Question Pilo’s study: The wide range of thyroid hormone adaptation

Question-Pilo-14Doctors today are taught that “the” thyroid secretes T4 and T3 at a ratio of 14:1, or 15:1, and “the” human body converts 80% of its T3 supply from T4.

These ratios are now taught and passed on as if they are static, immutable laws of nature, like the law of gravity.

However, studies of complex thyroid hormone kinetics are based on small populations, like Pilo’s study of 14 people.  

The 14 people in Pilo’s 1990 study were diverse. They provided a range of T4 and T3 values and ratios.

This group of 14 people was also unique. They can’t represent all normal people, much less people with thyroid disease.

Therefore, they can’t provide anything like an immutable law about “the” T4 and T3 hormone secretion of a thyroid or “the” human body’s conversion rate of T4 to T3.

All they can provide is an experimental model.

This article was never intended to provide anything like a law or general principle, such as laws of physics, where we can be certain of the velocity of an object as it drops from a height.


  • We can gain valuable insight into thyroid hormone diversity between people by looking at the tables of data in Pilo’s 1990 study. Among these 14 people are a wide ranges of values and ratios. 
  • Pilo’s article never claims these 14 people are representative of the normal population. It doesn’t provide enough information to assess how representative or healthy they are in their hormone levels.
  • Pilo and his team do not explain or emphasize this human thyroid hormone diversity in 1990.
  • Other studies during the era examined the wide range of T4-T3 production and conversion. They also studied T3 loss in T4-treated thyroidless people. They learned a lot by putting thyroid health in the context of adaptation to thyroid disease.
  • Newer studies are renewing the focus on the wide range of T3 and T4 in “normal” people and in treated thyroid patients, especially the loss of T3 within T4 monotherapy.


Pilo et al, 1990 engaged in a study of only 14 “healthy” adults (9 men and 5 women) aged 19-65.

These fourteen people and the way their data have been interpreted are now the biased, partial foundation for the therapy of millions of patients.

The researchers measured their T3 and T4 levels using radioactive tracer procedures (they injected them with tiny amounts of T4 and T3 tagged with radioactive iodine atoms).

They measured these radioactive isotopes of hormones in blood samples using gel chromatography. Organic chemists still use these methods, but we no longer use them to measure thyroid hormones.  


Pilo and team present us with baseline data at the start of the study, a table of plain and simple lab test results.


Because they studied so few people, they are able to give us a row of data for each of the 14 subjects.

Therefore, by just scanning the columns, we can see the full range of variation for T4, T3, Free T4, Free T3, and TSH and how each individual represents a unique constellation of values.

Notice the range in numbers:

Their Free T4 ranged from 6.9 to 13.1 pg/mL, [which converts to 8.88 to 16.86 pmol/L].

Their Free T3 ranged from 2.6 to  5.7 pg/mL, [which converts 4.00 to 7.7 pmol/L].

Their Free T3:T4 ratio in blood ranged from 1.84 to 3.46 pg/mL, or 0.38 to 0.65 pmol/L.

You can see a wide range of human variation in their blood levels.

But how wide were their ranges in comparison to normal healthy people? How representative were these 14 people?  


No reference ranges were given in Pilo’s article, so we can’t even begin to answer the question about how representative they are.

(NOTE: We tried very hard to find one. The authors cite methods in a 1987 article, which then refers you to see the methods section in a 1983 article. Back then they didn’t test Free hormones, just the Total T4 and T3, for which they gave a reference range, but its basis was highly questionable. For each of 68 so-called “healthy” patients — whose characteristics were not described — they tested 6 different blood samples on 5 different manufacturers’ lab test kits and then pooled their data. This was not necessarily a valid method to develop a reference range. It’s hard to know whether technologies and assays changed over 7 years).

Of course, a lot of controversy surrounds the use of reference ranges: Are we calculating them on a mathematical model that is accurate and useful to therapy? Are they based on blood samples of people that are truly free of diseases and medication influences on their hormone levels? We can’t begin to go into the controversy about how mechanistically these ranges are now applied to diagnostic judgment.

But regardless of these limitations of reference ranges, they can help us understand how a person or a small group relates to a larger, supposedly healthy control population.

Here, we have no idea how widely this 1990 study’s participants’ T3, T4 and T3:T4 ratios ranged in relation to a larger control population.

Nor do we know whether these numbers were shifted lower or higher than such a control.

Not addressing this could lead people to draw incorrect and potentially harmful assumptions from this study.

For example, it was a well understood principle in 1990 that the higher T4 is, the less will convert to T3, and the lower T4 is the more will convert to T3. Pilo’s own research team knew this already in 1983.

Based on this principle alone, if these 14 people are not representative of the full range of T4 levels across a healthy population, then their rate of T3 conversion could be skewed high (if T4 was low on average), or T3 conversion rates could be skewed low (if T4 was higher on average).


