What lab tests are used to diagnose Polycystic Ovary Syndrome (PCOS)?

Laboratory Diagnosis of Polycystic Ovary Syndrome

Measure total testosterone (TT) and free testosterone (FT) as first-line laboratory tests using liquid chromatography-tandem mass spectrometry (LC-MS/MS), as these demonstrate the highest diagnostic accuracy for biochemical hyperandrogenism in PCOS. 1

First-Line Androgen Testing

Primary androgen markers:

• Total testosterone (TT) should be measured first, showing pooled sensitivity of 74% and specificity of 86% for PCOS diagnosis 1, 2
• Free testosterone (FT) demonstrates superior sensitivity of 89% with specificity of 83% 1
• Calculated free testosterone (cFT) should be assessed by equilibrium dialysis or ammonium sulfate precipitation, or calculated using Free Androgen Index (FAI) 1
• LC-MS/MS is the mandatory preferred assay method, showing superior specificity (92%) compared to direct immunoassays (78%) 1, 2

The 2025 International PCOS Guidelines explicitly recommend TT and FT as first-line tests because they provide the best balance of sensitivity and specificity. 1 Direct immunoassays are problematic due to low accuracy at the low serum concentrations typical in women, leading to false positives. 1

Second-Line Androgen Testing

If TT or FT are not elevated but clinical suspicion remains high:

• Androstenedione (A4) can be measured (sensitivity 75%, specificity 71%) 1
• DHEAS can be measured (sensitivity 75%, specificity 67%) 1
• Note that A4 and DHEAS have poorer specificity than TT/FT and should only be used as adjunctive tests 1
• DHEAS levels also help rule out non-classical congenital adrenal hyperplasia 2

Essential Exclusionary Laboratory Tests

To exclude other causes of hyperandrogenism and menstrual irregularity:

• Thyroid-stimulating hormone (TSH) to rule out thyroid disease 2, 3
• Prolactin level (morning resting serum) to exclude hyperprolactinemia 2, 3
• 17-hydroxyprogesterone if congenital adrenal hyperplasia is suspected 4

Metabolic Screening Tests

All women with suspected PCOS require metabolic assessment:

• Two-hour oral glucose tolerance test with 75g glucose load to screen for glucose intolerance and type 2 diabetes 2, 3
• Fasting lipid panel including total cholesterol, LDL, HDL, and triglycerides 2, 3
• These tests are critical because PCOS carries increased risk for metabolic syndrome, type 2 diabetes, and cardiovascular disease 2, 5

Supplementary Hormonal Tests

Additional tests that may support diagnosis but are not required:

• LH and FSH measured between days 3-6 of menstrual cycle, with LH/FSH ratio >2:1 suggesting PCOS (present in 55% of cases) 2, 6
• Mid-luteal progesterone (<6 nmol/L indicates anovulation) 2
• Anti-Müllerian Hormone (AMH) shows high diagnostic accuracy (threshold ≥35 pmol/L has 92% sensitivity and 97% specificity), but should not yet be used as a single diagnostic test 2, 3

Critical Pitfalls to Avoid

Common laboratory errors:

• Never rely on direct immunoassays for testosterone measurement—they have significantly lower specificity (78%) compared to LC-MS/MS (92%) and produce false positives 1, 2
• Do not use ultrasound findings alone for diagnosis—polycystic ovarian morphology is present in up to one-third of normal reproductive-aged women 3
• Screen for androgen-secreting tumors if testosterone is very high (>150-200 ng/dL) or symptoms have rapid onset 2, 4
• Consider Cushing's syndrome screening if patient has buffalo hump, moon facies, hypertension, abdominal striae, or easy bruising 2, 3

Diagnostic Algorithm

Step 1: Measure TT and FT using LC-MS/MS 1

Step 2: If LC-MS/MS unavailable, calculate FAI (Total testosterone/SHBG × 100) 1

Step 3: If TT/FT normal but clinical suspicion high, measure A4 and DHEAS 1

Step 4: Measure TSH and prolactin to exclude mimicking conditions 2, 3

Step 5: Perform metabolic screening with oral glucose tolerance test and lipid panel 2, 3

The diagnostic accuracy varies by which PCOS criteria are applied—Rotterdam criteria show higher sensitivity (77%) but lower specificity (83%) compared to NIH criteria (sensitivity 51%, specificity 94%). 2 This means Rotterdam criteria will identify more cases but with more false positives, while NIH criteria are more conservative.