Free T4 Blood Test (T4, Free, Direct): What Your Number Means

At a Glance

Thyroxine — T4 — travels through the blood almost entirely stuck to carrier proteins. Only about 0.03%, roughly one molecule in three thousand, is unbound and able to enter tissues. Free T4 measures that fraction.

Reports print it several ways — Free T4, FT4, T4, Free (Direct), Thyroxine, Free — for the same measurement.

Then the part almost nobody is told. There is no single normal range for free T4, and it changes over time as well as between laboratories. One university hospital's interval was 0.6 to 1.2 ng/dL until a new analyzer moved it to 0.9 to 1.7 in August 2026 — a 1.4 went from above range to mid-range with nothing about the patient changing. This record stores Labcorp's 0.82 to 1.77.

So the number alone is not the finding. The number read against the interval printed beside it is — and alongside your TSH.

What your number means

Each value below is read against this record's stored interval — 0.82 to 1.77 ng/dL, or 10.55 to 22.78 pmol/L. Use yours if it differs.

Two things this table cannot do. It cannot say whether a value near an edge means anything — that depends on your TSH, your symptoms, and when you last took thyroid medication. And it does not transfer to a different interval: against the 0.6 to 1.2 some assays used until recently, the 1.4 here was raised.

The conversion is fixed: ng/dL × 12.87 = pmol/L.

How to read your result

Free T4 is not read alone. Endotext: assessment "requires both a sensitive serum TSH assay and a valid T4 estimate, with interpretation based on the relationship between these two values."

The fifth row is worth slowing down for. Central hypothyroidism is usually described as "low TSH with low free T4", but the UK guidelines are more accurate: TSH "can be low, within or mildly above the reference range" — so neither a normal nor a mildly raised TSH excludes it.

The last row needs its order stated, because the order is the point. The 2026 European Thyroid Association guideline names that combination — "elevated T4 but TSH within the reference interval" — as a reason to consider interference, checking medications and supplements first. Rare causes such as thyroid hormone resistance or a TSH-secreting pituitary tumor are investigated only after interference has been definitively excluded.

Why two laboratories disagree about your number

Not rounding. When the IFCC compared commercial free T4 assays against the reference procedure, only 2 of 17 came within 10%; 15 read low, by up to 42%.

That is why the University of Iowa moved its adult range from 0.90–1.70 to 0.80–1.80 in one revision, and why the two current intervals named above are both Roche assays and still differ — 0.82–1.77 and 0.9–1.7. The retired assay's own test record puts it outright: results from assays using "different manufacturers, instruments, or methods are not comparable."

Units and specimen change the number too. The same hormone is ng/dL in the US and pmol/L elsewhere — 1.0 ng/dL is about 12.9 pmol/L, so a value near 13 is unremarkable in one system and impossible in the other. Free T4 is also measured on dried blood spot and in urine. A number without its unit and specimen is not a result.

Practically: do not compare a free T4 from one laboratory against another's, or against a range found online. A trend holds within one laboratory and one method.

What "Direct" and "Non-Dialysis" mean on your report

Neither says anything about whether your result is high, low or normal — and they are not interchangeable, which is where reports mislead people.

"Non-Dialysis" is the more informative, but it rules one thing out rather than naming another: the result did not come from equilibrium dialysis. It does not say which method did — ultrafiltration with mass spectrometry is also direct and also non-dialysis. If that matters, the laboratory's method documentation answers it.

"Direct" is laboratory naming and cannot tell you the method. Labcorp calls its Roche immunoassay T4, Free, Direct. Quest uses it the other way: its equilibrium-dialysis test is T4, Free, Direct Dialysis, its routine immunoassay simply T4, Free. In the specialist literature "direct" describes methods that physically separate free from bound hormone. So the word marks the quick assay at one national laboratory and the reference method at another: check the test's full name.

What an assay assumes is the part that matters. Quest describes its own Centaur immunoassay as one that "assumes that the T4 binding capacity is normal" — results "can be impacted when the albumin or thyroid binding globulin levels are abnormal." Equilibrium dialysis makes no such assumption, and the gap is not small: Quest reports that assay ran "generally 40% to 60% lower" than dialysis, and recommends repeating by dialysis when a result does not fit the clinical picture. Pregnancy, severe illness, low albumin and inherited binding-protein variants are where they part company.

If you remember only one thing

Your free T4 is only meaningful against the interval printed on your own report, and only alongside your TSH. A value someone else calls normal may be high on your assay, and a value inside your range can still be wrong if you took biotin, were acutely unwell, or swallowed a levothyroxine tablet a few hours before the draw.

Two people, same result

Two reports both read 1.4 ng/dL.

The first was screened at a routine physical; her laboratory prints 0.82–1.77 and her TSH is 2.1. Both mid-interval. Nothing here to chase.

The second belongs to someone taking 5,000 micrograms of a hair-and-nails biotin supplement daily. His laboratory prints the same 0.82–1.77, so his 1.4 is inside the interval too — but his TSH came back suppressed. An upper-range free T4 with a low TSH is the pattern of thyrotoxicosis — and, on a susceptible assay, what a large biotin dose manufactures, pushing free T4 up and TSH down at once. Which it is cannot be told from the numbers. And even if the pattern is real, thyrotoxicosis has several causes — TSH-receptor antibodies confirm Graves', not these two numbers.

Same number, same interval — the second value on the report, and what else is in the blood, decide it.

What this result cannot tell you

Whether your thyroid is over- or underactive. That is the free T4 and TSH pair.

