Total vs free testosterone: what each measures and which one to rely on

Sex hormone–binding globulin holds testosterone so tightly that the bound share is unavailable to cells. When SHBG is high — with age, hyperthyroidism, liver disease, HIV, estrogen, some anticonvulsants, or simply a lean build — more of the total is locked up: the total can read 400 ng/dL while the free is below range, and the man has symptoms of low testosterone. When SHBG is low — obesity, insulin resistance, type 2 diabetes, hypothyroidism, nephrotic syndrome, androgen or steroid use — less is bound, and a total of 250 can go with a normal free value and no deficiency. This is why the Endocrine Society and AUA guidelines ask for a free (or bioavailable) testosterone whenever the total is near the threshold or an SHBG-altering condition is present.

Total testosterone is the better-standardized measurement — immunoassays are calibrated against mass spectrometry through the CDC's standardization program, and the diagnostic thresholds (300 ng/dL in the AUA guideline, 264 in the Endocrine Society's harmonized range) were set on totals. So the sequence is: two fasting morning totals on separate days; if both are clearly low or clearly normal, the question is answered; if either is borderline, or SHBG is likely abnormal, add SHBG and a calculated or dialysis free testosterone, and read that against its own reference range (roughly above 47–65 pg/mL depending on the method).

The repeat is not a formality: 30% of men with one low morning value are normal on the second, and a value drawn in the afternoon, after a poor night's sleep, during an illness or after a large meal runs low in everyone.

Three methods appear on lab reports. Equilibrium dialysis physically separates the free hormone and is the reference method. The calculated free testosterone (Vermeulen or Sodergard formula) uses the total, SHBG and albumin and agrees well with dialysis; most large laboratories report it. The direct 'analog' free testosterone immunoassay, still offered by some labs, under-reads badly and tracks SHBG rather than the true free fraction — the Endocrine Society advises against relying on it. If your report does not say which method was used, assume the least reliable one and ask.

Women's testosterone levels are a tenth of men's, below the accurate range of most total-testosterone immunoassays, so for PCOS, hirsutism and suspected androgen excess the total should be measured by LC-MS/MS and the free testosterone or free androgen index (total ÷ SHBG × 100) calculated. A low SHBG, common in PCOS and insulin resistance, is often the reason a woman with a normal total has androgen symptoms.