PCOS is diagnosed when a clinician finds two of three Rotterdam features—irregular or absent ovulation, androgen excess, or polycystic ovarian morphology—after excluding look-alike conditions. No single testosterone, AMH, insulin, or ultrasound result can confirm it on its own.
- Rotterdam criteria PCOS requires any 2 of 3 features: ovulatory dysfunction, androgen excess, or polycystic ovarian morphology after exclusion testing.
- One hormone test cannot diagnose PCOS because testosterone and SHBG vary by assay, cycle timing, medicines, and body weight.
- Irregular cycles are fewer than 8 periods per year or cycles longer than 35 days in adults; adolescents need different timing rules.
- Ultrasound threshold is at least 20 follicles in one ovary or ovarian volume of at least 10 mL when modern transvaginal imaging is used.
- Thyroid testing with TSH is routine because hypothyroidism can cause irregular periods, weight change, and altered SHBG.
- Prolactin above 25 ng/mL often needs a calm, repeat morning sample before it is blamed on PCOS.
- 17-hydroxyprogesterone is used to screen for non-classic congenital adrenal hyperplasia, a treatable PCOS mimic.
- Metabolic screening should include glucose assessment and lipids even when body size is small and periods are the main symptom.
How the Rotterdam framework actually establishes PCOS
PCOS diagnosis criteria use the Rotterdam framework: two of three features must be present, and other endocrine causes must be reasonably excluded. The three features are ovulatory dysfunction, clinical or biochemical hyperandrogenism, and polycystic ovarian morphology; having all three is not required.
The Rotterdam criteria were introduced in 2003, but the clinical work is more nuanced than ticking two boxes. A person with cycles every 45 to 90 days and clear facial terminal-hair growth may meet criteria without an ultrasound, while someone with regular 28-day cycles, acne alone, and a multifollicular scan may not. The 2023 international guideline led by Teede et al. retains the two-of-three approach while stressing careful exclusion of mimics.
Ovulatory dysfunction means ovulation is infrequent or unpredictable, not merely that a period feels different. In adults, cycles longer than 35 days, fewer than 8 periods annually, or gaps exceeding 90 days are clinically meaningful; a cycle diary is often more useful than a single progesterone result. Our women's hormonal health guide explains why cycle timing changes almost every reproductive hormone.
Thomas Klein, MD, has seen a common and frustrating pattern: a patient is told she has PCOS after one scan, then later discovers untreated thyroid disease or high prolactin. The diagnosis should explain the whole pattern over time, not simply attach a label to a lab flag. PCOS can coexist with another condition, so exclusion testing is not a formality.
The three features in plain language
Ovulatory dysfunction concerns cycle frequency; hyperandrogenism concerns androgen effect or level; and polycystic ovarian morphology is an imaging pattern, not proof that someone has ovarian cysts. Rotterdam criteria apply only after pregnancy, thyroid dysfunction, hyperprolactinemia, and non-classic congenital adrenal hyperplasia have been considered.
Symptoms that count—and symptoms that do not prove PCOS
Irregular periods, unwanted terminal hair growth, persistent acne, and scalp hair thinning can support PCOS, but none proves it by itself. The most diagnostically useful symptom history records cycle spacing, hair-growth distribution, medication use, and how quickly changes appeared.
Cycles that are 21 to 35 days apart are usually considered regular in adults, although regular bleeding does not guarantee ovulation every month. A mid-luteal progesterone greater than 3 ng/mL, or 9.5 nmol/L, supports recent ovulation, but the sample must be taken about 7 days before the next expected period—not automatically on “day 21.”
Clinical hyperandrogenism means terminal, coarse hair in androgen-sensitive areas rather than the fine hair found on forearms or thighs. A modified Ferriman-Gallwey score of 4 to 6 may be meaningful depending on ethnicity, although self-removal can make scoring unreliable; acne after age 25 and female-pattern scalp thinning are less specific clues. For lab context, see our guide to high DHEA in women.
Rapid progression changes the urgency. New deep voice change, marked muscle change, or quickly advancing hair growth over 3 to 6 months is not typical uncomplicated PCOS and warrants prompt assessment for an androgen-secreting source. In my experience, patients often normalize these changes because they have heard PCOS is common; that is precisely when a clinician should take a closer look.
