Targeting TSH? It’s not as specific or as sensitive as people claim.

Patients and doctors are often told there’s an ideal “target” for TSH levels. Some believe that in thyroid therapy, your TSH should be below the upper reference limit, some say it should be below 2.5, and a minority will even say it should be suppressed entirely whether or not you are a thyroid cancer patient.

Some will say, with more wisdom and discretion, that a good TSH level also depends on how much thyroid tissue you have, or what type of thyroid therapy you are on.

Patients often say in sympathy with a fellow patient who is suffering with a TSH of 9.0 and being refused thyroid therapy because it was not yet 10.0, “That’s awful!  I felt horrible when my TSH was 6.6.” That’s another statement that shows that people believe that TSH levels to some degree correlate with their overall health status. Is it reasonable to believe that your TSH makes you feel thyroid symptoms more than T4 or T3 can?

In general, medical education and policy promotes the practice of treating the TSH as a health outcome in thyroid therapy by making TSH the primary “target” or endpoint.

A doctor administers thyroid therapy, TSH is normalized, and this is supposed to mean the therapy is successful, without concern for where FT3 is within the context of FT4.

Patients and doctors often hear the mantra repeated, in various words:

“TSH is the most specific and sensitive indicator of thyroid hormone sufficiency.”

Notice the concepts of “specificity” and “sensitivity.”

This is special terminology that is used to validate a diagnostic test. If we want to understand whether this claim is true or false, we have to understand this terminology AND the science that agrees or disagrees with the claim.

• First, I’ll lay out these two criteria (specificity and sensitivity) of a good diagnostic test. 
• Then I’ll outline three ways in which science proves that the TSH is not very likely to meet these criteria.
• Finally, I’ll go over some circumstances in which TSH might actually be a useful indicator, in context.

SPECIFICITY

Specificity is a negative idea. You can trust the test to rule out a diagnosis of a disease:

“The ability of a test to correctly classify an individual as disease- free is called the test′s specificity.”
(Parikh et al, 2008)

A good trick to remember this is to think about being protected by your “special” status. Being in the “normal” or healthy range of this test result means that you can confidently rule out a specific disease / disorder.  

For example, during thyroid therapy, if your “TSH = normal” result correctly classifies you as free of hypothyroidism, then none of your tissues should be in a hypothyroid or thyrotoxic state.

If TSH is not specific enough, when you are truly hypothyroid the TSH will stay normal and won’t rise above reference range to indicate your hypothyroidism.

SENSITIVITY

In contrast, “sensitivity” is a positive idea. You can trust the test to identify who has a disease:

“Sensitivity is the ability of a test to correctly classify an individual as ′diseased.′” (Parikh et al, 2008).

As a mnemonic, think of a sensitive “sensor” like a good smoke detector. If the smoke detector has good “sensitivity,” it will ring when there’s really smoke from a fire.

If your “TSH = low” result incorrectly classifies you as “thyrotoxic” (disease / disorder) when you are actually euthyroid, that’s a problem. It’s like a smoke detector ringing even when your shower is making too much steam!

JUDGING SPECIFICITY AND SENSITIVITY

The word “sensitive” can be confusing because of its other meanings in English.

If the TSH is “hyper-sensitive” in the sense that it is oversensitive to other chemicals or hormones and TSH falls below range for reasons other than thyrotoxicosis, that’s a failure of sensitivity; the low TSH is incorrectly classifying you as diseased if you are truly euthyroid.

If the TSH is dull or “insensitive” and can’t rise when hypothyroidism exists, it means that it cannot achieve a result that assesses its failure to achieve “sensitivity” as a test. It mainly fails the “specificity” test, because the results were actually “normal” but your normal TSH failed to indicate any special protected status — it didn’t rule out your true hypothyroidism.

In addition, a person has to be fair and reasonable, not an absolute perfectionist, when judging a test that measures disease status indirectly. There is no test that is 100% specific or sensitive without being synonymous with the health outcome itself.

