The best PCOS blood test panel is not one lab but a timed set: testosterone with SHBG, DHEAS, TSH, prolactin, glucose or HbA1c, and morning 17-hydroxyprogesterone. Most baseline hormones work best on cycle days 2-5, while progesterone is checked about 7 days after ovulation.
- Cycle timing matters: most baseline PCOS hormone labs are best drawn at 7-10 a.m. on cycle days 2-5; progesterone belongs about 7 days after ovulation.
- Total testosterone in adult premenopausal women is often roughly 15-70 ng/dL; values above 150-200 ng/dL are unusual for routine PCOS and need urgent review.
- SHBG below about 30 nmol/L often increases free androgen exposure even when total testosterone stays within range.
- Prolactin is commonly under 25 ng/mL in nonpregnant women; 25-50 ng/mL usually deserves a repeat fasting morning sample after rest.
- TSH is typically around 0.4-4.0 mIU/L in adults; abnormal thyroid function can mimic cycle irregularity and hair changes blamed on PCOS.
- 17-hydroxyprogesterone above 200 ng/dL in an early-morning follicular sample raises concern for nonclassic congenital adrenal hyperplasia and often leads to ACTH stimulation testing.
- HbA1c of 5.7-6.4% indicates prediabetes, but a normal HbA1c does not rule out insulin resistance in PCOS.
- Progesterone above 3 ng/mL supports recent ovulation; repeated values below 1 ng/mL suggest persistent anovulation.
The short list: which blood tests actually help in suspected PCOS
The most useful PCOS blood test panel includes total testosterone, free testosterone or SHBG, DHEAS, TSH, prolactin, HbA1c or glucose, and a morning 17-hydroxyprogesterone. Most are best drawn between 7 and 10 a.m., ideally on cycle days 2-5 if you bleed at all, while progesterone is checked about 7 days after ovulation. No single lab confirms PCOS; blood work documents hyperandrogenism and rules out thyroid, prolactin, adrenal, and insulin-related lookalikes. If you already have results, Kantesti AI can interpret them alongside cycle timing and symptoms.
PCOS is diagnosed when 2 of 3 Rotterdam features are present after other causes are excluded: oligo-anovulation, clinical or biochemical hyperandrogenism, or polycystic ovarian morphology. The 2023 international guideline led by Teede and colleagues still uses that framework, which is why labs sit beside symptoms and ultrasound rather than replacing them; our women's hormone guide gives the broader cycle context.
In our analysis of more than 2 million uploaded panels across 127+ countries, the commonest miss is an incomplete workup: testosterone ordered alone, with no SHBG, TSH, prolactin, or 17-hydroxyprogesterone. When Kantesti's neural network sees that pattern, it flags the report as potentially under-interpreted instead of pretending the diagnosis is settled.
I still remember a 24-year-old with acne, chin hair, and 60-day cycles whose total testosterone was 41 ng/dL—technically normal at her lab. Her SHBG was 17 nmol/L and calculated free testosterone was clearly high, which changed the conversation from 'your hormones are fine' to 'this really may be PCOS.'
As of March 31, 2026, the practical takeaway is simple: order enough labs to exclude thyroid disease, hyperprolactinemia, and adrenal causes on the first pass. Readers who want the background on our clinical team can see About Us.
Best cycle timing for each PCOS hormone test
Most PCOS hormone tests are most informative on cycle days 2-5 because LH, FSH, estradiol, and 17-hydroxyprogesterone are easiest to compare at that baseline point. Progesterone should be measured roughly 7 days before the next period, not automatically on 'day 21' unless you cycle every 28 days.
Cycle timing matters because LH, FSH, estradiol, and 17-hydroxyprogesterone swing across the month. Early-follicular testing—usually days 2-5—gives the cleanest baseline, and if you are also checking glucose, insulin, or lipids, our fasting rules are worth following closely.
Prolactin is the test I repeat most often. A morning sample after 20 minutes seated quietly, avoiding hard exercise, nipple stimulation, and sex for about 24 hours, is far more reliable than a rushed afternoon draw; if a report uses abbreviations you do not recognize, our lab abbreviation guide can help.
