PCOS Open Controversies: What Experts Still Disagree About

The core disagreement, stated plainly

PCOS is not one disease with one test. It is a diagnosis reached by combining features, irregular ovulation, biochemical or clinical signs of excess androgen, and/or a particular ovarian appearance on ultrasound, using one of three overlapping but distinct rule sets developed by different professional bodies since 1990. Because the criteria disagree, the population labeled "PCOS" in one clinic or one study is not the same population labeled "PCOS" in another. That single fact underlies most of the downstream controversies about risk and treatment described below: metabolic risk, mental health burden, and treatment response all vary by which features a given patient actually has, not just by whether she carries the diagnosis.

The most current international guideline (Teede et al., 2023, endorsed by more than 40 societies) uses the Rotterdam criteria as its diagnostic base but sets a higher ultrasound follicle threshold than the original 2003 definition to account for modern high-resolution imaging. Clinics still using older equipment or older thresholds can therefore diagnose polycystic ovarian morphology in patients who would not meet the updated standard, and the reverse is also true. No large prospective study has yet validated a single morphologic threshold that holds across ultrasound equipment generations, which is why "how many follicles is too many" remains an active methodological argument rather than a settled fact.

Why three different diagnostic definitions still exist

NIH 1990 requires both hyperandrogenism and irregular or absent ovulation. It does not consider ovarian appearance on ultrasound at all.

Rotterdam 2003 broadened the definition to any two of three features: irregular ovulation, hyperandrogenism, or polycystic ovarian morphology on ultrasound. This is the framework used by most guidelines today, including the 2023 international guideline, but it also means a woman can be diagnosed with PCOS without any measurable androgen excess (so-called phenotype D, ovulatory dysfunction plus ovarian morphology, but no hyperandrogenism).

Androgen Excess Society (2006) requires androgen excess as a mandatory feature, plus at least one of the other two criteria. This narrower definition excludes phenotype D entirely.

Studies applying these three rule sets to the same group of women have found that prevalence and the mix of included patients change substantially depending on which criteria are used, with Rotterdam consistently capturing the largest and most metabolically varied group. The exact percentage difference reported by any one study is a number worth confirming in the primary literature rather than repeating as fixed fact, because it depends heavily on the population studied.

This matters clinically because a woman who meets Rotterdam criteria through ovulatory dysfunction and ultrasound findings alone, without hyperandrogenism, likely has a different long-term metabolic profile than a woman who meets all three features. Guidelines increasingly recommend that clinicians document which specific features a patient has, not just whether she "has PCOS."

The androgen measurement problem

Hyperandrogenism is one of the few features most diagnostic frameworks agree matters, yet measuring it reliably is harder than it sounds. Total testosterone, calculated free testosterone, and DHEA-S are all used, sometimes interchangeably, across different guidelines and labs.

Standard immunoassay testosterone testing, the method most community labs use, is known to lose precision at the low concentrations typical of female serum. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) is the more accurate reference method, and the 2023 international guideline recommends it where available. In practice, most clinical settings worldwide still rely on immunoassay platforms, which means the same patient could plausibly receive a "hyperandrogenism" result in one lab and a "normal" result in another depending on which assay was used. This is an acknowledged limitation in the guideline literature, not a fringe concern.

Anti-Müllerian hormone (AMH) has been proposed as a surrogate for ultrasound-based ovarian morphology, since it correlates with antral follicle count and does not depend on operator skill or transducer frequency. Some researchers have argued for replacing the ultrasound criterion with an AMH cutoff to standardize diagnosis. The 2023 international guideline does not go that far, citing insufficient standardization of AMH assay cutoffs across manufacturers and labs. This is an area of active methodological work rather than a resolved question.

Is the metabolic risk about PCOS, or about obesity that co-occurs with it?

This is one of the most consequential unresolved questions, because it changes how aggressively a clinician should screen for diabetes and cardiovascular risk in a given patient.

What looks reasonably well established: insulin resistance is common across PCOS phenotypes, including in some women who are not overweight, and several large observational cohorts have reported an elevated risk of type 2 diabetes in women with PCOS even after statistical adjustment for body mass index. That adjusted signal is the basis for treating PCOS as an independent metabolic risk factor rather than simply a marker for obesity-related risk. The exact magnitude reported varies across studies and should be checked against the specific cohort in question rather than quoted as a universal hazard ratio.