Even though the study didn’t focus on TSH, it has great value in helping us assess how representative this untreated population of 14 people might have been.

In an untreated population with healthy thyroid tissue, TSH plays a regulatory role over T4 and T3 secretion.

Pilo et al’s 14 patients had a TSH ranging from 1 to 2.

Today’s TSH reference ranges are approximately 0.3 to 4.0 mU/L, which means their patients covered about 21% of the spread of many TSH reference ranges today.

In the early 2000s there was a heated debate over where the upper limit of the TSH reference range should be set. This was a controversy even more heated than the T4:T3 ratios we are now examining. That debate arbitrarily settled on a compromise based on the need to manage population health: they needed to avoid overdiagnosing and overtreating hypothyroidism, so they compromised on a range that cuts off around 4.0-4.5 instead of 2.5, even though the vast majority of healthy people maintain a TSH below 2.5.

Even if you take the most narrow TSH range proposed, no higher than 2.5 mU/L, Pilo’s tiny population’s TSH range was about 50% narrower than that.

Here’s how it can skew the results.

Even in 1990, it was well known that higher amounts of TSH stimulation on the thyroid gland raises its ratio of T3 output in relation to its T4 secretion.

Would the amount of T3 secreted be far higher if the sample included people with what we now consider a “normal” TSH near 4.0, like it is in people who are recovering from illness, and people whose gland is mildly failing but is going to be perfectly fine with an extra kick of stimulation? Yes, of course their healthy T3 would be higher than the norm.

Would the amount of T3 be significantly reduced in people whose TSH was lowered by fasting or calorie restriction, exhausting exercise, or hot climates? Yes, of course their healthy T3 would be much lower than the norm.


Moving away from simple blood test results, let’s look at what Pilo and his team arrived at in his results regarding T4-T3 conversion.

The human variation in the conversion ratios of T4 into T3 were also wide.

Pilo et al’s Table 4 gives different calculations based on three models, not four, not two, but a nice, tidy three. It appears we are supposed to be most impressed by the complexity of their six-compartment model, which happens to yield the highest mean.


In the complex 6-compartment model, the lowest conversion rate in one patient is 16.9, and the highest rate in another patient is 42.9.

That means in Person 1, 42% of their secreted T4 got converted to T3.

That means in Person 2, only 17% of their secreted T4 got converted to T3.

That’s a large variation from person to person.

Which test subject was healthier?

Why should we believe the mathematical average of conversion rate, 27.3, is the most healthy state for the human body at the low end of the range (16.9), and for the human body at the high end of the range (42.9)?

The lesson to learn from the data is human variation and diversity.

Thyroid hormone conversion varies from person to person, and probably from time to time in their lives. There is no static ratio.

How could anyone know whether the data from these 14 people applies in theory to all healthy human beings? What can this study say about “the” human thyroid gland?


Why did this article get canonized when other articles in the era emphasized thyroid hormone diversity and ranges more than ratios?

Other studies during the era showed the T3 hormone adaptation to various influences on the thyroid hormone system.

Even some of Pilo’s own research team members did prior studies that featured human thyroid hormone fluctuation and diversity.

In 1978, when Bianchi et al studied “triiodothyronine (T3) kinetics in normal subjects, in hypothyroid, and hyperthyroid patients,” they said that “a constant ratio between the turnover rate and the total hormone concentration (i.e. by definition MCR) was not found in our subjects.” They didn’t find a constant ratio, but rather, “the fraction of free T3 increases as the total thyroid hormones plasma concentration goes up.”

In 1983, when Bianchi expanded their mathematical methods for estimating T3 thyroid hormone kinetics, they found that conversion rates varied widely between healthy controls and “sick euthyroid” people. However, they also learned that treated hypothyroid patients on T4 monotherapy didn’t gain much of an increase to their conversion rate to compensate for their lost thyroid gland.

This 1983 study also confirmed the principle that the more T4 a person has, the less T3 the person gets out of it, throughout the entire wider range of normal, sick euthyroid, and T4-treated thyroidless people.

Then, in 1987, when Bianchi et al studied the effect of diseases in thyroid hormone binding, they found that thyroid hormone production and peripheral T4-T3 conversion made extreme adaptations to try to compensate for these diseases.

The thyroid hormone system is essentially an adaptive system, not a static one. It enables “homeostasis,” not “stasis.”

As Hoermann and colleagues have explained as recently as 2016, there is variation in “homeostatic” control among the hormones across the entire euthyroid range in response to early and progressive changes in the system. In the event of imminent thyroid gland failure causing a reduced secretion of FT4 per unit of TSH, the efficiency of conversion of T4 into T3 increases in the attempt to maintain FT3 stability. In such situations, T3 stability takes priority over the maintenance of the prior set point.

Back in the decade before Pilo’s 1990 study, Laurberg’s 1984 article examined this homeostasis. He explored the question of why the human body increases its T3 production and T3 conversion in untreated hypothyroidism.

