Thyroid blood values: which ones really count
TSH comes from the pituitary gland, not from the thyroid. And the number in the bracket next to it describes a group of strangers. Both change how you read your result.
All guides from the thyroid cluster
Every lab report carries two kinds of numbers. One kind gets talked about, that is your value. The other almost never does, those are the limits in the bracket. Yet this second kind helps decide whether a person becomes a patient.
You are holding a lab sheet. At the top it says TSH, next to it a number, and behind that in brackets something like 0.4 to 4.0.
Your number is 3.6. So you are inside. All good, says the call from the practice.
And you sit there and think: then why do I feel the way I feel. Or the other way round: your value is 4.4, so just outside, and suddenly there is a word in the room that was not there yesterday.
Many people know this moment, and in my consultations it is one of the most frequent reasons for a conversation. And almost every time we then talk about something that is not on the sheet: what this value actually measures, and where the two numbers in the bracket come from.
This is the entry article of our thyroid series. It does not answer which number is good, it answers what a number can say at all.
Why TSH alone is often not enough
If you have blood taken for the thyroid in Germany, you usually get exactly one number back. TSH. That is not carelessness, it is intent: TSH is the most sensitive single test we have. It just measures something different from what most people assume.
TSH comes from the head, not from the neck
TSH stands for thyroid stimulating hormone. It is produced in the pituitary gland, a pea sized gland on the underside of the brain. The thyroid does not produce it. It receives it.
Picture a heating system with a thermostat. The thyroid is the radiator. TSH is not the temperature in the room. It is the position of the dial. If the thermostat turns up high, that means: from my point of view the room is too cold.
So a high TSH value first of all means: the brain is asking more loudly. A low one means: the brain is asking more quietly. Both are statements about the assessment, not directly about the supply.
There is one more feature that makes this value so sensitive. The relationship between free T4 in the blood and TSH does not run in a straight line, it runs roughly logarithmically. A small change in thyroid hormone can trigger a large change in TSH. That is exactly why TSH reacts early when something shifts. And that is exactly why TSH numbers can look more dramatic than the difference in hormone levels often is.
A working group around Rudolf Hoermann reviewed in 2015 how the feedback loop between the pituitary gland and the thyroid actually works. Their finding: the balance is strongly individual, TSH also intervenes in the conversion of T4 into T3, and it can neither be defined normatively nor serve as a precise marker for a balanced thyroid state.
Important for perspective: the authors themselves say that hormone concentration in tissue has hardly ever been measured directly in humans.
Hoermann R et al. Front Endocrinol. 2015;6:177. DOI: 10.3389/fendo.2015.00177 [Mechanism Review]This criticism comes from inside, not from outside
It would be easy to turn this into a front line. They do it wrong, we do it right. That would only be dishonest, because the criticism of the pure TSH view stands in the guidelines themselves.
In its 2014 guideline the American Thyroid Association names as an explicit research need: better biomarkers for a balanced thyroid state are required to complement TSH measurement. And it writes that newer insights into physiology could explain why some people do not feel well on standard therapy. That is remarkably open for a professional society, since it relativizes its own leading value.
TSH is not a bad value. It is an excellent screening test, it finds early when something shifts in the feedback loop, and it is cheap and standardized worldwide.
What it is not: a status report. A screening test answers whether you should look more closely, not how you are doing. Separating those two questions takes a lot of pressure out of conversations about lab values.
From the perspective of Clinical Psychoneuroimmunology one more layer comes in: the pituitary gland sits at the same junction as the stress axis and the sex hormone axis, and what happens up there rarely happens in isolation. More on this in Cortisol and the HPA axis and in The thyroid and female hormones.
And now you know why a single value from the head cannot tell the whole story of the neck.
The panel: which value answers which question
One thing first, because it causes confusion surprisingly often. Thyroid values are not part of a normal blood count. The small and the large blood count measure blood cells. TSH, free T4, free T3 and the antibodies are hormone and immune values and have to be requested specifically.
So if somebody told you your blood count was fine, that says nothing about the thyroid yet. And now to the actual question: not which values exist, but what each one answers, and what it does not.
How loudly the brain is asking the thyroid.
Whether the feedback loop is out of balance. The most sensitive screening test we have.
How well the tissue is supplied, and whether your symptoms come from the thyroid.