Therefore, the ultimate questions for thyroid patients and doctors are questions of degree:

• “To what degree is a normalized TSH, as it is often used currently in therapy as a test on its own, SPECIFIC enough to RULE OUT tissue hypothyroidism in the treated thyroid patient?”
• “To what degree is a TSH below reference range SENSITIVE enough to correctly classify us as DISEASED, i.e. “thyrotoxic” or overtreated with thyroid hormone?

Now, here’s some of what the science says.

1) TSH IS LOCALIZED AND BIASED

In 2019, Dr. Bianco and colleagues reviewed research about molecular action of thyroid hormones.

They admitted that the TSH is a “specific” and “sensitive” indicator only of the hypothalamus and pituitary glands’ local tissue T3 sufficiency, insufficiency, or excess.

First I’ll let Bianco and team speak about the hypothalamus, and then I’ll explain it. 

He says,

“The idea that circulating T4 is detected by the hypothalamus and the pituitary gland via the D2 pathway [Deiodinase type 2 — which converts T4 to T3] has been challenged over the years.” …

“D2 activity accelerates under low serum T4 conditions, whereas high serum T4 levels result in loss of D2 activity.”

However, higher serum T4 does not result in loss of D2 activity in the hypothalamus.

“As a result, the hypothalamus remains exquisitely sensitive to elevations or drops of circulating T4, in contrast to what is observed in other tissues”
(Bianco et al, 2019, p. 1016).

To summarize what he explained,

As bloodstream T4 rises higher in its reference range, most tissues in the body progressively slow down their T4-T3 conversion.

(The rest of the body’s conversion rate slows down because the abundance of T4 inactivates Deiodinase Type 2 (D2) — an inactivation process called ubiquitination.)

But meanwhile, the “T4 superconverer” glands — the hypothalamus and pituitary — keep on converting T4 efficiently.

These two tissues convert T4 into T3 at their own unique rate. 

The hypothalamus is the boss; its secretion of TRH hormone tells the pituitary how much TSH to secrete. Pituitary TSH is a response to local tissue T3 supply activating receptors within the tissues of the hypothalamus. 

As a result, the TSH secretion of the pituitary signal that its own tissues are euthyroid (normal). It could care less if other tissues in the body are hypothyroid.

Bianco is clearly identifying an inconsistency between T3 sufficiency in the hypothalamus + pituitary compared to other tissues.

Let’s put this into perspective with hypothetical TSH numbers:

• Your pituitary TSH of 2.5 may tell the doctor “I’m euthyroid!”
• But the rest of your body says clearly, through its circulating Free T3 + T4 and / or your symptoms and health outcomes, “I’m hypothyroid.”

(In this case, the TSH test fails to be specific; you are unprotected from tissue hypothyroidism by a normalized TSH.)

• Alternatively, your TSH of 0.03 may say “I’m thyrotoxic! Stop producing, Mr. Thyroid!”
• But the rest of the body is saying “We disagree, Mr. Pituitary and Sir Hypothalamus!  We’re euthyroid because we’re not getting as much T3 out of our T4 supply.”

(In this case, the TSH test fails to be appropriately sensitive to the body’s supply; it is sensing localized T3 excess at its receptors while your body senses appropriate levels of T3 reaching its receptors.)

The degree to which the TSH is biased too low while on thyroid medication will vary from person to person, but it will likely be biased low, not biased in the other direction.

TSH is biased lower especially in T4 monotherapy because bloodstream T4 is significantly higher in these patients than in the healthy-thyroid population (Gullo et al, 2011).

2) TSH IS NOT SO SPECIFIC AFTER ALL

Next, TSH is not specific as an indicator because it is influenced and co-regulated by more than just thyroid hormone negative feedback.

TSH is hyper-sensitive to health factors beyond thyroid hormones, which makes it vulnerable to hypo-secretion.

If TSH hypo-secretes (lowered TSH), it means TSH cannot rise to fulfill its role as a “sensitive” indicator of hypothyroidism, nor is a normal TSH “specific” enough to rule out hypothyroidism. 

“Primary hypothyroidism” (thyroid failure) is usually defined in a way that separates it from from “central hypothyroidism” (TSH hypo-secretion). 

However, people with primary hypo are not protected from central hypo — either severe or mild, temporary or permanent, chemically-induced or caused by other health conditions (Beck-Peccoz et al, 2017).

