Evaluation of Suspected PCOS with Normal Testosterone and High SHBG
In a patient with normal free and total testosterone but elevated SHBG, measure androstenedione (A4) and DHEAS as second-line androgen markers, while simultaneously evaluating for alternative diagnoses that cause elevated SHBG, particularly thyroid disease and conditions causing chronic anovulation without hyperandrogenism. 1, 2
Understanding the Clinical Picture
Your patient presents with an atypical hormonal pattern for PCOS:
- Normal total and free testosterone argues against typical PCOS but does not exclude it, as approximately 30% of women with confirmed PCOS have normal testosterone levels 2
- Elevated SHBG is unusual for PCOS, where SHBG is typically decreased, particularly in those with insulin resistance and higher androgen levels 2, 3
This pattern suggests either:
- A non-hyperandrogenic PCOS phenotype (if Rotterdam criteria are being applied)
- An alternative diagnosis mimicking PCOS
- A condition causing elevated SHBG that masks underlying hyperandrogenism
Immediate Next Steps in Androgen Assessment
Second-Line Androgen Testing
Measure androstenedione (A4) and DHEAS using LC-MS/MS if available, as these may be elevated when testosterone is normal 1, 2:
- Androstenedione has 75% sensitivity and 71% specificity for PCOS diagnosis, though lower than testosterone 1
- DHEAS has 75% sensitivity and 67% specificity, and is particularly valuable in women under 30 years 2
- Both markers have poorer specificity than testosterone/free testosterone and should be interpreted cautiously 1
Critical threshold: If androstenedione >10.0 nmol/L, consider adrenal or ovarian tumor rather than PCOS 2
Calculate Free Androgen Index (FAI)
Calculate FAI = (Total Testosterone/SHBG) × 100, which has 78% sensitivity and 85% specificity for PCOS 1, 4:
- FAI may reveal hyperandrogenism masked by elevated SHBG 4
- However, FAI interpretation is problematic when SHBG is elevated (it's most reliable when SHBG is low) 2
- A cutoff of FAI ≥4.97 provides 71.4% sensitivity and 85.2% specificity 4
Essential Workup to Exclude Alternative Diagnoses
Mandatory Laboratory Tests
Measure TSH to exclude thyroid disease, as hyperthyroidism increases SHBG and causes menstrual irregularity 2, 5:
- Thyroid dysfunction is a common mimic of PCOS
- Elevated SHBG is characteristic of hyperthyroidism
Measure prolactin (morning resting level) to exclude hyperprolactinemia 2, 5:
- Hyperprolactinemia causes anovulation without hyperandrogenism
- Levels >20 μg/L are abnormal 2
- If elevated, confirm with 2-3 samples at 20-60 minute intervals to exclude stress-related elevation 2
Perform 2-hour oral glucose tolerance test with 75g glucose load to assess metabolic status 2, 5:
- All women with suspected PCOS require metabolic screening regardless of androgen levels 2
- Insulin resistance can be present even without hyperandrogenism
Obtain fasting lipid panel (total cholesterol, LDL, HDL, triglycerides) 2, 5
Assess Ovulatory Status
Measure mid-luteal progesterone (day 21 of 28-day cycle or 7 days before expected menses) 2:
- Levels <6 nmol/L confirm anovulation 2
- This helps establish whether oligomenorrhea represents true anovulation
Consider measuring LH and FSH on days 3-6 of menstrual cycle 2:
- LH/FSH ratio >2 suggests PCOS, though this is abnormal in only 35-44% of PCOS cases 2
- Normal or low-normal FSH with elevated LH supports PCOS 2
Ultrasound Evaluation
Perform transvaginal ultrasound with ≥8 MHz transducer to assess for polycystic ovarian morphology 1, 2, 5:
- Gold standard: ≥20 follicles (2-9mm diameter) per ovary (87.64% sensitivity, 93.74% specificity) 1, 6
- Alternative marker: Ovarian volume >10 mL if accurate follicle counting is difficult 1, 6
- Critical caveat: Up to one-third of reproductive-aged women without PCOS have polycystic ovarian morphology on ultrasound 2
Do not use ultrasound if patient is <8 years post-menarche or <20 years old due to high false-positive rates 2, 6
Diagnostic Algorithm for This Specific Scenario
If A4 or DHEAS are elevated:
- Biochemical hyperandrogenism is confirmed (though with lower specificity than testosterone) 1
- Combined with oligo-anovulation, this meets Rotterdam criteria for PCOS 2
- Polycystic ovarian morphology on ultrasound would confirm the diagnosis but is not required 2
If A4 and DHEAS are also normal:
- Consider non-hyperandrogenic PCOS phenotype (oligo-anovulation + polycystic ovarian morphology without hyperandrogenism) if using Rotterdam criteria 2
- Strongly consider alternative diagnoses:
- Hypothalamic amenorrhea (assess for excessive exercise, low body weight, stress) 2
- Thyroid disease (if TSH abnormal)
- Hyperprolactinemia (if prolactin elevated)
- Primary ovarian insufficiency (check FSH if age-appropriate)
If TSH is abnormal:
- Treat thyroid disease first and reassess after achieving euthyroid state
- Menstrual irregularity may resolve with thyroid treatment alone
Important Clinical Pitfalls
Do not use AMH levels for PCOS diagnosis despite elevated levels in PCOS, as there is no standardization, no validated cut-offs, and significant overlap between women with and without PCOS 2, 6:
Do not diagnose PCOS based on ultrasound findings alone, as polycystic ovarian morphology is present in up to one-third of normal women 2
Remember that PCOS can be diagnosed with normal biochemical androgen levels if clinical hyperandrogenism (hirsutism, acne, alopecia) plus oligo-anovulation are present 2:
- Clinical assessment remains paramount 2
- Physical examination for hirsutism using modified Ferriman-Gallwey score is essential 2
Elevated SHBG in the context of suspected PCOS should prompt investigation for causes of increased SHBG:
- Hyperthyroidism
- Estrogen therapy or oral contraceptives (if recently discontinued)
- Hepatic disease
- Anorexia nervosa
If patient is on hormonal contraception (including progestin-only implants), remove or allow to expire before hormonal testing, as these suppress the hypothalamic-pituitary-ovarian axis and make PCOS hormone levels unreliable 2