Question Pilo’s Study: Thyroid Hormone Reductionism

Question-Pilo-theoryMany aspects of biology are incredibly complex. People love simplicity. It’s necessary for a model to reduce a system down to its basic principles and proportions.

Pilo’s research study published in 1990 is now cited as the definitive proof of where human beings’ T3 thyroid hormone supply comes from.

Pilo’s paper is what many people point to when they say that “the” thyroid secretes T4 and T3 at a molar ratio of 14:1.

Pilo is often who they cite when they claim that 20% of our T3 comes from the thyroid gland, and 80% of our T3 comes from peripheral conversion of T4.   

This suffers from the scientific pitfalls of reductivism.

This article will examine the complexity and theoretical nature of Pilo et al’s 1990 study and demonstrate that it, and other studies like it, are not designed to establish laws for thyroidal production and secretion. They are only designed to establish methods for understanding thyroid hormone secretion and conversion as an overall process.


  • Scientists make models at the risk of reductionism.  At the micro-level, reductionism reduces a complex whole to each tiny part. At the macro-level, it oversimplifies the complexity.
  • Pilo et al wrote a theoretical paper. They proposed a complex methodology. They used human data as raw material to test theoretical and mathematical models of thyroid hormone circulation. They knew they had to make assumptions.
  • Other leading thyroid scientists have noticed Pilo’s theoretical abstraction and complexity. They preferred not to cite Pilo’s article for its ratios.
  • Others were just as reductionist. They still choose to give average ratios based on their own research studies on even smaller groups.
  • Why do these articles get canonized when others emphasize thyroid hormone fluctuations and ranges more than ratios?   
  • Why are we forgetting that treated thyroid patients can’t fit within this theoretical model, even if their medication tries to make some numbers fit?


In an essay on scientific reductionism by Martyn Shuttleworth, he explains reductionism.

He starts by saying “One form of scientific reductionism follows the belief that every single process in nature can be broken down into its constituent parts and can be described scientifically.”

This is just one extreme of scientific reductionism, the desire to analyze every tiny detail of a complicated process.

The opposite extreme within the continuum of scientific reductionism is to “reduce” complexity by choosing one or two components of a complex model that appear to have an overarching or controlling role within it.

Shuttleworth explains this second type of reductionism as an oversimplified model. “One area that uses reductionism extensively is computer modeling. For example, if a scientist designs a computer program to model and predict weather patterns, they cannot possibly include every single permutation of such a vast and complicated system. Instead, they simplify many of the elements to allow the program to work without losing the accuracy.

This is related to ‘Black Box’ science, where part of a system is regarded as a box. The scientist knows that any data inputted into the box results in a certain output, so they do not need to model every last process within the box.”

So one end of the spectrum of reductionism generates endless complexity and “reduces” things down to every little separate detail, and the other end of it oversimplifies and “reduces” it to a machine with only one or two knobs.

Either way you do it, reductionism ends up with a similar problem — it can overlook and underestimate key details, and it sacrifices a larger perspective of the whole.  

It is best explained by the metaphor of the group of blinded scientists each touching a different part of the elephant. The scientist touching the tail says “it’s a rope.” The one touching the ear says “It’s a fan.” The tusk only, “It’s a spear.” The leg, “It’s a tree.”

Thyroid hormone synthesis, secretion and conversion is not only complex process in itself, but because it is needed and used everywhere in the human body, it interacts in complex ways with the health of each organ and with environmental influences like diet, stress, exercise, toxins and even ambient temperature.

The thyroid hormone system is the perfect storm of complexity. Given this complexity, the human temptation is to oversimplify thyroid hormones and reduce them to TSH and T4.

Shuttleworth says, “Chaotic systems, such as turbulence, weather patterns and even the behavior of crowds are difficult to explain by the process of scientific reductionism.”

Pilo’s 1990 study suffers from both types of reductionism outlined by Shuttleworth, but those who apply Pilo’s study use it to reduce thyroid hormone conversion to two simple averages.


Pilo and his team really valued the mathematics. It took them four and a half pages to describe the various aspects of computational wizardry. They strongly emphasized the theoretical arguments that informed their mathematics.

As you can see in their complex visual models and their explanations around their math, they were theorizing the various “pools” of blood or plasma that get transferred between “compartments,” or tissues and organs within the human body.  


Pilo and team’s mathematical formulas are well beyond most people’s understanding, but we don’t need to explain them to understand what is going on.

Even a non-mathematics expert can pick up a lot of the logic of their grammar and the emphasis of their argument.

Very little was certain: They continually expressed reliance on prior assumptions, theoretical models, estimates, and ranges of error for each aspect of calculation.

Though they tried to be precise, the size and complexity of the model was so overwhelming that it required simplifications at the level of details.


Their language reveals this stress.

