From T4 to T3: the conversion that so much depends on

You are sitting there with a lab report. TSH normal. Free T4 normal. Maybe a free T3 underneath, rather towards the lower edge.

And below that the sentence you know: your thyroid is fine.

And still there are cold fingers in June. A run up you need in the morning. A fog around eleven that no amount of coffee clears away.

In my practice I hear this pattern often. And almost every time we then talk about something that does not appear on the lab report at all: about what becomes of the T4 after it has left the thyroid.

Because T4 is not yet the finished hormone. For active T3 to be made from it, a single iodine atom has to be removed, at exactly the right position. Three enzymes do this work, and they sit distributed across the whole body. They are called deiodinases.

1. The thyroid supplies the raw material, not the finished hormone

Picture a restaurant. There is a storeroom, and there are kitchens. The storeroom supplies ingredients. What ends up on the plate is decided by the kitchens.

The thyroid is the storeroom. It mainly releases thyroxine, T4 for short, carrying four iodine atoms. T4 is stable, it lasts a long time in the blood and it makes a good reserve. At the receptors in the cell nucleus it is barely active itself.

The active molecule is called triiodothyronine, T3 for short, and it carries three iodine atoms instead of four. One atom of difference, and almost the entire effect hangs on it.

Here is the decisive point: the greater part of your T3 is not made in the thyroid. It is made in the tissues, out of the T4 that was delivered there.

Before anything can be converted, the hormone has to get into the cell. There are dedicated transporters in the membrane for that, among them the MCT, OATP and LAT families. Only inside does it become clear what happens next. If T4 meets an activating enzyme, T3 is formed. If it meets an inactivating one, reverse T3 is formed, a molecule that is silent at the receptor.

And now you know why a normal T4 value still says nothing about how much active hormone arrives in your muscle.

2. Three enzymes, three roles: D1, D2 and D3

Remember three names. Deiodinase type 1, type 2 and type 3, or D1, D2 and D3 for short.

All three do the same craft: they remove an iodine atom. The difference lies in which position they remove it from and where they sit. And that difference decides whether an active or a silent molecule is left at the end.

D1, the supplier. It sits above all in liver and kidney, works at high throughput and releases its product mainly into the blood. When you think of the T3 level in serum, you are largely thinking of this enzyme.

D2, the self supplier. It sits in the endoplasmic reticulum and produces T3 mainly for the cell it is inside. Brain, pituitary, muscle and brown adipose tissue cover a considerable share of their own needs this way.

D3, the brake. Out of T4 it makes reverse T3, out of T3 it makes T2. Not a defect but a safety valve, especially abundant in the placenta and in tissues under inflammation or a lack of oxygen. What the reverse T3 value can do is covered in Reverse T3: what the value can do.

Read that last sentence again. Plenty of T4 can turn down the activity of the enzyme that converts T4 into T3. That is not a disturbance, that is regulation.

The same group described in 2018 that the catalysis itself triggers this switch, and that the fine tuning usually runs through opposing changes of D2 and D3. The body rarely turns only one dial. And the enzymes respond to many signals, among them NF-kappaB, bile acids, growth factors and HIF-1alpha, the switch for a lack of oxygen.

And now you know why the question is not whether enough material is there, but under which conditions the enzymes are working.

3. Nothing runs without selenium, and more selenium still does not always shift something

Anyone searching online for an impaired T4 to T3 conversion lands on selenium within two clicks. Almost always with a product link next to it. The biochemical part of that story is correct. The rest is more complicated, and both halves belong together.

The active site of all three deiodinases contains selenocysteine, that is cysteine in which sulfur has been swapped for selenium. It is precisely this selenium atom that attacks the iodine. Selenium here is not a booster, it is building material.

Josef Köhrle from the Berlin Charité has described this several times: selenium enables the synthesis and turnover of thyroid hormones, and it has two roles. It builds the converting enzyme, and it builds the glutathione peroxidases, which may protect the glandular tissue from oxidative stress.

