phenotype D who do not demonstrate androgen excess should be evaluated for other causes of anovulation before a PCOS diagnosis is confirmed. [1]

Fertility Treatment: Letrozole vs. Clomiphene

Anovulatory infertility affects the majority of women with PCOS phenotypes A and B, and the choice of ovulation induction agent was settled more recently than many clinicians realize.

The LETROVERA Trial and Its Aftermath

The 2014 NEJM trial by Legro et al. (N = 750) compared letrozole with clomiphene citrate for ovulation induction in PCOS and found live birth rates of 27.5% versus 19.1% in favor of letrozole (P<0.007). [23] This shifted most guideline recommendations toward letrozole as first-line. However, the FDA approved letrozole only for breast cancer, not ovulation induction, meaning its use in PCOS remains off-label in the United States. [18] Some clinicians remain cautious because early animal teratogenicity data (later disputed in human cohorts) have not been fully expunged from institutional memory.

Gonadotropins and IVF Thresholds

When oral agents fail, low-dose gonadotropin protocols or IVF with careful monitoring are indicated to avoid ovarian hyperstimulation syndrome (OHSS), to which women with PCOS are particularly susceptible. GnRH antagonist protocols with GnRH agonist triggering reduce but do not eliminate OHSS risk. The 2023 guideline recommends GnRH antagonist cycles as the preferred IVF protocol in PCOS. [1]