The lean PCOS phenotype complicates simple BMI-based risk stratification. A meaningful minority of women who meet PCOS criteria have a body mass index in the normal range, and several smaller studies have found that these women can still show abnormal glucose handling on formal testing (oral glucose tolerance testing) despite normal fasting glucose and normal weight. This suggests that screening tools relying on BMI alone will miss metabolic risk in a subset of PCOS patients, though the precise proportion affected differs across the populations studied (reported estimates for the lean phenotype's share of all PCOS cases also vary by ethnicity and by which diagnostic criteria were applied).

What is not established: whether PCOS itself, independent of weight and insulin resistance, meaningfully raises the risk of hard cardiovascular events such as heart attack or stroke. Available prospective cohort studies have generally been underpowered for these outcomes, with limited follow-up duration and confounding from oral contraceptive use (common in PCOS populations and independently associated with cardiovascular risk in some contexts). No PCOS-specific randomized cardiovascular outcome trial exists. Clinicians should treat the elevated lipid and insulin markers seen in PCOS as a reason for monitoring and lifestyle counseling, not as proof of a specific increased heart attack or stroke rate that has been directly measured in PCOS populations.

No single first-line treatment fits every patient

This is a genuine and durable disagreement, not a gap that will close with one more trial, because the "right" first-line treatment depends heavily on the patient's goals (cycle regulation vs. androgen symptoms vs. fertility vs. metabolic risk) and on which phenotype she has.

Combined oral contraceptives remain the most widely prescribed option for menstrual irregularity and androgen-related symptoms (acne, hirsutism) in women not trying to conceive, and the 2023 international guideline supports this use. The tradeoff: some data suggest OCP use can modestly worsen insulin sensitivity markers in some women, which is a relevant consideration in patients who already have metabolic risk factors, though this needs to be weighed against the OCP's clear benefit for cycle regulation and androgen symptoms.

Metformin improves menstrual frequency and appears to modestly improve pregnancy-related outcomes according to a Cochrane systematic review, but the same review found no conclusive evidence that it outperforms oral contraceptives specifically for androgen-related symptoms. The 2023 guideline positions metformin as an addition to lifestyle intervention, particularly when insulin resistance is present, rather than as a universal first-line monotherapy.

Lifestyle intervention (modest weight loss of roughly 5 to 10 percent in women with overweight or obesity) consistently improves menstrual regularity, androgen levels, and insulin sensitivity in trial data, and is guideline-endorsed as a foundational intervention. The practical limitation is durability: at least one randomized trial found that ovulation-related benefits from diet and exercise alone were substantially reduced by twelve months without added pharmacologic support, which is a meaningful caveat for patients told to "just lose weight" without a maintenance plan.

GLP-1 receptor agonists: real signal, premature guideline adoption

Semaglutide and liraglutide (GLP-1 receptor agonists, approved by the FDA for type 2 diabetes and, at higher doses, for chronic weight management) are increasingly used off-label in women with PCOS and obesity, given the overlap between insulin resistance, weight, and PCOS symptoms. This use is off-label for PCOS specifically: no GLP-1 agent carries an FDA indication for PCOS, and as of this writing the 2023 international PCOS guideline does not include GLP-1 agents in its treatment algorithm.

Small randomized trials comparing liraglutide with metformin in women with PCOS and overweight have reported greater reductions in free testosterone and greater weight loss with liraglutide over short follow-up periods (weeks, not years), but these trials have been too small to assess fertility or ovulation outcomes. Semaglutide's weight-loss magnitude is well documented in large general-obesity trials (not PCOS-specific), and separate trials evaluating semaglutide specifically in PCOS populations are registered and ongoing on ClinicalTrials.gov. The guideline development group's stated reason for not yet including GLP-1 agents is the absence of a PCOS-specific trial powered for reproductive or androgen outcomes, not doubt about the drugs' general metabolic effects. A clinician recommending a GLP-1 agent for PCOS today is making an off-label, extrapolated decision based on obesity and diabetes trial data plus small PCOS-specific pilot data, and that distinction should be explained to the patient.

Inositol: widely used, not guideline-endorsed

Myo-inositol and D-chiro-inositol are naturally occurring insulin-sensitizing compounds sold in the United States as dietary supplements, which places them outside the FDA drug approval pathway. Several meta-analyses of small randomized trials have reported improvements in fasting insulin and testosterone levels with myo-inositol supplementation compared with placebo, and some work has suggested a particular ratio of myo-inositol to D-chiro-inositol may produce better outcomes than either compound alone, though trial designs and outcome measures vary enough across this literature that a single confident effect size is not appropriate to quote.