A Swedish randomized trial from 2024 points in the same direction. 414 older people with a low selenium status received selenium yeast together with coenzyme Q10 or placebo over 48 months. Afterwards the free T3 was significantly higher, the free T4 lower, and the rise in TSH was smaller than under placebo. One limitation: the selenium share cannot be cleanly separated from the coenzyme Q10.

And now the other half. This paragraph is missing from almost every popular page on the topic.

Zinc I would only quote with both studies together. In a small Japanese investigation from 1994, 13 of 134 people screened had a low free T3 with a normal T4, and nine of them a zinc deficiency. After twelve months of zinc, free T3 and T3 came back into the reference range and reverse T3 fell. In a second investigation in 62 patients with chronic liver and bowel disease, low zinc likewise went along with a low T3, but giving zinc did not change these values. An honest summary: mechanistically plausible, human studies thin.

Iron comes up constantly in conversion discussions, yet it sits at a different point of the chain. Thyroid peroxidase, which attaches iodine to the carrier protein inside the thyroid, is a heme enzyme. Heme contains iron, and when iron is short the activity can fall. Iron therefore belongs more to the production of the hormone than to its conversion.

That is not hair splitting, it changes the question. In intervention studies in children with goiter in Ivory Coast, the response to a dose of iodine was weaker when iron deficiency anemia was present, and iron therapy afterwards improved that response. A systematic review from 2024 therefore recommends measuring iron status and ferritin as a standard part of the workup in hypothyroidism and especially in Hashimoto. More on this in Iron deficiency, thyroid and sleep.

Supply levels and the values to measure are covered in Selenium, zinc, iron and vitamin D, and iodine in Putting iodine in perspective.

And now you know why my answer to the selenium question starts with a question back: what is your starting point?

4. What shifts the conversion: stress, energy shortage, illness

There is a kind of exhaustion that feels different from tiredness. It does not come from a lack of sleep, and it does not go away after a holiday.

Many people know this pattern. When blood is drawn in exactly that phase, you sometimes see a low free T3 with an unremarkable TSH.

Then the quick verdict comes: stress blocks the conversion. That sentence does have a basis. It is just rarely backed up, and the study that backs it is almost fifty years old and beautifully clean.

Seen through the lens of Clinical Psychoneuroimmunology this is a typical picture. The stress system does not act at one single point, it shifts priorities. A hormone that drives basal metabolic rate is not a priority in an alarm state.

Think for a moment about what that means. T3 drives basal metabolic rate and heat production. When energy gets scarce, throttling makes sense. A body that could not do this would be worse off during a period of hunger.

People who live in a deficit for months sometimes see exactly this pattern in the lab and read it as a thyroid problem. It could simply be the answer to the deficit. More on this in Calorie deficit: why it is true and still not enough.

That the energy shortage is the point and not the method is shown by a Dutch follow up study in 66 women. Whether they lost weight by restriction or through gastric bypass made no difference. After three weeks TSH fell, reverse T3 and free T4 rose, and the rise in reverse T3 went along with the inflammatory markers interleukin 6 and 8. After three months free T4 and reverse T3 were back at baseline. Here the stress route and the inflammation route meet.

In women a second layer comes in, because the cycle axis reacts sensitively to a lack of energy. I describe that interplay in The thyroid and female hormones.

With any severe illness the thyroid values shift. The picture is called non-thyroidal illness syndrome or low T3 syndrome: low T3, low to low normal T4, TSH normal or slightly reduced. The more marked the drop in T3, the more severe the illness, as an Italian follow up observation in 45 critically ill patients showed in 1992. These numbers must not wander, though: this is about people in hospital, not about tiredness in everyday life.

Chronic low grade inflammation is the quieter relative of this situation. In autoimmune processes it runs in the background, and cytokines may act on the same enzymes. Why the immune system works against its own thyroid in Hashimoto in the first place is covered in Hashimoto and the immune system.

And now you know why, with a shifted pattern, I first ask about energy, sleep and inflammatory load and not about a supplement.

5. Liver and gut: what is possible there, and what is not

If deiodinase 1 sits mainly in liver and kidney, one question follows that is rarely asked in a thyroid conversation: how is your liver doing? Not only in the sense of liver values, but in the sense of workload.