Whether an abnormal value will persist. Many are repeated — an unexpected result, a borderline one, one that could have been distorted. But a repeat is not a precondition for a result to mean something: a clear pattern with symptoms is assessed, not deferred.

Whether the assay was fooled. Biotin on a susceptible platform, inherited binding-protein variants and interfering antibodies all produce a wrong number that looks right.

Anything reliable if you were acutely unwell. NICE: do not test during an acute illness unless you suspect the illness is caused by it.

How well your levothyroxine dose is set — usually. In primary hypothyroidism TSH answers that. In central hypothyroidism it is the other way round: the Endocrine Society recommends against using TSH to adjust replacement there, and free T4 guides it instead.

Common interpretation mistakes

Treating a value near the edge as a diagnosis. Intervals are built so that a portion of healthy people fall outside them.

Assuming a high free T4 with a low TSH means Graves' disease. That pair indicates thyrotoxicosis, not its cause — UK guidance confirms Graves' with TSH-receptor antibodies. It is also the pattern biotin, a recent levothyroxine dose and several assay interferences produce.

Treating a normal free T4 as proof that symptoms are not thyroid-related. NICE notes people "may feel well even when their thyroid function tests are outside the reference range", and that treatment changes may help some whose tests are inside it. What the evidence does not support is an "optimal" free T4 narrower than your laboratory's interval.

Reading your number against a range from the internet, another laboratory, or last year's test elsewhere. A trend holds within one method only.

Questions your doctor may ask

Do you take biotin or any hair, skin and nails supplement — and when last?

Do you take levothyroxine, and how many hours before the draw?

Were you acutely unwell or in hospital when this was drawn?

Are you pregnant, or taking estrogen or an oral contraceptive?

Any amiodarone, lithium, phenytoin, carbamazepine or immunotherapy?

Any earlier results from this same laboratory to compare against?

Read together with

TSH — read together with free T4; neither stands alone.

Free T3 — the receptor-active hormone, added when a low TSH needs explaining.

T4, Total (Thyroxine) — bound plus unbound; it moves when binding proteins move.

Free Thyroxine Index — an older calculated estimate.

Thyroid Peroxidase Antibodies — what separates autoimmune thyroid disease from other causes.

Clinical pearls

Biotin is the interference to ask about first — on the assays it affects. The 2026 ETA guideline scopes it to streptavidin–biotin-based assays, where a large dose pushes free T4 up and TSH down at once. Susceptibility belongs to the platform, not to free T4 testing as such. There is a published literature of patients investigated, and some treated, for Graves' that turned out to be a supplement. How long to stop depends on dose and assay: the FDA said at least 8 hours above 5 mg, the Synthroid label at least 2 days, and Quest 72 hours for massive doses — which it defines as 5,000 to 10,000 micrograms, exactly what those supplements contain. Only 72 hours covers the longest. Follow what the laboratory asks.

A levothyroxine tablet raises free T4 for hours. The American Thyroid Association describes "a transient peak in serum T4 and free T4 levels of about 15% magnitude about 4 hours after" the dose. UK guidelines judge this "not clinically significant" for dose decisions — both hold, because in primary hypothyroidism the dose is judged by TSH. A patient drawn hours after the tablet can show a free T4 at the top of the interval with an appropriate TSH.

A high free T4 in someone who feels well may be their albumin. Familial dysalbuminemic hyperthyroxinemia — a mutant albumin binding T4 ten times more tightly, in roughly 1 in 10,000 people of European ancestry — reads falsely high on immunoassays, normal by dialysis.

In pregnancy the interval matters more than the method. The 2026 ATA guideline calls analog free T4 "an appropriate marker of thyroid function in the first half of pregnancy" and asks for a laboratory- and trimester-specific range as "the preferred method." Where it is hard to interpret, TSH takes priority.

Clinical Takeaway

Free T4 measures the roughly 0.03% of thyroxine not bound to carrier proteins — by LOINC a quantitative serum measurement with no method specified, and method is precisely what determines the interval: 15 of 17 assays in the IFCC comparison read low against the reference procedure, by up to 42%. Interpretation is the TSH–free T4 relationship rather than either value alone, and a normal TSH does not exclude central hypothyroidism. Quest's routine Centaur immunoassay assumes normal binding-protein capacity and ran 40% to 60% below equilibrium dialysis in its own comparison; "direct" and "non-dialysis" are laboratory labels rather than method descriptions, so a result's assumptions belong to its named assay. Before acting on an abnormal value, exclude high-dose biotin on a susceptible assay, recent levothyroxine dosing, acute non-thyroidal illness and binding-protein anomalies — often with a repeat, though a clear pattern with symptoms is assessed rather than deferred.

In one sentence

Free T4 is the small unbound fraction of thyroxine, and your number means what the interval printed beside it and your TSH say it means — not what a number alone says.

Bottom line

Find the reference interval printed next to your result and read your number against that. Intervals differ between laboratories by more than the width of the abnormal zone, and change when a laboratory changes analyzer.

Then find your TSH on the same report. The two together say more than either alone; the patterns worth naming are visible only in the pair.

If your value is outside the interval, the next steps are ordinary rather than alarming: the things that produce a wrong number get excluded, often with a repeat. High-dose biotin, a levothyroxine tablet taken a few hours before the draw, and a test during an acute illness are the ones worth checking, and all three are free to rule out. If you are pregnant, say so — free T4 is then read against a pregnancy-specific interval.