Why one blood test cannot diagnose PCOS
No blood test diagnoses PCOS because androgen concentrations overlap substantially between PCOS and people without PCOS. Results also change with assay method, time of day, hormonal contraception, SHBG concentration, and whether ovulation occurred recently.
Total testosterone is usually the first biochemical androgen test, but liquid chromatography–tandem mass spectrometry is more accurate at the low concentrations typical in women than many direct immunoassays. A result just above a laboratory limit should therefore be repeated or confirmed before it defines hyperandrogenism. Kantesti is an AI blood test analyzer that reads testosterone, SHBG, and related markers as a pattern rather than declaring PCOS from one value.
Free testosterone can be measured by equilibrium dialysis or estimated from total testosterone, SHBG, and albumin; an isolated “free testosterone” immunoassay is often the weakest result on a panel. Low SHBG can raise calculated free testosterone even when total testosterone sits within range, particularly with insulin resistance, hypothyroidism, obesity, or liver disease. Read more about the causes of low SHBG.
Combined oral contraception suppresses ovarian androgen production and raises SHBG, so it can make biochemical testing falsely reassuring. The 2023 guideline suggests stopping the pill for at least 3 months if biochemical assessment is essential and pregnancy prevention is safely addressed; this is a shared decision, not a universal instruction. Our detailed guide covers PCOS testing after birth control.
PCOS hormone testing: what clinicians order and why
PCOS hormone testing is mainly used to document androgen excess and exclude other disorders, not to find a single diagnostic number. Most initial panels include total testosterone, SHBG or calculated free testosterone, TSH, prolactin, and 17-hydroxyprogesterone.
A practical early-morning panel often includes total testosterone, SHBG, albumin, DHEA-S, prolactin, TSH, and 17-hydroxyprogesterone. DHEA-S is produced mainly by the adrenal glands and is less cycle-dependent than testosterone, which makes it useful when the source of androgen excess is uncertain. A markedly elevated total testosterone, often above 150 to 200 ng/dL depending on method, deserves urgent specialist review rather than routine PCOS follow-up.
LH and FSH are frequently ordered, but an LH:FSH ratio is not part of Rotterdam criteria and should not be used to diagnose PCOS. LH varies across the cycle, and an elevated ratio can occur in people without PCOS; very high FSH instead raises concern for reduced ovarian function. Our LH and FSH interpretation guide helps place these results in cycle context.
Kantesti AI interprets hormone panels by checking units, reference intervals, medicines, and linked thyroid or prolactin results before it identifies questions for a clinician. Kantesti is an AI blood test interpretation platform designed to organize multi-marker laboratory patterns, not replace a diagnostic consultation. This distinction matters most when symptoms and results point in different directions.
Why thyroid disease is ruled out before diagnosing PCOS
TSH testing is routine in a possible PCOS evaluation because hypothyroidism can cause infrequent periods, weight change, infertility, and low SHBG. A thyroid result does not diagnose PCOS, but an abnormal one may explain the presentation or coexist with it.
Most laboratories use a TSH reference interval near 0.4 to 4.0 mIU/L, though intervals and pregnancy targets differ. A raised TSH with low free T4 supports overt primary hypothyroidism; a raised TSH with normal free T4 is subclinical hypothyroidism and needs individualized management. Teede et al. (2023) recommend excluding thyroid disease in all people being assessed for PCOS.
Thyroid disease can alter SHBG before it produces dramatic symptoms. Hypothyroidism often lowers SHBG and may increase calculated free testosterone, whereas hyperthyroidism tends to raise SHBG and can mask biochemical hyperandrogenism; this is one reason I do not interpret androgen values without thyroid context. See our explanation of T4, T3, and TSH.
Biotin supplements can interfere with some immunoassays, creating a pattern of falsely low TSH and falsely high free T4 or T3. Doses of 5,000 to 10,000 mcg daily are common in hair products, so tell the laboratory and clinician; many advise pausing biotin for 48 to 72 hours, although local assay guidance wins.