No periods at all? You usually do not need to wait for a spontaneous bleed. Most exclusion labs can be drawn any morning with the date documented, though progesterone is rarely helpful unless we are specifically asking whether ovulation happened.
Birth control changes the picture more than many patients realize. Combined oral contraceptives suppress LH and ovarian androgen production while raising SHBG, so a biochemical workup is often best done after 6-12 weeks off treatment if pregnancy risk and symptoms make that reasonable.
If you are postpartum or breastfeeding
Breastfeeding can keep prolactin elevated for months, and that alone can delay ovulation. In that setting, I interpret any prolactin value in context and often defer definitive PCOS labeling until lactation is winding down.
Androgen labs that matter most: testosterone, SHBG, DHEAS, androstenedione
The androgen tests that change decisions most are total testosterone, SHBG or calculated free testosterone, DHEAS, and sometimes androstenedione. A total testosterone above about 150-200 ng/dL in a premenopausal woman is unusual for routine PCOS and deserves urgent endocrine review.
The best first androgen test is total testosterone measured by LC-MS/MS, because standard immunoassays are noisy in the female range. Azziz and colleagues pointed this out years ago in JCEM, and the problem still shows up daily; many labs quote a reference interval around 15-70 ng/dL, but method-specific ranges differ and some European labs report in nmol/L.
Low SHBG is often the missing piece. When SHBG falls below about 30 nmol/L, free androgen exposure rises even if total testosterone looks 'normal,' and that is why our SHBG deep dive matters so much in suspected PCOS.
A DHEAS level is useful because it samples adrenal androgen production. Mild elevation can occur in PCOS, but values above roughly 700-800 µg/dL make me think harder about an adrenal source, and androstenedione can catch biochemical hyperandrogenism when testosterone misses it.
What worries me most is tempo. If facial hair, deepening voice, or muscle change appears over 6-12 months and total testosterone lands at 160 ng/dL, I stop calling it 'probably PCOS' and start excluding tumor or ovarian hyperthecosis.
Why calculated free testosterone can mislead
Calculated free testosterone is only as good as the total testosterone and SHBG assays feeding it. Kantesti AI cross-checks unit conversions because a result entered as ng/mL instead of ng/dL can create a spectacularly false alarm.
LH, FSH, estradiol, progesterone, and AMH: useful context, not a standalone diagnosis
LH and FSH can support the story, but they do not diagnose PCOS. The old LH:FSH ratio above 2:1 is neither sensitive nor specific; I still see it overused.
An LH:FSH ratio above 2:1 can occur in PCOS, but plenty of confirmed cases have a ratio near 1:1, and many non-PCOS cycles drift higher as well. That is why I treat the ratio as context, not a criterion.
Early-follicular FSH is often around 3-10 IU/L and estradiol about 25-75 pg/mL, though lab ranges vary. If estradiol is already above 80-100 pg/mL on day 3, it can suppress FSH enough to make ovarian reserve look better than it really is.
Midluteal progesterone above 3 ng/mL supports that ovulation happened recently. Values below 1 ng/mL on repeated untimed checks strongly suggest anovulation, and Kantesti's neural network explains this timing problem in the same way our clinical standards team does.
Then there is AMH. Many women with PCOS have AMH above 4-5 ng/mL, but assay variability is real and guideline groups still do not recommend it as a universal standalone diagnostic test; if you want help parsing units and reference comments, see our guide to reading blood results.
The 'day 21 progesterone' myth
A day-21 progesterone only makes sense in a 28-day cycle. In a 40-day cycle, the more useful draw is closer to day 33, because the goal is to sample the mid-luteal window, not a calendar number.
How thyroid disease and prolactin disorders can mimic PCOS
Every suspected PCOS workup should include TSH, free T4 when TSH is abnormal, and prolactin. A TSH outside roughly 0.4-4.0 mIU/L or a prolactin above 25 ng/mL can explain irregular cycles without PCOS.
Thyroid disease can mimic PCOS because both can disrupt ovulation. Our low TSH guide explains the hyperthyroid pattern. The mirror-image hypothyroid pattern is covered in our high TSH guide.