The thyroid is one gland. The hypothalamus is another, and the pituitary a third.

Nothing in biology says you can’t have relative degrees of failure in two or three organs at the same time. (See “Screening for central hypothyroidism during thyroid therapy“)

As for the nature of their failure, it is hard for many doctors to admit without your having physical damage to these glands, perhaps from a motor vehicle accident. This is why central hypothyroidism is so rarely diagnosed.

Milder forms of hypothalamus and pituitary failure occur when the TSH is suppressed just enough to fail as a diagnostic tool.

These organs can fail because their natural roles in the human body are very different from being a monitor of thyroid status.

Let’s think about the natural job description of these two central glands, hypothalamus and pituitary.

The hypothalamus and pituitary process data about thyroid hormone supply (and the need for future hormone supply) in a way that no other set of organs do. This is because of their role in regulating a living, flexible thyroid gland.

Nature designed TSH as a feed-forward impulse. This means that TSH secretion anticipates the needs of the body in the near future.

It is not just a closed feedback loop in which TSH takes orders from thyroid hormones. It’s an open loop that takes orders from many hormonal and environmental cues about trends and future dangers.

If the hypothalamus and pituitary were not also specific or sensitive to additional factors that say whether more T3 and T4 are needed tomorrow or next month, they would be unable to fulfill their physiological role as co-regulators of another gland (the thyroid). 

There’s a lot of complex input coming into these glands. Many substances and factors such as leptin, TSHR thyroid antibodies, rexinoids, glucocorticoids, can co-regulate the TSH and suppress its secretion. 

It’s no surprise if wires can get crossed. Some inputs can overwhelm others. 

According to Bianco and colleagues’ review, at the molecular level, TSH secretion is a response not only to thyroid hormone, but also a multitude of other health and environmental factors that can reduce D2 expression and change thyroid hormone receptor sensitivity in the hypothalamus and pituitary (Bianco, 2019).

(NOTE: Bianco and team know well and teaches the principles that lead to a falsely normalized or falsely low TSH in many states of health. However, they have not yet brought themselves to say anything directly against the TSH test and the use of TSH normalization as therapeutic target in treated hypothyroidism. This is unfortunate both for the sake of scientific integrity and for the many patients whose therapy they could improve by expressing the logical conclusions of such a scientific review.)

Illnesses can alter the “normal” TSH-FT4-FT3 relationships in many directions.

If you want to see a list of health conditions in which TSH secretion seems to be confused or clueless about the state of thyroid hormones flowing through the bloodstream and supplying the rest of the body, take a peek at this table in Chatzitomaris et al, 2017. 

You’ll see the most consistent positive or negative columns (red or blue) associated with “diseases” are FT3 and TT3. 

When your body is diseased, T3 levels are often too high or too low. 

[Table reproduced within Copyright “fair use / fair dealing” for the purposes of criticism or review.]

Anyone looking at this table would say that it’s mainly T3 (sometimes amplified by T4) that serves as a “sensitive” indicator of disease. 

When any of these diseases or health conditions strike, the doctors will say “No — you can’t possibly be hypothyroid because your TSH is not elevated.”

They are simply looking at the wrong hormone.

They are not judging your thyroid status correctly by your T3 (+T3 via T4 conversion) levels. It’s the loss of T3 that makes any tissue in your body hypothyroid. T4’s incapacity to signal at the nuclear receptor and mitochondria means that FT4 being high can never counterbalance a too-low FT3.

Notice that some of the conditions listed in Chatzitomaris and team’s table are not diseases, but benign states like “caloric deprivation” and “exhausting exercise.” Yes, dieting and exercise, if overdone, can make things worse for suffering hypothyroid patients on therapy.

If doctors tell the obese and fatigued thyroid patient to “eat less and exercise more,” and they take their doctor’s advice by fasting and exercising too vigorously, then their TSH might not be capable of rising to reveal how hypothyroid they have become.