Prolactin: the often missed PCOS mimic
Prolactin should be checked because hyperprolactinemia can suppress ovulation and cause irregular or absent periods that resemble PCOS. A single mild elevation is common and usually requires a properly prepared repeat test before any imaging is considered.
Many labs quote an upper prolactin limit around 20 to 25 ng/mL, or about 425 to 530 mIU/L, for non-pregnant women. Sleep, pain, exercise, sex, a difficult sample collection, and medicines such as metoclopramide, antipsychotics, and some antidepressants can transiently elevate it. A quiet morning sample after 20 to 30 minutes of rest is a sensible first repeat.
Prolactin above 100 ng/mL, or approximately 2,120 mIU/L, is less likely to be explained by stress alone and commonly prompts endocrine assessment; headache or visual field symptoms make the timeline more urgent. Levels between 25 and 50 ng/mL are where clinical context does the heavy lifting. Our review of slightly elevated prolactin gives a practical retesting plan.
Macroprolactin is a large prolactin-antibody complex that can produce a high laboratory result with little biological effect. If periods are regular and there is no milk discharge, infertility, or low sex-hormone pattern, polyethylene glycol screening for macroprolactin can prevent unnecessary scans. That small laboratory detail has saved more than a few of my patients a great deal of anxiety.
Adrenal conditions to exclude: 17-hydroxyprogesterone and beyond
An early-morning 17-hydroxyprogesterone test screens for non-classic congenital adrenal hyperplasia, an inherited adrenal condition that can mimic PCOS. It is particularly relevant with early-onset hair growth, severe acne, certain family histories, or androgen excess without typical metabolic features.
For screening, 17-hydroxyprogesterone is generally collected before 8 AM and, when possible, in the early follicular phase. A basal level above 200 ng/dL, or about 6 nmol/L, commonly triggers an ACTH stimulation test, though assay-specific cutoffs vary. Legro et al. (2013) specifically recommend excluding non-classic congenital adrenal hyperplasia in PCOS assessment.
The stimulated result is more definitive: a 60-minute 17-hydroxyprogesterone concentration above 1,000 ng/dL, or 30 nmol/L, supports non-classic 21-hydroxylase deficiency in many protocols. This is not a minor distinction because treatment, genetic counselling, and pregnancy implications differ. Our guide to high 17-hydroxyprogesterone explains sample timing in more detail.
DHEA-S above roughly 700 µg/dL, or 19 µmol/L, can raise concern for a major adrenal androgen source, but no single cutoff is absolute across laboratories. Sudden virilization, severe biochemical elevation, or Cushing-like signs should move the work-up beyond PCOS. For symptoms and screening logic, review our Cushing syndrome testing guide.
What counts as a positive PCOS ultrasound diagnosis
PCOS ultrasound diagnosis requires either at least 20 follicles in one ovary or an ovarian volume of at least 10 mL, using a modern transvaginal probe with at least 8 MHz frequency. The scan identifies polycystic ovarian morphology, not “cysts” that need removal.
The 20-follicle threshold is for adults and depends on imaging quality; older equipment cannot reliably support it. If follicle counts are not reliable, ovarian volume of 10 mL or more in either ovary is the accepted alternative, provided there is no dominant follicle, corpus luteum, or cyst distorting the measurement. The 2023 international guideline advises against routine ultrasound for diagnosis in adolescents.
A scan is unnecessary when irregular cycles and hyperandrogenism are already documented, because those two features meet Rotterdam criteria after exclusions. This avoids a common delay: patients are sometimes told they cannot receive a diagnosis without imaging, despite having two clear clinical features. Kantesti is an AI-powered blood test analysis tool that can help users bring organized hormone and metabolic results to that discussion, while imaging remains a clinician-led assessment.
Polycystic morphology is common in people without PCOS, especially in adolescence and after stopping hormonal contraception. A report of “multiple follicles” does not establish the syndrome unless the clinical and laboratory picture fits. If pelvic pain is the main issue, clinicians also look for other explanations; our endometriosis testing guide covers why imaging and blood markers have limits.