A normal prolactin for nonpregnant women is commonly under 25 ng/mL, although assay ranges differ. I, Thomas Klein, MD, usually repeat any value between 25 and 50 ng/mL as a fasting morning sample after rest, because venipuncture stress alone can bump it.
Persistent prolactin above 50 ng/mL deserves a medication review first. Antipsychotics, metoclopramide, some antidepressants, and even chest-wall irritation can elevate it, while prolactin above 100 ng/mL makes a pituitary adenoma much more likely than PCOS.
One subtle trap is macroprolactin—biologically less active prolactin that can make the number look scary without causing classic symptoms. Our physicians on the medical review team ask for macroprolactin when the story and the number do not match.
When prolactin and TSH are both off
Combined TSH elevation and prolactin elevation is common in untreated hypothyroidism because TRH can stimulate both pathways. Treating the thyroid first can normalize both without a pituitary scan.
Insulin, glucose, and metabolic labs that change long-term risk
Metabolic labs matter because PCOS raises lifetime risk of prediabetes, type 2 diabetes, dyslipidemia, and fatty liver even when periods are the first complaint. An HbA1c of 5.7-6.4% indicates prediabetes, yet many young women with PCOS still have a normal A1c and an abnormal 2-hour glucose tolerance test.
A normal fasting glucose does not clear the metabolic side of PCOS. Fasting glucose 70-99 mg/dL and HbA1c below 5.7% are reassuring, but a 2-hour OGTT of 140-199 mg/dL still shows impaired glucose tolerance and is common in younger women with PCOS.
Fasting insulin is useful for pattern recognition, not diagnosis. Values above roughly 15 µIU/mL or a research HOMA-IR above 2.5 often line up with insulin resistance, but there is no globally standardized cutoff and I never diagnose off insulin alone.
Lipids tell you what the ovaries cannot. Triglycerides above 150 mg/dL and HDL below 50 mg/dL often travel with insulin resistance. Our review of HbA1c cutoffs explains glycemic thresholds. A separate guide to lipid panel interpretation helps with triglycerides and HDL. Our ALT guide covers the liver side.
Liver enzymes matter because metabolic dysfunction-associated steatotic liver disease clusters with PCOS. In women, a persistent ALT above about 25 U/L can be a meaningful early clue even when the lab's printed upper limit is 35 U/L, and if you already have those labs, our AI blood test analyzer can pull the pattern together in about 60 seconds.
Lean PCOS still needs metabolic testing
Lean PCOS is real. I have seen marathon runners with BMI under 22 kg/m² and completely normal fasting glucose who still failed a 75 g OGTT at 2 hours.
17-hydroxyprogesterone, DHEAS, and the red flags that suggest something other than PCOS
A morning follicular-phase 17-hydroxyprogesterone is the single most useful blood test for distinguishing PCOS from nonclassic congenital adrenal hyperplasia. A 17-OHP below 200 ng/dL makes NCAH less likely, while higher values usually trigger ACTH stimulation testing.
The screening sample for 17-hydroxyprogesterone should be drawn between about 7 and 9 a.m. in the follicular phase. A value below 200 ng/dL usually argues against nonclassic CAH, 200-800 ng/dL is gray-zone territory, and higher numbers usually prompt ACTH stimulation.
Androgen-secreting tumors usually look louder than PCOS. Total testosterone above 150-200 ng/dL, DHEAS above 700-800 µg/dL, or rapid virilization over a few months should move imaging and endocrine referral to the front of the list.
Cushing syndrome is not a routine screen in every acne-and-irregular-period case, and that saves a lot of unnecessary testing. I order cortisol testing when there is easy bruising, wide purple striae, proximal muscle weakness, or new hypertension—not just because the cycles are irregular.
Not every hair or fatigue complaint in suspected PCOS is hormonal. Low ferritin under 30 ng/mL can worsen shedding, and our fatigue lab list is useful when the symptom profile is broad. The iron side is covered in our ferritin ranges. Low vitamin D under 20 ng/mL can also muddy the picture, and our vitamin D chart is a helpful reference. If symptoms are driving the question more than the diagnosis name, our test selector can narrow what to ask for.
Why a random afternoon 17-OHP can fool you
A random afternoon 17-OHP can be misleading because adrenal steroid output follows a circadian rhythm. I trust a 7-9 a.m. follicular sample much more than an untimed result tucked into a general chemistry panel.
