No, testosterone levels are NOT always elevated in PCOS
Testosterone levels are elevated in only approximately 70-75% of women with PCOS, meaning 25-30% of PCOS patients have normal testosterone levels despite meeting diagnostic criteria. This is a critical clinical pitfall that can lead to missed diagnoses if clinicians rely solely on testosterone measurement.
Understanding the Diagnostic Reality
The most recent 2025 International PCOS Guidelines meta-analysis demonstrates that total testosterone has a sensitivity of only 74% for detecting biochemical hyperandrogenism in PCOS 1. This means roughly one in four women with PCOS will have normal total testosterone levels. Free testosterone performs slightly better with 89% sensitivity, but still misses approximately 11% of cases 1.
Why Some PCOS Patients Have Normal Serum Testosterone
Several mechanisms explain normal circulating testosterone in PCOS:
Follicular hyperandrogenism: Research shows that 64.9% of PCOS patients with normal serum androgen levels display abnormally high follicular fluid androgen levels 2. The ovarian microenvironment may be hyperandrogenic even when systemic levels appear normal.
Phenotypic heterogeneity: The Rotterdam criteria recognize four distinct PCOS phenotypes. One phenotype (oligoanovulation + polycystic ovaries without overt hyperandrogenism) accounts for a significant subset of patients who have milder endocrine abnormalities 3.
Age-related changes: Testosterone levels in PCOS women decrease by approximately 50% around ages 42-47, though they remain elevated compared to controls in younger and older age groups 4.
Diagnostic Approach When Testosterone is Normal
If initial total testosterone (TT) or free testosterone (FT) measurements are not elevated, measure androstenedione (A4) and DHEAS 1. The 2025 guidelines explicitly recommend this stepwise approach, though A4 and DHEAS have poorer specificity (71% and 67% respectively).
Recommended Testing Algorithm:
- First-line: Measure total testosterone AND free testosterone using mass spectrometry (LC-MS/MS) when available
- If mass spectrometry unavailable: Use free androgen index (FAI) as an alternative
- If TT and FT are normal: Consider measuring androstenedione and DHEAS
- Remember: PCOS diagnosis requires only 2 of 3 criteria (hyperandrogenism, ovulatory dysfunction, polycystic ovaries), so biochemical hyperandrogenism is not mandatory 5
Critical Clinical Pitfalls
Do not exclude PCOS based solely on normal testosterone levels. The 2013 Endocrine Society guidelines explicitly state that PCOS can be diagnosed with ovulatory dysfunction plus polycystic ovaries alone, without any evidence of hyperandrogenism 5. Historical data confirms that when using clinical features plus ultrasound, total testosterone was abnormal in only 70% of PCOS cases 6.
Key Points for Practice:
- Clinical hyperandrogenism (hirsutism, acne, androgenic alopecia) can establish the diagnosis even with normal biochemical androgens 5
- Half of women with PCOS may have normal free testosterone concentrations 7
- The combination of oligomenorrhea/amenorrhea plus polycystic ovary morphology on ultrasound is sufficient for diagnosis without any androgen elevation 3
- Always exclude mimicking conditions (thyroid disease, hyperprolactinemia, nonclassic congenital adrenal hyperplasia) regardless of testosterone levels 5
The bottom line: Testosterone elevation is common but not universal in PCOS. Relying exclusively on testosterone measurement will miss 25-30% of cases. Use the Rotterdam criteria's "2 of 3" approach, incorporating clinical assessment, ovulatory function, and ovarian morphology alongside—but not dependent upon—androgen measurements.