Despite the volume of trial data, the 2023 international guideline does not classify inositol as a recommended first-line therapy, citing heterogeneous outcome measures and a lack of long-term (beyond about a year) safety data. This is a case where patient demand and small-trial enthusiasm have outpaced what guideline bodies consider sufficient evidence for a formal recommendation. That does not mean inositol is unsafe; it means the evidence base has not yet cleared the bar guideline committees use for a positive recommendation.

Mental health: a documented burden with an unresolved cause

Depression and anxiety are reported substantially more often in women with PCOS than in age-matched women without the condition, based on pooled data across multiple studies. The 2023 international guideline responded to this by recommending routine screening for anxiety and depression, using validated tools such as the PHQ-9 or GAD-7, at every PCOS assessment visit.

What is not established is why the association exists. One line of genetic epidemiology work has not found strong evidence that testosterone levels causally drive depression risk, which would suggest that the psychosocial burden of PCOS symptoms (hirsutism, acne, weight change, infertility, and the stigma attached to all of these) may be doing more of the work than androgen biology itself. If that is correct, treatments that reduce visible symptoms and support fertility goals may do more for mental health than androgen-lowering therapy alone, but this causal question has not been resolved and should not be presented to patients as settled.

Despite the guideline recommendation for routine screening, audits of specialty PCOS clinics have found that formal psychological screening is documented in only a minority of new patient visits, suggesting a real gap between guideline recommendation and everyday practice. Patients who feel their mental health symptoms are being dismissed in the context of a PCOS diagnosis have grounds to ask directly for depression and anxiety screening as part of their workup.

Should PCOS even be called "polycystic ovary syndrome"?

This is a real and ongoing debate, not a fringe complaint. The name implies ovarian cysts are central to the condition, but the follicular structures seen on ultrasound in PCOS are immature follicles, not pathologic cysts, and a meaningful subset of women who meet diagnostic criteria never show this ultrasound finding at all. Patient and clinician surveys have found that a majority of both groups consider the current name confusing or outdated relative to current understanding of the condition. Proposed alternatives include names built around metabolic-reproductive dysfunction or androgen excess, but none has achieved consensus adoption.

The 2023 international guideline acknowledged calls for renaming but concluded that consensus was insufficient to recommend a change, in part because renaming would affect insurance coding, decades of published literature, and patient community identity built around the existing term. Clinicians are instead encouraged to explain the naming mismatch to patients at diagnosis, since the word "cyst" can cause unnecessary alarm.

Fertility treatment: letrozole has largely replaced clomiphene, but with a regulatory wrinkle

For women with PCOS and anovulatory infertility, a major multicenter randomized trial (Legro et al., published in the New England Journal of Medicine in 2014) found higher live birth rates with letrozole than with clomiphene citrate for ovulation induction, a result that shifted most fertility guidelines, including the 2023 international PCOS guideline, toward recommending letrozole as first-line. Readers and clinicians should confirm the exact reported live birth rates against the original trial publication rather than treating a secondhand summary as definitive.

The regulatory wrinkle: the FDA has approved letrozole for breast cancer treatment, not for ovulation induction, so its use for PCOS-related infertility in the United States is off-label even though it is the guideline-preferred agent. This is a common and accepted situation in reproductive endocrinology, but patients are not always told that a guideline-preferred fertility drug is being prescribed off-label, and they should be, since off-label status has implications for informed consent and sometimes for insurance coverage.

When oral agents fail, gonadotropin protocols or IVF are used, with particular caution around ovarian hyperstimulation syndrome (OHSS), a risk to which women with PCOS are more susceptible than the general fertility population. GnRH antagonist-based protocols are generally preferred in PCOS IVF cycles specifically to reduce this risk.

What is established, what is plausible, and what is not established

Established: PCOS is diagnosed using overlapping but distinct criteria sets that produce different patient populations; insulin resistance is common across most PCOS phenotypes; combined oral contraceptives, metformin, and lifestyle intervention each have a defined and guideline-supported role depending on the patient's goals; letrozole is now guideline-preferred over clomiphene for ovulation induction in PCOS, though it is used off-label; depression and anxiety occur more often in women with PCOS than in the general population.