Look at the table again to see this interesting contradiction:

• How many times is TSH inappropriately elevated when FT3 is normal or elevated? In three of the “Type 2 Allostasis” conditions. That’s because in Type 2 states, TSH is not low enough, and it’s driving the thyroid gland’s T4-T3 secretion higher. 
• In all of the “Type 1 Allostasis” health conditions, TSH is either “normal” or low. 
• Due to its many other hypersensitivities, TSH can be “exquisitely INsensitive” to hypothyroidism in tissues beyond the hypothalamus. 

It seems to many people like the poor TSH is confused, when it’s really just trying to do its job. 

In these states, it is not accurately indicating thyroid hormone status, because that’s not the only role of TSH in the human body. TSH is responding to many health factors beyond your thyroid hormones. It is trying to respond to the current health status or the disorder in the body.

An unnatural health condition missing from this table is “hypothyroidism caused by poor T4-T3 conversion while being treated with T4-monotherapy.” This is a state in which TSH can sometimes be normal, low or even suppressed while T3 is chronically below reference range (see Gullo et al, 2011 and Larisch et al, 2018).  Sadly, a subset of treated thyroid patients are maintained in this biochemical state also found in state of critical illness and heart failure. Nobody is doing research studies to track the health effects of this chronic hormone deficiency over years and decades.

3) THE TSH RANGE IS TOO WIDE FOR ANY INDIVIDUAL.

I’m not sure why this research is being overlooked, in the early years of this century it was discovered that a TSH reference range of 0.3 – 5.00 mU/L was approximately 50% wider than it is for a healthy individual over 12 months (Andersen et al, 2002).

Andersen and team found “the individuality index was below the critical limit of 0.6 for all thyroid function tests demonstrating that reference ranges based on the group are insensitive to significant changes in the individual.”

Notice that Andersen and team use the word “INsensitive.” This is the absolute opposite of “sensitive.”

Let’s think about how this affects interpretation of TSH test results. When judging a healthy-thyroid individual’s TSH, reliance on the population-wide statistical reference range is going to be too vague and generalized about 50% of the time.

In Andersen and team’s study, the population’s TSH reference range was 0.3 to 5.00. That’s a range that is 4.70 mU/L wide, and the mid-point is 2.20. 

• Person A who is not on thyroid therapy will have a TSH BELOW 2.2 when they are healthy (the vast majority of healthy adults fall in this category)
• Person B who is not on thyroid therapy will have a TSH ABOVE 2.2 when they are healthy. This pattern is less common. It occurs in advanced age, during the phase of recovery from illness, in euthyroid people with iodine deficiency, and during the non-symptomatic earliest stages of autoimmune thyroid gland failure. A slightly higher TSH is found in men than in women.
• Person C who is not on thyroid therapy might have a TSH mid range but not close to either end when they are healthy.

It is very important to keep in mind that this study was done in thyroid-healthy, untreated people. A similar study design would have revealed a much narrower, lower range of TSH and especially T3 in thyroidless people with thyroid disease who are on T4 monotherapy (Gullo et al, 2011; Larisch et al, 2018).

The TSH reference range falls “below the critical limit” of sensitivity — one test is NOT sensitive enough, and TSH alone fails the critical test of being a sole indicator of any individual even in the thyroid-healthy population. 

CONNECT THE DOTS

Now, add up the scientific proofs above — #1, #2, and #3, and ponder point #3B.

What do you get?

• The TSH test result is not indicative of thyroid hormone sufficiency outside the hypothalamus and pituitary.
• The TSH can be significantly lowered by factors beyond thyroid hormone supply in blood, even in people who aren’t being treated with thyroid hormone.
• The TSH range is not narrow enough to meet “critical limit” criteria for thyroid testing, since the population-wide range is far too wide to fit even a healthy-thyroid individual.

This reveals a shocking degree of failure. The degree cannot be expressed in a number.

The most shocking thing, worthy of a separate post on its own, is that the TSH test has still not been validated as a “surrogate endpoint” for use in clinical trials on thyroid therapy. (Validation of TSH as a biomarker for body-wide thyroid hormone sufficiency would require a special study design that has never yet been conducted in thyroid therapy.)

PUT TSH IN ITS APPROPRIATE CONTEXT

Of course TSH can still be a useful test in thyroid therapy — when it’s put in context.