Why PCOS diagnosis is different in teenagers
Teenagers require both persistent ovulatory dysfunction and hyperandrogenism for a PCOS diagnosis; ultrasound and AMH should not be used diagnostically in adolescence. Normal pubertal maturation can produce irregular cycles and multifollicular ovaries for several years after the first period.
In the first year after menarche, irregular cycles are expected and do not define PCOS. From 1 to under 3 years after menarche, cycles shorter than 21 or longer than 45 days are considered irregular; from 3 years onward, cycles shorter than 21, longer than 35 days, or fewer than 8 yearly warrant evaluation. Any single cycle longer than 90 days after the first post-menarche year needs assessment.
The guideline recommends reassessment around 8 years after menarche for adolescents with one feature but not both, sometimes described as “at risk” rather than diagnosed. This language is not evasive—it prevents a lifelong label from being applied during normal puberty. Our article on hormone tests for teenage girls discusses safe timing and interpretation.
Acne alone is common during puberty, whereas severe acne resistant to standard treatment or progressive terminal hair growth offers stronger evidence of androgen effect. In a teen with irregular periods, pregnancy testing, TSH, prolactin, and 17-hydroxyprogesterone may still be appropriate. A normal ultrasound does not settle the question either way.
AMH is useful context, not a shortcut diagnosis
Anti-Müllerian hormone may substitute for ultrasound to define polycystic ovarian morphology in adults under the 2023 guideline, but it cannot be used as a standalone PCOS test. A high AMH result must be interpreted with cycles and androgen status, and it is not recommended for adolescents.
AMH reflects the number of small growing follicles, so it often runs higher in PCOS—but it also varies with age, assay manufacturer, contraception, and ethnicity. There is no universal diagnostic AMH cutoff; one laboratory may flag 5.0 ng/mL while another uses a different assay and interval. This is one of those areas where context matters more than the number.
Do not combine AMH and ultrasound to count two Rotterdam features, because both measure the same morphology domain. A person needs at least one additional feature: ovulatory dysfunction or hyperandrogenism. For the reverse problem—low results—our low AMH guide explains what AMH can and cannot predict.
AMH is not a direct measure of egg quality, an immediate fertility verdict, or a reason to diagnose PCOS in someone with no symptoms. As of October 2, 2026, major guidance remains cautious because assay standardization is incomplete. I tell patients that AMH is a clue from a crowded room, not a headcount with perfect precision.
Metabolic tests that belong in a PCOS evaluation
Everyone with PCOS should have glycaemic status assessed at diagnosis and repeated every 1 to 3 years according to risk, regardless of BMI. A 75-g oral glucose tolerance test is the most accurate assessment for dysglycaemia in PCOS.
A fasting glucose of 100 to 125 mg/dL, or 5.6 to 6.9 mmol/L, indicates impaired fasting glucose; a 2-hour 75-g glucose result of 140 to 199 mg/dL, or 7.8 to 11.0 mmol/L, indicates impaired glucose tolerance. Diabetes is diagnosed at fasting glucose of 126 mg/dL or more, 2-hour glucose of 200 mg/dL or more, or HbA1c of 6.5% or more when confirmed appropriately.
HbA1c is convenient but can miss post-meal dysglycaemia that an oral glucose tolerance test detects. Teede et al. (2023) state that the 75-g test is the most accurate option in PCOS, while fasting glucose and HbA1c are reasonable alternatives when it cannot be done. Our fasting insulin guide explains why insulin values are not diagnostic on their own.
A fasting lipid profile should include total cholesterol, LDL-C, HDL-C, and triglycerides at diagnosis. Triglycerides of 150 mg/dL or more, 1.7 mmol/L, and HDL-C below 50 mg/dL, 1.3 mmol/L, in women are metabolic syndrome components, but they are not Rotterdam criteria. Kantesti AI can show trends across repeated metabolic panels, which is usually more informative than reacting to one non-fasting triglyceride result.
How to prepare for PCOS blood tests without distorting them
For the clearest PCOS hormone testing, collect an early-morning sample, record cycle day, and list every hormonal medicine and supplement. Fasting is useful for glucose and lipid planning but is not required for every androgen or thyroid test.
