Free vs Total Testosterone in PCOS: Why One Can Be Normal and the Other Not

What is the actual difference between free and total testosterone?

Total testosterone counts every testosterone molecule in a blood sample, whether or not it is biologically active; free testosterone counts only the roughly 1–2% that is not bound to a carrier protein and can actually reach a receptor. Sex hormone-binding globulin (SHBG), a liver-made protein, binds the great majority of circulating testosterone and holds it inactive. A 2025 diagnostic meta-analysis of 2,857 women (1,650 with PCOS) compared how well each androgen measure identifies biochemical hyperandrogenism and found calculated free testosterone reached 89% sensitivity against 74% for total testosterone alone — meaning total testosterone, used by itself, misses roughly one in four true cases. That finding is the evidence base the 2023 international guideline used when it named total and free testosterone, together, as the first-line tests for PCOS — not total testosterone on its own, and read alongside the rest of the standard PCOS blood panel.

The reason the two numbers split apart is SHBG. When SHBG is low, more testosterone circulates unbound, so free testosterone climbs even while total testosterone sits inside its printed reference range. When SHBG is high, the opposite happens: more of an already-elevated total testosterone gets locked up, so free testosterone can look ordinary. Total testosterone alone cannot tell you which situation you are in — you need SHBG in the same blood draw to know what your total testosterone number actually means.

Why can total testosterone be normal while free testosterone is not?

SHBG suppression by circulating insulin is the single most common reason total testosterone reads normal in a person who has real biochemical hyperandrogenism. Insulin directly lowers the liver’s SHBG output, and the insulin-resistant PCOS phenotype often carries the lowest SHBG of the four Rotterdam phenotypes. With less SHBG competing for testosterone, a total testosterone concentration that would be unremarkable in someone with typical SHBG frees up enough active hormone to drive hirsutism and acne. This is precisely the pattern behind the site’s most common diagnostic near-miss: a total testosterone report that comes back “normal” while the person has visible clinical androgen excess.

What is biochemical vs. clinical hyperandrogenism?

Clinical hyperandrogenism is what a clinician can see or a patient can report — a hirsutism score, cystic jawline acne, androgenic hair thinning — while biochemical hyperandrogenism is what a blood test measures. The two do not always agree, and either one alone satisfies the hyperandrogenism leg of the Rotterdam criteria. A person can have visible hirsutism with androgens that measure inside a lab’s normal range, usually because hair follicles in androgen-sensitive areas convert testosterone to the far more potent dihydrotestosterone locally, at the follicle, in amounts a blood draw never sees. The reverse also happens: androgens can measure above range with little or no visible hair or skin change, especially early, or in phenotypes where the hyperandrogenism is adrenal rather than ovarian in origin.

This is why neither category outranks the other in the diagnostic criteria. A clinician who orders bloodwork, sees a normal total testosterone, and concludes “no hyperandrogenism” is skipping both the free testosterone question and the clinical exam that the guideline treats as equally valid evidence.

What does the free androgen index actually measure — and how reliable is it?

The free androgen index (FAI) is a ratio, not a direct measurement: total testosterone divided by SHBG, multiplied by 100. It became popular because it needs no extra assay beyond the two numbers labs already report, but it is a cruder tool than a properly calculated free testosterone value. The original 1999 comparison of free-testosterone estimation methods found that FAI’s relationship to true bioavailable testosterone shifts as SHBG shifts — meaning FAI becomes systematically less reliable at exactly the high- and low-SHBG extremes where the diagnostic question matters most. A calculated free testosterone value, derived from total testosterone and SHBG using an equilibrium-based formula, tracked the gold-standard equilibrium dialysis result far more consistently across that same SHBG range.

A separate 2021 meta-analysis limited to FAI alone, pooling seven studies, found similar numbers — 76% sensitivity and 82% specificity for distinguishing PCOS from controls, with an area under the curve of 0.84. Set against the table above, FAI performs close to total testosterone on raw accuracy, but not as well as calculated free testosterone on sensitivity — the measure most likely to catch a case total testosterone alone would miss. FAI is still useful as a quick screen precisely because it costs nothing extra to calculate; it is a poorer tool than a calculated free testosterone for a borderline case that will determine treatment. The two weaker performers in the table, androstenedione and DHEA-S, are adrenal rather than ovarian androgens — worth adding only when total and free testosterone are not elevated, per the guideline, not as a substitute for either.

What does high SHBG do to your results, and who does it affect?

High SHBG suppresses free testosterone toward normal even when total testosterone is genuinely elevated, which is the mirror image of the low-SHBG problem and gets far less attention. Three situations reliably raise SHBG enough to matter: combined hormonal contraceptive use, which can raise SHBG several-fold and stays elevated for weeks after stopping; hyperthyroidism; and, less predictably, the leaner end of the PCOS population, where insulin resistance — and its SHBG-suppressing effect — is milder or absent. In any of these situations, a total testosterone that would flag as hyperandrogenic in someone with average SHBG can produce a calculated free testosterone that reads inside range, because more of that same total testosterone is bound and inactive.

What should you actually ask for?

Ask for total testosterone and SHBG drawn together, so free testosterone can be calculated rather than guessed at from total testosterone alone, as one part of the full PCOS diagnostic work-up. If your total testosterone comes back normal but you have visible hirsutism, acne, or androgenic hair loss, ask specifically whether SHBG was run and whether a calculated free testosterone or FAI was generated from it — many standard panels report total testosterone and SHBG as separate line items without ever computing the ratio a clinician needs to interpret them together. If the calculated result is still borderline and it will change your diagnosis or treatment, a repeat total testosterone by LC-MS/MS is worth requesting, since assay method affects the input to every one of these calculations. If glucose intolerance is also on the table, the OGTT with insulin is the companion test most work-ups run alongside this one.

Note: in May 2026, PCOS was renamed polyendocrine metabolic ovarian syndrome, or PMOS, by a global consensus of more than 50 organisations. The hormone measurements and the reasoning in this article are unchanged by the rename; this article uses PCOS because that is still the term most readers search.

Reading this by phenotype

In the insulin-resistant phenotype, low SHBG is the norm, so free testosterone and FAI run ahead of total testosterone — expect the free measure to be the one that confirms hyperandrogenism, not the total. In the lean, hyperandrogenic phenotype, SHBG is often normal or high, so total and free testosterone tend to move together, and a genuinely elevated total testosterone is less likely to be masked. In the ovulatory-but-hyperandrogenic phenotype, androgens run above range on either measure while cycles stay regular — the SHBG question matters less here because the total testosterone elevation is usually clear enough on its own. In the non-hyperandrogenic phenotype, every androgen measure in this article can come back inside range regardless of SHBG, because the diagnosis in that phenotype does not rest on androgens at all.

Who this doesn’t work for

Calculating free testosterone from total testosterone and SHBG does not help if the total testosterone itself was run on a low-quality immunoassay at a concentration near the assay’s floor — the calculation only compounds an unreliable input. It also will not settle a borderline case drawn while on combined hormonal contraception, since the pill suppresses ovarian androgen output and raises SHBG at the same time, distorting both halves of the calculation in the same direction. And none of these androgen measures — total, free, or FAI — diagnose PCOS by themselves; they satisfy one leg of the Rotterdam criteria, and a person can be diagnosed, or excluded, without ever needing this specific number if the rest of the diagnostic work-up settles the question instead.