A Berlin follow up study in 45 alcohol dependent patients makes the picture more complete and less comfortable. Before detoxification T4 and thyroxine binding globulin were low, while T3, reverse T3 and TSH were not. The strongest rise in T3 fell between day 8 and the third month of abstinence. Important for honesty: the pattern under alcohol is not identical with the classic conversion picture, and recovery takes months.

And then the gut. Many popular guides state that around twenty percent of the conversion takes place there. I went looking for that number, and in the scientific literature it cannot be substantiated. That is why it does not stand here as a fact.

What can be substantiated is qualitative and interesting anyway. Camilla Virili and Marco Centanni described in 2017 how the gut flora takes part in the enterohepatic circulation of thyroid hormones and how it influences micronutrients such as selenium. In a second paper the same authors note that around 70 percent of lymphatic tissue sits at gut level and that peripheral thyroid homeostasis could be sensitive to changes in the gut flora. That is the level of mechanism and association, not the level of percentages. More on this in Leaky gut and intestinal permeability.

And now you know why I ask about alcohol with this topic, without that being a reproach.

6. Medications that have a say, and why that is not a reason to stop anything

This is the most delicate section in the article, so the most important statement comes first.

Amiodarone is a drug for heart rhythm disorders and at the same time the finest teaching case for this topic. It inhibits the 5'-deiodinase, that is exactly the activating step, so that T3 production in the tissue can fall. The lab picture: raised T4, raised reverse T3, low T3. Many people are nevertheless euthyroid on it. On top of that comes an iodine overload, because the molecule itself contains a lot of iodine. That is why regular monitoring of thyroid function is part of treatment with amiodarone.

Beta blockers have a say as well. Wiersinga and Touber studied eleven patients with hyperthyroidism and six with hypothyroidism on levothyroxine in 1977. Under propranolol plasma T3 fell in all of them, by the same percentage, and in hypothyroidism T4 and TSH rose. The authors attribute this to an inhibition of the peripheral conversion. Beta blockers are widespread for good reasons, in hyperthyroidism even deliberately for symptom control. This finding is a puzzle piece for interpreting a lab value, not an argument against the therapy.

Glucocorticoids you know from section four. What dexamethasone did in 1977 applies in principle to therapeutic glucocorticoids too. A shifted pattern there is a known accompanying effect and not a sign that the therapy is wrong. Among the antithyroid drugs, propylthiouracil is interesting because it additionally inhibits deiodinase 1. I write that deliberately without a study reference of my own: despite a targeted search I found no cleanly verifiable primary study for it, the point rests in the deiodinase reviews.

Further reading: Levothyroxine and lingering symptoms, Graves disease and hyperthyroidism and Desiccated thyroid extract.

And now you know why the first question with a strange thyroid finding is often: what are you currently taking?

7. Why TSH can be normal while things look different elsewhere

This is the question most people come with. And the one where I have to be most careful, because a cleanly documented mechanism and an unproven transfer sit right next to each other here. So, in three steps.

Step one, the mechanism. TSH comes from the pituitary, and how much is released depends on how much active hormone arrives in the hypothalamus and pituitary. Deiodinase 2 sits there and makes its own T3 locally. Remember the on off switch: D2 is switched off through ubiquitin, and plenty of T4 speeds that up. If the hypothalamus ubiquitinates its D2 less strongly than other tissues do, it stays active under T4 for longer than the rest.

The fT3 to fT4 ratio and the thing with the 60 percent

There are calculators online that tell you where in the reference range your values sit, in percent. Alongside that it is often claimed that a good value lies above 60 percent and that a difference of more than 5 percentage points points to an impaired conversion. I would neither dismiss that nor adopt it, but put it in context.

What the ratio can do: it describes a relation, and relations are sometimes more informative than single values. The specialist literature uses it for exactly that, and the European guideline from 2012 explicitly names it as a monitoring parameter.

What it cannot do: for the widespread percentage cut offs there is no validating study, no endpoint evidence, no cohort. A cut off without validation is a thinking aid, not a diagnostic. On top of that, fT3 and fT4 vary with the time of day and with the laboratory method. Which values are worth measuring at all is covered in Thyroid blood values: which ones really count.