How to recognize real PCOS patterns versus common lookalikes
Real PCOS patterns usually show mild to moderate androgen excess with normal or near-normal thyroid and prolactin tests, not dramatic hormone spikes. When I review a panel with total testosterone 58 ng/dL, SHBG 19 nmol/L, A1c 5.9%, and prolactin 14 ng/mL, that feels like classic insulin-driven PCOS.
Pattern one is classic insulin-driven PCOS: cycles every 45-70 days, total testosterone 58 ng/dL, SHBG 19 nmol/L, A1c 5.9%, triglycerides 198 mg/dL, and prolactin 14 ng/mL. That combination tells me the androgen story is real and the metabolic piece needs attention on day one.
Pattern two is the thyroid masquerade. A woman with fatigue, cold intolerance, dry skin, cycles every 50 days, TSH 8.6 mIU/L, and prolactin 34 ng/mL can look like PCOS at first glance, but normal androgens usually point us back to thyroid-first treatment.
Pattern three is the urgent adrenal or ovarian signal. If hirsutism worsens quickly and DHEAS is 840 µg/dL or testosterone 188 ng/dL, I worry less about labels and more about speed; that is the patient I would not leave in a routine follow-up queue.
Pattern four is lean anovulatory PCOS, which many websites barely mention. I, Thomas Klein, MD, see women with BMI 21 kg/m², A1c 5.2%, normal total testosterone, SHBG 26 nmol/L, AMH 6.8 ng/mL, and repeated progesterone below 1 ng/mL—not dramatic, but very real.
Kantesti AI is particularly useful when results arrive across several PDFs from different labs. You can see that kind of multi-report reasoning in our real patient cases. If you already have results, our AI-powered blood test interpretation can process them quickly.
A practical PCOS lab checklist by situation: regular cycles, no periods, birth control, postpartum
The right PCOS blood test order depends on cycle status, medications, and age. If you are having periods, order most baseline hormones on days 2-5; if you have gone more than 90 days without bleeding, most exclusion labs can be drawn any morning and the date simply documented.
If cycles are present, my standard morning panel on days 2-5 is total testosterone, SHBG or free testosterone, DHEAS, TSH, prolactin, 17-OHP, and often LH, FSH, estradiol, glucose, A1c, lipids, and ALT. If pregnancy is possible, add a serum β-hCG before you over-interpret anything.
If you have had no bleeding for more than 90 days, do not wait months for a 'perfect' cycle day. Draw the exclusion labs now, document amenorrhea, and use progesterone later only if the question becomes 'did ovulation happen' rather than 'what is causing the irregularity?'
Combined hormonal contraception changes testosterone and SHBG enough to blur the picture. When it is safe, I prefer androgen testing after 6-12 weeks off the pill; a levonorgestrel IUD usually distorts androgen labs less, but it can still confuse cycle tracking.
Postpartum and perimenopausal cases deserve extra skepticism. Lactation can keep prolactin high for months, while perimenopause can push FSH above 10-15 IU/L and make a long-standing PCOS pattern look different; our annual lab checklist helps with baseline screening.
Bottom line: the right order set is personal, not one-size-fits-all. If you want a quick second read on a real report, try Try Free AI Blood Test Analysis. If you are curious how the model reasons through assay differences, our technology guide shows the logic.
What I would not over-order at the start
I rarely start with a huge fertility panel unless the history points there. A focused first pass is usually enough: androgen testing, thyroid, prolactin, 17-OHP, and metabolic screening answer most of the clinically important questions.
Research publications and medical review
Methodology matters. A testosterone result measured by LC-MS/MS and tied to cycle day carries more clinical weight than an untimed immunoassay number floating alone.
Interpretation quality depends on timing, assay method, and clinical review. Kantesti's medical content is reviewed by physicians, and our broader approach to AI blood test interpretation is built around context rather than isolated flags. Across 2M+ users, 75+ languages, and 127+ countries, that context-first approach matters more than ever.
Recommended citation: Kantesti AI Research Team. (2026). Serum Proteins Guide: Globulins, Albumin & A/G Ratio Blood Test. Zenodo. https://doi.org/10.5281/zenodo.18316300. Discovery copies are indexed on ResearchGate. A parallel listing also appears on Academia.edu.