For 17-hydroxyprogesterone, before 8 AM and early in the cycle is ideal because levels rise later in the day and after ovulation. Prolactin is best repeated after a calm 20- to 30-minute rest, while strenuous exercise the evening before can produce a distracting mild elevation. A baseline blood test checklist can prevent avoidable redraws.
Do not stop prescribed contraception, thyroid medicine, steroids, or psychiatric treatment solely to obtain “natural” results without speaking to the prescriber. For androgen testing on combined hormonal contraception, a 3-month washout may be considered only if the result will change management and effective contraception is arranged. Pregnancy testing should come first when a period is late, because pregnancy is a common cause of amenorrhea.
Bring prior laboratory reports, even if they are from another country or use different units. Kantesti is an AI lab test interpretation service that can translate units and display longitudinal patterns from uploaded reports, but a clinician must still confirm a diagnosis and decide whether retesting is safe. The platform's approach is described in our AI technology guide.
How clinicians combine symptoms, labs, and imaging
Clinicians diagnose PCOS by checking whether two Rotterdam domains persist after exclusions, then separately assessing fertility goals, endometrial protection, and metabolic risk. The diagnostic label is only the start of a plan; two people meeting the same criteria may need very different care.
Consider three typical patterns. Irregular cycles plus biochemical hyperandrogenism meets criteria without ultrasound; regular cycles plus polycystic morphology does not; and hyperandrogenism plus morphology may meet criteria even if bleeding looks monthly, provided ovulation is assessed when uncertainty remains. A progesterone level drawn roughly 7 days before expected bleeding can clarify ovulation in the last scenario.
If periods occur fewer than 4 times a year, discuss endometrial protection rather than simply waiting for the next cycle. The goal is generally to avoid prolonged unopposed estrogen exposure; exact treatment depends on pregnancy plans, contraindications, and the clinician's assessment. Heavy or unpredictable bleeding is a separate symptom pathway—our guide to heavy menstrual bleeding explains urgent warning signs.
Kantesti reviews laboratory patterns with physician oversight and flags when a thyroid, prolactin, androgen, or glucose result needs clinical follow-up rather than a self-diagnosis. The methodology and guardrails are available through our medical validation information. Thomas Klein, MD, emphasizes that a good PCOS assessment leaves you with an explanation and a next appointment, not just a long list of abnormal flags.
When suspected PCOS needs urgent or specialist assessment
Rapid virilization, severe headache with vision changes, pregnancy possibility, or very high androgen results need prompt medical assessment rather than a routine PCOS work-up. Most PCOS evaluations are outpatient, but a few symptom patterns point elsewhere and should not wait.
Seek urgent care for possible pregnancy with one-sided pelvic pain, fainting, shoulder pain, or heavy bleeding; PCOS does not protect against pregnancy or ectopic pregnancy. A home pregnancy test can turn positive before a missed period, while serum hCG may detect pregnancy earlier. Our pregnancy test timing guide explains false-negative timing.
Arrange expedited endocrine or gynecology review for new voice deepening, clitoral enlargement, rapidly progressive hair growth, or total testosterone in a markedly elevated range such as above 150 to 200 ng/dL. A prolactin result above 100 ng/mL with headache or visual symptoms also warrants timely evaluation. These findings are uncommon, but overlooking them is far more consequential than repeating a routine panel.
For non-urgent uncertainty, bring a 6- to 12-month cycle record, photographs or dates documenting symptom change, medicines, and complete reports to an appointment. Kantesti's medical content is reviewed with input from our Medical Advisory Board, but personalized diagnosis remains the responsibility of your treating clinician. That boundary is protective, not bureaucratic.
Frequently Asked Questions
Can PCOS be diagnosed with one blood test?
No. PCOS cannot be diagnosed with one blood test because Rotterdam criteria require two of three domains—ovulatory dysfunction, hyperandrogenism, and polycystic ovarian morphology—after other causes are excluded. Total testosterone, free testosterone estimates, SHBG, DHEA-S, TSH, prolactin, and 17-hydroxyprogesterone each answer different clinical questions. A testosterone result just above the laboratory range should often be repeated or confirmed with a more accurate method before it is treated as biochemical hyperandrogenism.