The “T3 paradigm” we espouse is not about merely replacing the TSH with the FT3 test. That would be like replacing one “monotesting” policy with another.  You can’t say “Test ONLY T3!”  It is always unwise to interpret Free T3 without measuring Free T4, because a higher T3 can compensate for lower T4, but a low T3 can’t compensate for high-normal T4. 

In addition, the TSH result still provides some relevant information because it is one set of tissues’ response to thyroid hormones. You can also usefully look at total cholesterol levels and SHBG (sex-hormone binding globulin) as signals of T3 supply to the liver.

Instead of judging any single hormone in isolation, a person needs to listen to several, because the thyroid hormones affect all cells and tissues throughout the human body. 

Free T3, Free T4, clinical signs and symptoms and health status, plus the TSH can always be useful together as a chorus of biomarkers in assessing tissue euthyroidism and guiding therapy decisions.

Nevertheless, here are a couple of general principles about TSH levels alone:

• When TSH is higher than mid-range, it can strongly confirm an individual patient’s symptoms of hypothyroidism and the need for a dose increase or a change in thyroid therapy modality.
  • It is highly unlikely that any treated person’s healthy metabolic set point / equilibrium is located above the midpoint of the TSH reference range because TSH is biased lower in people with less thyroid tissue who are on thyroid therapy, and because the vast majority of the thyroid-healthy population has a TSH less than 2.5 mU/L.
  • Especially in cases of depression, a TSH above 2.5 is strongly indicative of tissue hypothyroidism in the brain in T4-treated people (Talaei et al, 2017)
• In patients who still partly depend on their own thyroid gland, having some TSH can help supply varying amounts of T3 as well as T4 to the body. It’s far more convenient and natural to have your “pharmacy in your neck” regulating part of your daily dose through a continual intravenous feed, rather than to have to rely 100% on medication absorbed in doses through the GI tract.
• In patients without any functional thyroid tissue on T4 monotherapy, freedom from hypothyroid symptoms is most often achieved when TSH is below the normo-thyroid reference range. (Larisch et al, 2018) 
• Any relative change in TSH can be very useful an indicator of the hypothalamus and pituitary’s distinctive response to a therapy adjustment. If the TSH does not change but other parameters do, one can use scientific knowledge to interpret what is going on in the treated patient’s body.
• You can detect mild or moderate degrees of central hypothyroidism or TSH hypersecretion if you enter TSH, T3 and T4 test results into the free clinical endocrinology app, the SPINA-Thyr program, while a person is on no therapy or T4-monotherapy. This can be helpful to diagnose pituitary disorders, resistance to thyroid hormone (RTH), the effect of TSHR antibodies on the ultrashort feedback loop, and thyroid hormone metabolism disorders like the partial SBP2 deiciency.

REFERENCES

Bianco, A. C., Dumitrescu, A., Gereben, B., Ribeiro, M. O., Fonseca, T. L., Fernandes, G. W., & Bocco, B. M. L. C. (2019). Paradigms of Dynamic Control of Thyroid Hormone Signaling. Endocrine Reviews, 40(4), 1000–1047. https://doi.org/10.1210/er.2018-00275

Gullo, D., Latina, A., Frasca, F., Le Moli, R., Pellegriti, G., & Vigneri, R. (2011). Levothyroxine Monotherapy Cannot Guarantee Euthyroidism in All Athyreotic Patients. PLoS ONE, 6(8). https://doi.org/10.1371/journal.pone.0022552

Larisch, R., Midgley, J. E. M., Dietrich, J. W., & Hoermann, R. (2018). Symptomatic Relief is Related to Serum Free Triiodothyronine Concentrations during Follow-up in Levothyroxine-Treated Patients with Differentiated Thyroid Cancer. Experimental and Clinical Endocrinology & Diabetes: Official Journal, German Society of Endocrinology [and] German Diabetes Association, 126(9), 546–552. https://doi.org/10.1055/s-0043-125064

Parikh, R., Mathai, A., Parikh, S., Chandra Sekhar, G., & Thomas, R. (2008). Understanding and using sensitivity, specificity and predictive values. Indian Journal of Ophthalmology, 56(1), 45–50. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2636062/