Plausible but not settled: that PCOS confers cardiovascular risk independent of body weight and insulin resistance at the level of hard events (heart attack, stroke); that a specific myo-inositol to D-chiro-inositol ratio meaningfully outperforms either compound alone; that GLP-1 receptor agonists will eventually earn a PCOS-specific guideline indication once adequately powered trials complete; that symptom burden rather than androgen biology is the primary driver of the PCOS-depression association.

Not established: a single universally accepted ultrasound follicle threshold for polycystic ovarian morphology that holds across equipment generations; a validated AMH cutoff to replace ultrasound criteria; a single first-line pharmacologic treatment that outperforms the others for every PCOS phenotype; any PCOS-specific hard cardiovascular outcome trial.

Readers should treat this article, and any PCOS source, with appropriate skepticism toward precise percentages and hazard ratios repeated without a direct citation trail. Where this draft describes a study finding without a verifiable specific source attached, that finding should be confirmed against the primary publication before being used for an individual clinical decision.

Decision framework: what your phenotype and goals should actually change

Guidelines do not currently publish a formal phenotype-specific treatment algorithm. The framework below is the HealthRX.com clinical team's synthesis of the diagnostic and treatment evidence discussed above, intended to help a patient and clinician structure the conversation, not to replace an individualized treatment plan.

Step 1: Confirm which specific features you have, not just the label. Ask your clinician which of the three Rotterdam features (ovulatory dysfunction, biochemical or clinical hyperandrogenism, polycystic ovarian morphology on ultrasound) you actually meet, and whether the ultrasound was performed with a transducer meeting current-generation frequency standards.

Step 2: State your primary goal before choosing a treatment. "I want regular cycles and less acne," "I want to get pregnant this year," and "I'm worried about my long-term diabetes risk" lead to different first-line choices even within the same phenotype. A treatment that is appropriate for cycle regulation (an oral contraceptive) is not the treatment you would choose if pregnancy is the near-term goal, and a metabolic-focused plan (metformin, lifestyle change, or an off-label GLP-1 discussion) does not address hirsutism or acne on its own.

Step 3: Know when a finding should trigger a second opinion or additional workup rather than a standard PCOS treatment plan. Rapidly progressive hirsutism, very high testosterone levels, or signs of virilization (voice deepening, significant hair loss pattern change) warrant evaluation for an androgen-secreting tumor rather than routine PCOS management. Amenorrhea alone, without other PCOS features, warrants ruling out pregnancy and other causes before assuming PCOS. These are exceptions where the standard PCOS pathway should not be assumed.

Step 4: Ask directly about off-label status. If letrozole, a GLP-1 agonist, or high-dose inositol combinations are recommended, ask specifically whether the use is FDA-approved for your indication, guideline-recommended off-label, or investigational. This single question resolves most of the confusion patients report about why a "standard" PCOS treatment is not listed on the drug's official label.

Step 5: Ask when screening for depression, anxiety, glucose intolerance, and lipid abnormalities was last done, since guideline-recommended screening for these is inconsistently applied in practice even when the diagnosis is well established.

Evidence sources and verification note

This draft describes findings from named trials, cohort studies, and systematic reviews discussed in the PCOS literature, along with recommendations from the 2023 international evidence-based PCOS guideline (Teede et al.). Specific numeric results (hazard ratios, odds ratios, percentage reductions) are described in general terms here because the precise source-to-claim linkages inherited from earlier drafts of this article could not be independently verified and are flagged for confirmation during medical review rather than presented as exact figures. Regulatory and approval status statements (for example, letrozole's FDA-approved indication, and inositol's status as a dietary supplement rather than an approved drug) should be confirmed against current FDA labeling at Drugs@FDA before publication, since drug labels and approved indications can change over time. Ongoing GLP-1 trials in PCOS populations can be searched directly at ClinicalTrials.gov.

This article does not provide individualized diagnostic or dosing guidance. A clinician should confirm which diagnostic criteria apply to an individual patient, order and interpret androgen and metabolic testing directly, and tailor any treatment discussion to that patient's fertility goals, symptom profile, and medical history. Sudden or severe symptoms such as significant unexplained bleeding, signs of virilization, or symptoms of ovarian hyperstimulation during fertility treatment warrant prompt medical evaluation rather than waiting for a routine follow-up.