What are the three Rotterdam criteria for PCOS?
The three Rotterdam criteria are irregular or absent ovulation, clinical or biochemical hyperandrogenism, and polycystic ovarian morphology on ultrasound or, in selected adults, an elevated AMH result. PCOS is diagnosed when two of these three features are present after pregnancy, thyroid dysfunction, hyperprolactinemia, and non-classic congenital adrenal hyperplasia have been excluded. In adults, cycles longer than 35 days or fewer than 8 periods yearly support ovulatory dysfunction. You do not need all three features or an ultrasound to meet the criteria.
What ultrasound result confirms polycystic ovarian morphology?
With modern transvaginal ultrasound, polycystic ovarian morphology is defined as at least 20 follicles in at least one ovary or an ovarian volume of 10 mL or more in at least one ovary. The 20-follicle threshold assumes a transvaginal probe with frequency of at least 8 MHz, so older imaging equipment may rely more on ovarian volume. The scan does not show dangerous “cysts”; it shows many small follicles. Ultrasound is not recommended as a diagnostic test for PCOS in adolescents because this appearance is common during normal puberty.
Which tests rule out conditions that look like PCOS?
Initial exclusion testing usually includes TSH for thyroid disease, prolactin for hyperprolactinemia, and early-morning 17-hydroxyprogesterone for non-classic congenital adrenal hyperplasia. A 17-hydroxyprogesterone concentration above about 200 ng/dL, or 6 nmol/L, often leads to ACTH stimulation testing. Pregnancy testing is necessary with missed periods, and severe or rapidly progressive androgen symptoms may require testing for adrenal disease, Cushing syndrome, or an androgen-secreting source. The exact panel depends on symptoms, medicines, age, and examination findings.
Can you have PCOS with normal testosterone?
Yes. A person can meet PCOS diagnosis criteria with normal total testosterone if they have clinical hyperandrogenism, such as significant terminal hair growth, plus ovulatory dysfunction or polycystic ovarian morphology. Total testosterone may remain normal when SHBG is low and calculated free testosterone is higher, and assay limitations can miss modest elevations. Conversely, acne alone with normal cycles and a normal androgen panel does not establish PCOS. Clinicians interpret symptoms, repeat results, and medication effects together.
Is AMH high enough to diagnose PCOS?
No. AMH is not high enough on its own to diagnose PCOS because it reflects follicle number and varies by age, laboratory method, ethnicity, and hormonal contraception. The 2023 international guideline allows AMH to define the morphology component in adults, but the person still needs ovulatory dysfunction or hyperandrogenism to meet two Rotterdam domains. There is no single universal AMH cutoff such as 5 ng/mL that applies across all laboratories. AMH should not be used for PCOS diagnosis in adolescents.
Get AI-Powered Blood Test Analysis Today
Join over 2 million users worldwide who trust Kantesti for instant, accurate lab test analysis. Upload your blood test results and receive comprehensive interpretation of 15,000+ biomarkers in seconds.
📚 Referenced Research Publications
Klein, T., Mitchell, S., & Weber, H. (2026). Multilingual AI Assisted Clinical Decision Support for Early Hantavirus Triage: Design, Engineering Validation, and Real-World Deployment Across 50,000 Interpreted Blood Test Reports. Kantesti AI Medical Research.
Klein, T., Mitchell, S., & Weber, H. (2026). A Pre-Registered, Rubric-Based Automated Technical Benchmark of the Kantesti Blood-Test Interpretation Engine on 100,000 Synthetic Test Cases. Kantesti AI Medical Research.
📖 External Medical References
⚕️ Medical Disclaimer
This article is for educational purposes only and does not constitute medical advice. Always consult a qualified healthcare provider for diagnosis and treatment decisions.
E-E-A-T Trust Signals
Experience
Physician-led clinical review of lab interpretation workflows.
Expertise
Laboratory medicine focus on how biomarkers behave in clinical context.
Authoritativeness
Written by Dr. Thomas Klein with review by Dr. Sarah Mitchell and Prof. Dr. Hans Weber.
Trustworthiness
Evidence-based interpretation with clear follow-up pathways to reduce alarm.