One note that belongs exactly here. If you take levothyroxine and your free T3 looks low, that is a reason for a conversation and not a reason to change the dose yourself. Dose adjustments belong with medical supervision and with monitoring of the values, because too much thyroid hormone carries its own risks as well, for the heart and for the bones.

On symptoms with unremarkable values only two sentences here, because there are two separate articles for that. Symptoms are real even when the lab is unremarkable, and they often have several sources at the same time. In detail in Normal values and symptoms anyway and Functional hypothyroidism.

And now you know why a ratio can be interesting without making a diagnosis.

8. The gene variant DIO2 Thr92Ala, and what the studies show

There is a gene variant that about half the world population carries in at least one copy. It affects, of all things, the gene for deiodinase 2. And the studies on it contradict each other. So I will tell them in the order in which they appeared.

2009 started the debate. Panicker and colleagues examined variants in all three deiodinase genes in 552 people on levothyroxine. The rarer CC genotype of rs225014 occurred in 16 percent and went along with worse scores on a questionnaire about psychological wellbeing. The same carriers improved more on the combination of T4 and T3 than on T4 alone. The variant had no influence on hormone levels in the blood. The authors themselves wrote that the results require confirmation.

2017 brought the enzymatic confirmation. Castagna and colleagues compared 140 patients who had thyroid surgery, before and after the operation. Carriers of the altered variant had significantly lower free T3 values afterwards. Cell experiments showed that the altered enzyme differs in protein stability and converts less T4 into T3.

The best explanation for the contradiction comes from the mouse lab. Gabriel de Almeida and colleagues compared two genetically distant mouse strains in 2024, each with and without the Ala92 variant. In one strain carriers and non carriers behaved the same. In the other, carriers showed higher cholesterol and a fatty liver, and at an ambient temperature of 30 degrees most of that disappeared again. A follow up paper in 2025 additionally found a goiter in the same strain. These are animal studies in mice, so a transfer to humans is not permissible.

Bianco and Kim summarized the state of the debate in 2018, from the pen of the group that has published most on it. Thr92Ala is, they write, the best studied deiodinase polymorphism. In the same text stands the decisive sentence: these associations have not been reproduced in all population studies.

What the guidelines say, and why that is understandable

Guideline endocrinology is cautious here, and it has good reasons for that.

The European guideline from 2012 notes that in 5 to 10 percent of people treated with levothyroxine who have a normal TSH, symptoms persist. Among the explanations it names accompanying autoimmune diseases and the inability of monotherapy to produce physiological concentrations in serum and tissues. So the conversion question is explicitly in there, as one of several explanations. Because the evidence for a superiority of the combination is not sufficient, monotherapy remains the standard.

The American guideline from 2014 comes to the same conclusion and itself calls for better biomarkers of the euthyroid state to supplement TSH measurement. That is an open acknowledgement that TSH alone does not answer the question.

This is exactly where the two views meet. The guideline asks: is it proven that a change of therapy helps on average? A functional view asks: which levers could influence the conversion in this individual person? Both questions are legitimate, and they need different levels of evidence.

That the concern is real is shown by a 2018 survey of 12,146 people: about 15 percent more of those on levothyroxine reported a reduced quality of life than control persons. An online survey run through patient organisations is strongly selective, though, so this is a mood picture and not evidence of efficacy.

What follows from this in practice

Now the part you have probably been waiting for. As a direction, not as a recipe. An article cannot replace a history taking, and everything concrete belongs in a medical conversation.

One last thought that goes beyond the biochemistry. The thyroid sits at the neck, where breath and voice pass through. Anthroposophic medicine traditionally sees in this a place where inside and outside are mediated. That is a way of looking from clinical tradition and not a study finding, and I mark it as such explicitly. More on this in The thyroid seen anthroposophically. It asks the same question that the physiology asks here: what happens between arriving and taking effect?

And now you know why I would rather talk about conditions than about supplements with this topic.

Frequently asked questions about T4 to T3 conversion

Where this topic connects

The conversion does not stand on its own. It hangs on energy, on cofactors, on stress and on what happens in the gut. Several paths lead onward from here.