Laurberg found that TSH stimulation as well as lower T4 levels were both involved in ensuring that enough T3 gets into peripheral tissues. This is a natural adaptation that keeps a person with a failing thyroid gland from experiencing hypothyroid symptoms, until their body is unable to adapt to too great a loss of T4.

However, Laurberg noted that people adapt poorly to the artificial situation of T4 monotherapy. Adding T4 and lowering TSH can’t reset thyroid hormone conversion for the treated thyroid patient.

Laurberg emphasized their wide range of T3 conversion, and mainly T3 loss. These are “patients receiving replacement or suppression therapy with T4 alone in whom serum T3 is 65%85% of the values expected from the serum T4/T3 ratio in normal subjects.”  


We see a wide range of T3 secretion and T3 from conversion in Pilo’s 14 people, but even they only represent a small slice of humanity.

In thyroid hormones, the range of natural adaptation is wide. There is no such thing as a static conversion rate or secretion rate. It can’t exist or humans would not survive. 

A lot of people were interested in these natural adaptations, at least before 1990.  

Why didn’t the ranges of function, the diversity of human adaptation, get noticed in Pilo’s 1990 article?

They were not focusing on them anymore. Why? Because T4 monotherapy was being promoted.

Once people’s minds became settled on “the” T4:T3 secretion rate and conversion rate in “the” human body, the diversity that remained was TSH. 

Now, it appeared to be more important to study diverse TSH and normalize people’s abnormal TSH. 

This is very similar to the way medicine has focused too much on high cholesterol and on lowering cholesterol levels, instead of looking at the complex systems that raise cholesterol. High cholesterol is not necessarily a bad thing, except in heart disease. Neither is a mildly higher TSH necessarily a bad thing, except one can be sure it is in thyroid disease and therapy.

As a result of this historical bias, this wide-ranging T3 deficit in T4-treated patients has not a concern for decades. Even the severe T3 loss in non-thyroidal illness has been treated as benign, because their TSH is normal so they can’t be classified as hypothyroid.

However, since about 2011, more and more studies have focused on T3 loss, especially this T3 inequity in T4-treated thyroid patients, which Laurberg noted in 1984.

Contemporary studies by Hoermann, Midgley, Larisch and Dietrich have continued to study the questions Pilo, Bianchi and their colleagues raised throughout their careers.

They have even updated the complex mathematical models that Pilo and colleages were working on. 

Why don’t we look at those new studies?

We need to look at the wide ranges of thyroid hormone adaptation in both health and disease, not just the average ratios.

– Tania S. Smith, with input from Linda Sanday


Bianchi, R., Mariani, G., Molea, N., Vitek, F., Cazzuola, F., Carpi, A., … Toni, M. G. (1983). Peripheral metabolism of thyroid hormones in man. I. Direct measurement of the conversion rate of thyroxine to 3,5,3’-triiodothyronine (T3) and determination of the peripheral and thyroidal production of T3. The Journal of Clinical Endocrinology and Metabolism, 56(6), 1152–1163.

Bianchi, R., Iervasi, G., Pilo, A., Vitek, F., Ferdeghini, M., Cazzuola, F., & Giraudi, G. (1987). Role of serum carrier proteins in the peripheral metabolism and tissue distribution of thyroid hormones in familial dysalbuminemic hyperthyroxinemia and congenital elevation of thyroxine-binding globulin. The Journal of Clinical Investigation, 80(2), 522–534.

Bianchi, R., Zucchelli, G. C., Giannessi, D., Pilo, A., Mariani, G., Carpi, A., & Toni, M. G. (1978). Evaluation of triiodothyronine (T3) kinetics in normal subjects, in hypothyroid, and hyperthyroid patients using specific antiserum for the determination of labeled T3 in plasma. The Journal of Clinical Endocrinology and Metabolism, 46(2), 203–214.

Hoermann, R., Midgley, J. E. M., Larisch, R., & Dietrich, J. W. (2016). Relational Stability in the Expression of Normality, Variation, and Control of Thyroid Function. Frontiers in Endocrinology, 7.

Laurberg, P. (1984). Mechanisms governing the relative proportions of thyroxine and 3,5,3’-triiodothyronine in thyroid secretion. Metabolism: Clinical and Experimental, 33(4), 379–392.

Pilo, A., Iervasi, G., Vitek, F., Ferdeghini, M., Cazzuola, F., & Bianchi, R. (1990). Thyroidal and peripheral production of 3,5,3’-triiodothyronine in humans by multicompartmental analysis. The American Journal of Physiology, 258(4 Pt 1), E715-726.

5 thoughts on “Question Pilo’s study: The wide range of thyroid hormone adaptation

Leave a public reply here, on our website.

This site uses Akismet to reduce spam. Learn how your comment data is processed.