Key takeaways
- The thyroid produces mostly T4; the active hormone T3 is made in peripheral tissues by deiodinase enzymes.
- D2 is the accelerator and D3 the brake, and their balance is regulated locally, tissue by tissue.
- TSH reports only what the pituitary makes of circulating T4 — excellent for primary thyroid failure, blind to conversion.
- Serum free T3 is not tissue T3, and the free T3 to reverse T3 ratio has no validated reference standard.
- A suppressed axis usually points upstream to energy deficit, inflammation, illness or poor sleep rather than to the thyroid.
Thyroid testing in most clinics is one number. TSH comes back inside the reference range, the patient is told the thyroid is fine, and the conversation ends. For most people that is correct and efficient. For a minority it is not, and the reason is structural rather than a matter of where the cut-off sits: TSH reports one relationship — what the pituitary makes of circulating thyroxine — and the axis has a step downstream of that which TSH cannot see.
The axis has four steps, not three
Written out, the hypothalamic–pituitary–thyroid axis looks like its siblings. The hypothalamus releases thyrotropin-releasing hormone; the anterior pituitary responds with thyroid-stimulating hormone; the thyroid produces hormone, which feeds back to suppress both levels above it. Three steps, a closed loop.
What makes thyroid different from the adrenal or gonadal axes is that the gland does not mainly produce the active hormone. It produces roughly 80% thyroxine (T4) and about 20% triiodothyronine (T3), and T3 is the molecule that binds nuclear receptors and changes transcription — several times more potent than T4. Most of the T3 the body uses is made outside the thyroid, in peripheral tissues, by removing one iodine atom from T4. So there is a fourth step, conversion, and it is regulated separately from everything above it.
Why TSH moves so dramatically
TSH is the most-ordered thyroid test because it is an excellent one for the job it does. The pituitary is exquisitely sensitive to thyroid hormone, and the relationship between free T4 and TSH is roughly log-linear — a small fall in free T4 produces a disproportionately large rise in TSH. That amplification makes TSH a good early detector of primary thyroid failure: the pituitary notices before the patient does.
It also creates two blind spots. The first is that the pituitary is measuring the hormone available to the pituitary, which relies on local conversion by an enzyme the pituitary is unusually rich in. The pituitary can therefore be well supplied while other tissues are not. The second is that the width of the population reference range for TSH is much larger than the range within which any single individual's TSH normally sits — each person occupies a narrow personal set point inside a broad population band, so a result can be "normal" and still be well displaced from that individual's own baseline (Hoermann et al., Front Endocrinol 2017). The TSH range debate covers where the cut-offs came from and why they are argued over.
The three deiodinases
Conversion is performed by a family of selenium-containing enzymes, and which one is doing the work determines what you get.
- D1 — mainly liver and kidney. Converts T4 to T3 and contributes to the circulating pool.
- D2 — brain, pituitary, skeletal muscle, brown fat. The main producer of intracellular T3, and the enzyme that most tissues rely on to supply themselves.
- D3 — placenta, brain, and induced in many tissues under stress. Removes a different iodine atom, producing reverse T3, which cannot activate the receptor. It also degrades T3 directly.
D2 and D3 are the accelerator and the brake, and they act on the same substrate (Bianco et al., Endocr Rev 2002). Their relative activity, tissue by tissue, sets how much active hormone that tissue actually experiences (Gereben et al., Endocr Rev 2008).
The uncomfortable part: local control
Here is what makes thyroid physiology harder than it looks on a lab report: deiodinase expression is regulated locally. A tissue can raise its own D2 activity to increase its internal T3 supply, or raise D3 to shut it down, without any of that appearing in a blood sample (Köhrle, Free Radic Biol Med 2022).
Serum free T3 measures a real thing — the circulating pool — but it is an imperfect proxy for what any given tissue receives. That cuts both ways. It is the mechanistic reason someone can feel hypothyroid with unremarkable blood work, and equally the reason it cannot be proven from blood that a person's symptoms are caused by inadequate tissue T3. Anyone claiming to measure tissue thyroid status from a serum panel is over-reading their data.
What pushes the balance toward the brake
D3 induction and D1 suppression are not malfunctions but an adaptive response — the body lowering its metabolic set point when conditions are poor. The pattern is starkest in serious illness, where circulating T3 falls, reverse T3 rises and TSH stays inappropriately normal; it is generally understood as adaptation rather than a thyroid disease requiring treatment (Fliers et al., Lancet Diabetes Endocrinol 2015).
Milder versions are far more common, and the triggers are mostly self-inflicted: sustained energy deficit, heavy training without matching intake, chronic inflammation, poor sleep, prolonged stress. Substrate matters too — the deiodinases are selenoenzymes and thyroid peroxidase needs iron, so a deficiency in either constrains the system independent of any thyroid pathology (Köhrle, Int J Mol Sci 2023). Selenium and thyroid and iodine status cover the two most relevant.
Reverse T3, and where the enthusiasm outruns the evidence
Reverse T3 is structurally almost identical to T3 and biologically inert at the receptor. Elevated reverse T3 genuinely indicates that D3 activity is high — that the system is throttling back. As a physiological signal, that is real.
What is not established is the clinical use it gets put to. The free T3 to reverse T3 ratio is widely quoted as a measure of "conversion status", with cut-offs presented as if validated. They are not: there is no agreed reference standard, assays vary, and no trial has shown that treating on the basis of it improves outcomes. Reverse T3 also rises with non-thyroidal illness, fasting and some medications, so an abnormal value frequently points somewhere other than the thyroid.
The defensible position is that a raised reverse T3 is a prompt to look for what is suppressing the system — under-eating, inflammation, illness, sleep debt — rather than a diagnosis in its own right or an automatic indication for T3 therapy. Reverse T3 and chronic stress works through that distinction.
The clinical pearl: TSH answers one question well — is the thyroid gland itself failing. It does not answer whether peripheral conversion is adequate, whether autoimmunity has started, or whether the axis is being suppressed by something outside the thyroid entirely. Those require different tests, and the honest limit is that none of them measure tissue thyroid status directly.
What to test, in what order, and what to expect
A defensible panel is TSH and free T4 first — the pair that identifies primary and central thyroid disease. Add free T3 to see the circulating active pool, thyroid peroxidase and thyroglobulin antibodies to detect autoimmunity, which often precedes any TSH abnormality by years (see thyroid antibodies), and reverse T3 only when there is a specific reason to ask whether the axis is being suppressed. Ferritin and vitamin D are worth having alongside, because iron deficiency produces an almost identical symptom set and is far more common. Free T3 versus free T4 covers how the two are read together.
What follows is a sequence, not a menu. If TSH is clearly abnormal, that is a thyroid problem and is treated as one. If TSH and free T4 are normal but free T3 sits low with reverse T3 raised, the useful question is not which medication but what is suppressing the axis — usually energy intake, sleep, inflammation or training load, all fixable without a prescription. Positive antibodies with normal function need watching rather than treating. Expect upstream fixes to work slowly: markers move over months, symptoms lag the numbers, and retesting sooner than about three months mostly generates noise.
Bottom line
The thyroid axis has four steps, and the fourth — peripheral conversion of T4 to T3 by deiodinase enzymes — is regulated separately from the pituitary feedback that TSH reports on. That is why TSH can be normal while the system is throttled back, and why a full panel including free T4, free T3 and antibodies sees patterns a single number cannot. The limits are worth stating plainly: serum T3 is not tissue T3, the reverse T3 ratio is not a validated test, and a suppressed axis usually reflects something upstream — under-eating, inflammation, illness, poor sleep — that is better addressed directly than medicated around.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