Recommended citation: Kantesti AI Research Team. (2026). C3 C4 Complement Blood Test & ANA Titer Guide. Zenodo. https://doi.org/10.5281/zenodo.18353989. Discovery copies are indexed on ResearchGate. A parallel listing also appears on Academia.edu.
As Thomas Klein, MD, I would make one plea: never let a single untimed testosterone result settle the PCOS question. The pattern—timing, prolactin, thyroid, adrenal screening, and metabolic risk—is what makes the diagnosis safe.
Frequently Asked Questions
Can PCOS be diagnosed by blood test alone?
No. PCOS is usually diagnosed when a patient has at least 2 of 3 findings—irregular or absent ovulation, clinical or biochemical hyperandrogenism, or polycystic ovarian morphology—after other causes are excluded. Blood tests are still central because they document androgen excess and rule out thyroid disease, prolactin disorders, and adrenal conditions such as nonclassic CAH. In practice, a good PCOS blood test narrows the diagnosis safely, but it does not replace history and imaging.
Which day of the cycle should I do blood tests for PCOS?
Most baseline hormone labs for suspected PCOS are best done on cycle days 2-5, ideally between 7 and 10 a.m. That timing is especially helpful for LH, FSH, estradiol, testosterone, and 17-hydroxyprogesterone. Progesterone is different: it is most useful about 7 days before the next period, not automatically on day 21. If you are not having periods, most exclusion labs can usually be drawn on any morning with the date documented.
Do I need to fast for a PCOS blood test?
Fasting is most useful when the panel includes glucose, insulin, and lipids, and many clinicians prefer at least 8-12 hours with water only. Hormone tests such as testosterone, TSH, and prolactin do not always require strict fasting, but a morning fasting sample reduces noise and makes the results easier to compare. I am especially picky about fasting and resting when a mildly high prolactin needs repeating. Water is fine; coffee can interfere with the metabolic side of the panel.
Can I test for PCOS while on birth control?
You can do some labs while on birth control, but biochemical androgen testing is often distorted. Combined hormonal contraception tends to lower ovarian androgen production, suppress LH, and raise SHBG, which can make testosterone look more normal than it really is. If it is clinically safe, many endocrinologists prefer androgen testing after 6-12 weeks off the pill. Thyroid, HbA1c, glucose, and many general labs remain interpretable while contraception is being used.
What testosterone level suggests something other than ordinary PCOS?
A total testosterone above about 150-200 ng/dL is more concerning for a non-PCOS cause, especially if symptoms are progressing quickly. A DHEAS above roughly 700-800 µg/dL also pushes clinicians to think about an adrenal source rather than routine PCOS. The number is not the whole story, though; rapid onset voice change, clitoromegaly, or dramatic hirsutism over 6-12 months matters just as much. In that setting, imaging and urgent endocrine review usually move up the list.
Is a normal HbA1c enough to rule out insulin problems in PCOS?
No. An HbA1c below 5.7% can look reassuring and still miss insulin resistance or even impaired glucose tolerance in younger women with PCOS. I have seen patients with A1c 5.2% and a clearly abnormal 2-hour OGTT in the 140-199 mg/dL range. That is why fasting glucose, lipids, and sometimes a formal glucose tolerance test give a fuller picture than A1c alone. A normal A1c is good news, but it is not the whole metabolic story.
Should prolactin be repeated if it is only a little high?
Usually yes. A prolactin level in the 25-50 ng/mL range is often repeated as a morning sample after 20 minutes of rest, because stress, exercise, poor sleep, and the blood draw itself can transiently raise it. If it stays elevated, medication review and sometimes macroprolactin testing are the next steps. Persistent values above 100 ng/mL are much more concerning for a pituitary adenoma than for PCOS.
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). Serum Proteins Guide: Globulins, Albumin & A/G Ratio Blood Test. Kantesti AI Medical Research.
Klein, T., Mitchell, S., & Weber, H. (2026). C3 C4 Complement Blood Test & ANA Titer Guide. Kantesti AI Medical Research.
⚕️ 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.