In a reproductive‑age woman being evaluated for polycystic ovary syndrome, are total testosterone, sex hormone‑binding globulin, free testosterone, human chorionic gonadotropin, and prolactin the appropriate labs to order?

Laboratory Testing for PCOS Evaluation

Your proposed panel is incomplete and includes one unnecessary test. The essential labs for PCOS assessment are total testosterone (preferably by LC-MS/MS), sex hormone-binding globulin (SHBG), and prolactin; however, you should measure TSH instead of HCG, and you must add a 2-hour oral glucose tolerance test and fasting lipid panel to properly evaluate metabolic complications. 1, 2

Core Androgen Assessment

First-line androgen testing should include:

• Total testosterone measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS), which demonstrates 74% sensitivity and 86% specificity for detecting hyperandrogenism 1, 2
• SHBG to calculate the free androgen index (FAI = total testosterone/SHBG ratio), which shows 78% sensitivity and 85% specificity 1
• Calculated free testosterone using the Vermeulen equation from high-quality total testosterone and SHBG measurements, which has the highest sensitivity at 89% with 83% specificity 1, 2

Direct immunoassays for free testosterone should be avoided because they have poor accuracy at the low concentrations typical in women (78% specificity versus 92% for LC-MS/MS). 1, 2

Essential Exclusion Tests

You must measure these hormones to rule out mimicking conditions:

• TSH to exclude thyroid disease as a cause of menstrual irregularity 1, 2
• Prolactin using morning resting serum levels to rule out hyperprolactinemia; women with PCOS have a 3.15-fold increased risk of elevated prolactin 1, 2

HCG is not part of the standard PCOS workup unless you are specifically excluding pregnancy before starting treatment. 1

Mandatory Metabolic Screening

All women with PCOS require metabolic assessment regardless of BMI:

• 2-hour oral glucose tolerance test with 75-gram glucose load to screen for impaired glucose tolerance or type 2 diabetes 1, 2
• Fasting lipid panel (total cholesterol, LDL, HDL, triglycerides) to assess cardiovascular risk 1, 2

Insulin resistance occurs in PCOS independent of body weight, making these tests essential even in lean patients. 1

Second-Line Androgen Tests (Only If First-Line Normal)

If total testosterone and calculated free testosterone are not elevated but clinical suspicion remains high:

• Androstenedione (A4) – 75% sensitivity, 71% specificity 1, 2
• DHEAS – 75% sensitivity, 67% specificity; particularly useful when SHBG is low or in women <30 years 1, 2

These have poorer specificity than testosterone measurements and should only be used as adjunctive tests. 2

Critical Diagnostic Considerations

PCOS can be diagnosed with normal laboratory values. The Rotterdam criteria require only two of three features: oligo/anovulation, clinical and/or biochemical hyperandrogenism, and polycystic ovarian morphology on ultrasound. 1 This means a woman with clinical hyperandrogenism (hirsutism, acne, or alopecia) plus irregular cycles can be diagnosed with PCOS even if all androgen levels return normal. 2

Total testosterone is abnormal in only 70% of women with confirmed PCOS, meaning 30% have normal testosterone levels despite having the condition. 2 This is why calculated free testosterone and FAI improve diagnostic sensitivity.

Timing and Method Considerations

• Testosterone should be measured in the morning (8-10 AM) due to diurnal variation 2, 3
• Testing should be performed off hormonal contraception because progestins suppress the hypothalamic-pituitary-ovarian axis and alter SHBG levels 1
• LC-MS/MS is mandatory when available; direct immunoassays have unacceptably high false-positive rates in women 1, 2, 3

Common Pitfalls to Avoid

Do not rely on LH/FSH ratio alone. An LH/FSH ratio >2 suggests PCOS but is abnormal in only 35-44% of women with the condition, making it a poor diagnostic marker. 2

Do not use AMH for diagnosis. Although AMH ≥35 pmol/L shows 92% sensitivity and 97% specificity in research, it should not replace ultrasound or serve as a standalone test due to lack of assay standardization and no validated cut-offs. 1, 2

Do not order insulin or C-peptide routinely. Physical signs of insulin resistance (elevated BMI, acanthosis nigricans) are sufficient for assessment; direct insulin measurement is not recommended. 2