Notice how they briefly refer to “flux” as they aim for mathematical simplicity by approximating a “steady state”: “From the flux balance in the endogenous steady state the following equations can be written…”  

In another section, their discussion of “fractional rate constants” of production and conversion imply a model that assumes a rate is constant, and yet we know that hormones are not secreted or converted at a constant rate. Imagining a constant rate provides a mathematical convenience.

Looking at Pilo et al’s reference list, one sees that they borrow a lot of their assumptions and models from 6 articles published in the 1980s by one researcher named Di Stefano and his colleagues, which appear to be rat models. 

They are in an intellectual discussion with someone focusing on rats, because at this level, the method is more important to them than the fact that it’s rats or humans being studied.

Are those still considered reliable studies, and to what degree are their models applicable between the rat and the human?  

What can an academic analyst of this article learn from this section of the article?

It’s not about these 14 people’s thyroid hormone levels.


The 14 research participants and their actual rates of secretion and conversion are not the main focus, and generalizing from these 14 to a larger population is not the main point.

Overall, much of the article seems to be engaging in a theoretical argument about their complex method. They are refining a way of modeling and calculating.

The 14 participants seem to have provided raw materials: blood samples and bodies through which a research team could demonstrate their mathematical wizardry.

To a large degree, this article is not about the humans. They are a collective group of lab rats, not unique individuals.

A lot of these scientists’ interest is in the technologies and test tubes, the theories and assumptions. We can imagine them having excited theoretical arguments about formulas on the chalkboard.

How does this make their findings become trustworthy biological facts?


This Pilo study was cited in a very limited way by a ground-breaking, comprehensive article on thyroid hormone conversion and action by Bianco and colleagues in 2002.

So Bianco and team noticed Pilo’s study, and yes, it also discusses the rate of T3 secretion and conversion and reduces them to average ratios.

Ah, but HOW was Pilo et al cited in their article?

Well, this is embarrassing.

Bianco and team rather dismissed Pilo’s article as one that discussed a narrow corner of a theoretical and hypothetical debate — the issue of where in the body most T4 is converted into T3.

Pilo’s study, they say, points out the “difficulties” of defining “which compartments” do T4-T3 conversion, because of the various “assumptions” that are possible in the theoretical models.

“Depending on the assumptions used,” say Bianco and team about Pilo’s article, “one can obtain estimates suggesting that as much as 81% or as little as 15% of T3 derives from rapidly equilibrating (D1-containing) tissues, with the remaining coming from slowly equilibrating (D2-containing) compartment.”

Basically, Bianco and team find Pilo far too theoretical as a basis for a ratio.

But they still give ratios.

As a basis for their own ratios, they cite mostly articles and chapters written by one of their own co-authors, P. Reed Larsen.

They don’t cite Pilo when they write about ratios. They give the ratio of T4/T3 found in thyroid tissue, not secretion, as 15:1, citing Izumi and Larsen, 1977. This study only measured gland samples from 6 healthy patients.

They don’t cite Pilo when they then give a ratio of secretion. They say it’s as low as 11:1. This ratio is from Larsen in 1975, based on thyroid tissue and blood from 11 healthy patients.

They don’t cite Pilo when they give a ratio of T4-T3 conversion. They rely a lot on one of the hugest textbook chapters on thyroid disease, again co-authored by Larsen, in 1998, and a 1996 article again co-authored by their team member, Larsen.

They pick their own favorite ratios based on their own averages from small populations.


Laurberg gravely cautioned us about the limitations of “kinetic” studies such as Pilo’s. Basically, he pointed out its reductionism.

He cautioned us not to rely too much on estimated rates and ratios based on averages among small populations.

Read this carefully:

“Kinetic studies are complicated,” he wrote, “and caution should be exercised when trying to develop models from mean values obtained from relatively few subjects in such studies.” (p. 388).

Why are we trusting such kinetic studies’ averages as “the” definitive experiments that discovered “the” T4:T3 ratios of secretion in “the” human body?

These are just mean (average) values.

They are the averages within a wide range of human variation.

They arise from relatively few people they studied.

The averages are meant to express an overall model in a complex system that is far wider in range and scope than those averages.

Don’t forget that most treated thyroid patients do not have the full ability to adapt within that full model, because they don’t have a fully functional thyroid gland. 

Don’t forget that treated thyroid patients are taking drugs that directly manipulate and distort their rates and ratios of thyroid hormones.

By trying to nail down our thyroid hormone ratios and rates to something static, even by trying to target an average found in a segment of a healthy population, medicine can keep some thyroid patients trapped in a state of disease. 

– Tania S. Smith, with input from Linda Sanday


Bianco, A. C., Salvatore, D., Gereben, B., Berry, M. J., & Larsen, P. R. (2002). Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases. Endocrine Reviews, 23(1), 38–89.

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.

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.

Shuttleworth, M. (2008, April 15). Scientific Reductionism – Reducing Complex Interactions in Research. Retrieved May 25, 2019, from

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s