Evaluation of Elevated DHEA-S and Testosterone in Suspected PCOS
Your patient's elevated DHEA-S (314) and testosterone (60) support biochemical hyperandrogenism, but you should prioritize measuring total testosterone and calculated free testosterone (or FAI) using LC-MS/MS as first-line tests, as DHEA-S has poor diagnostic specificity for PCOS and should only be considered when testosterone-based measurements are equivocal. 1, 2
Diagnostic Approach to Androgen Measurement
First-Line Testing Strategy
Measure total testosterone (TT) and free testosterone (FT) as your primary diagnostic tests for biochemical hyperandrogenism in PCOS, as these have the highest sensitivity (0.74 for TT, 0.89 for cFT) and specificity (0.86 for TT, 0.83 for cFT) with AUCs of 0.87 and 0.85 respectively 1
Calculate free testosterone using the Free Androgen Index (FAI) when direct measurement is unavailable, as FAI demonstrates sensitivity of 0.78 and specificity of 0.85 with an AUC of 0.87 1
Ensure SHBG levels are ≤170 nmol/L before relying on calculated free testosterone, as values above this threshold produce unreliable calculations due to excessive SHBG fluctuations 3
Role of DHEA-S in Diagnosis
DHEA-S has limited diagnostic utility with pooled sensitivity of only 0.75 and specificity of 0.67 (AUC 0.77), making it significantly inferior to testosterone-based measurements 1, 2
Consider DHEA-S only as a second-line marker when first-line testosterone tests are negative but clinical suspicion remains high, or when signs of virilization are present 2, 4
Age-adjusted reference ranges are essential for DHEA-S interpretation, as levels decline approximately 2% per year after age 20-30, and failure to use age-specific norms leads to overdiagnosis 2, 5
In women aged 20-29 years, elevated DHEA-S occurs in only 8-11% of PCOS cases (33% in non-age-adjusted studies), and is more common in non-classic phenotypes B and C than phenotype A 5, 6
Assay Method Considerations
LC-MS/MS vs Immunoassay
LC-MS/MS is the preferred method for androgen measurement due to superior accuracy, with higher sensitivity for calculated free testosterone (0.89 vs 0.74), FAI (0.89 vs 0.74), and androstenedione (0.86 vs 0.64) compared to direct immunoassays 1
For DHEA-S specifically, LC-MS/MS shows higher sensitivity (0.82) but lower specificity (0.57) compared to direct immunoassays (sensitivity 0.67, specificity 0.70), though differences are less pronounced than for testosterone 1, 2
Excluding Other Causes of Hyperandrogenism
When to Suspect Androgen-Secreting Tumors
Measure testosterone and DHEA-S to screen for neoplasms when severe virilization with recent onset is present, though the positive predictive value is only 9% for testosterone >8.7 nmol/L (250 ng/dL) 4, 7
Total testosterone >8.7 nmol/L has 100% sensitivity and 98% specificity for androgen-secreting tumors, but occurs in only 2.3% of hyperandrogenic patients 7
DHEA-S >16.3 μmol/L (6000 ng/mL) should prompt adrenal imaging, though the positive predictive value for adrenal tumors is extremely low in the absence of rapid virilization 7
Other Differential Diagnoses
Exclude non-classical congenital adrenal hyperplasia (NCCAH) by measuring basal or ACTH-stimulated 17-hydroxyprogesterone, particularly if oligomenorrhea began after menarche 4
Rule out Cushing's disease with overnight dexamethasone suppression test or 24-hour urinary free cortisol if signs of hypercortisolism accompany recent-onset hyperandrogenism 4
Check prolactin levels in women with recent-onset oligomenorrhea and mild hyperandrogenism to exclude hyperprolactinemia 4
Completing the PCOS Diagnosis
Rotterdam Criteria Application
Diagnose PCOS when two of three criteria are met: ovulatory dysfunction, clinical or biochemical hyperandrogenism, and polycystic ovary morphology on ultrasound or elevated anti-Müllerian hormone 1
Rotterdam criteria show higher sensitivity (0.77) but lower specificity (0.83) compared to NIH criteria (sensitivity 0.51, specificity 0.94) for biochemical hyperandrogenism diagnosis 1
Common Pitfalls to Avoid
Do not use DHEA-S as a first-line test for PCOS diagnosis due to poor specificity and low prevalence of elevation, as only 8.1% of PCOS patients have truly elevated age-adjusted DHEA-S 2, 5
Avoid relying on calculated free testosterone when SHBG >170 nmol/L, as this produces inaccurate results; use equilibrium dialysis for direct measurement instead 3
Do not screen for androgen-secreting tumors based solely on androgen levels without clinical features of rapid virilization, as the low frequency and poor positive predictive value make this cost-ineffective 7
Ensure blood samples are collected in the morning (8-10 AM) when testosterone levels are highest for accurate assessment 3