How much unbound thyroxine circulates in the blood.
Whether the thyroid produces enough. The second step after a conspicuous TSH.
How much of it becomes the active form inside the cells.
How much unbound T3 is in the blood right now.
When hyperthyroidism is suspected, a sensible additional question.
For routine measurement without a concrete question, data showing better outcomes are missing.
Antibodies against a key enzyme of hormone production.
Whether an autoimmune process could sit behind a raised TSH, and how high the probability is across decades.
How you are doing today and whether a therapy would bring anything.
Antibodies against the storage protein of the thyroid.
Together with TPO antibodies an additional hint towards an autoimmune process.
Positive in isolation, it was not significantly associated with dysfunction in NHANES III.
Antibodies that occupy the receiver for the TSH signal and can switch it on permanently.
When Graves disease is suspected, the central question, see Graves disease and hyperthyroidism.
Without a suspicion of hyperthyroidism, a value with no question behind it.
Storage protein from thyroid tissue in the blood.
In the aftercare of certain thyroid diseases, a question about the course over time.
It is not designed as a screening test for fatigue.
Building blocks and cofactors for hormone production and conversion.
Whether a construction site is open in the background.
A cause. All four show associations, whether cause or consequence is open.
The sequence the guideline foresees
The German DEGAM guideline on a raised TSH value in primary care describes a very lean path. First a repeat measurement of TSH. Then, if it is still raised, free T4. And then, where appropriate, the autoantibodies.
Behind that sits the consideration that every additional value without a concrete question produces new incidental findings, which then have to be worked up again. Anyone who has slept badly for three months because of an incidental finding knows the downside of too much diagnostics.
What a functional view asks in addition: whether someone with persistent symptoms might still benefit from a broader look, particularly at the accompanying values. That does not replace the guideline logic, it adds a second perspective to it.
What is actually established about the accompanying values
Honesty pays off here, because a lot of certainty is claimed online where there is none. So I separate: established by meta-analyses, mechanistically plausible with thin human data, and clinically observed without a strong study basis.
A Chinese working group evaluated iron status and thyroid values together in 723 pregnant women in 2018. In the first trimester, iron deficiency was associated with a higher risk of a low free T4, odds ratio 14.86 with a confidence interval of 2.31 to 95.81, and 3.36 in the second trimester.
An interval that wide means: the direction is recognizable, the size of the effect is not. More in Iron deficiency, thyroid and sleep.
Teng X et al. Thyroid. 2018;28(8):968-975. DOI: 10.1089/thy.2017.0491 [Cohort, n=723]Wichman and colleagues summarized sixteen controlled trials on selenium in chronic autoimmune thyroiditis in 2016. Under ongoing levothyroxine therapy, TPO antibodies fell by an average of 271 units after three months, although with low overall evidence quality and more frequent adverse effects in the selenium group.
Whether this movement in the titer means anything clinically is open, the authors themselves write. More depth in Selenium, zinc, iron and vitamin D.
Wichman J et al. Thyroid. 2016;26(12):1681-1692. DOI: 10.1089/thy.2016.0256 [Meta-analysis, k=16]A meta-analysis from 2015 pooled twenty case control studies. People with an autoimmune thyroid disease had lower vitamin D levels than comparison subjects, and a deficiency was around three times as common.
Case control studies compare two groups at one point in time and cannot show what came first: solid as an association, not load bearing as a statement about cause.
Wang J et al. Nutrients. 2015;7(4):2485-2498. DOI: 10.3390/nu7042485 [Meta-analysis, k=20]And one sentence that belongs to this whole section: the fact that a value is associated with something in a study is not yet a recommendation to take the corresponding substance. Selenium and iodine in particular have a narrow band between too little and too much, more on iodine in Putting iodine in perspective. Whether, what and how much may be sensible in an individual case therefore belongs at a table with medical guidance, and not in an order placed after reading a blog article.
For zinc I found no robust meta-analysis, so it stands here as a factor worth thinking about without any claim to numbers. In men a look at prolactin is also worthwhile, see Prolactin and the thyroid in men, and on cofactors in energy metabolism From calorie to energy.
Two values I deliberately leave out. How T4 becomes T3 in the first place is in From T4 to T3. And that there is a value called reverse T3 is only mentioned here, it is assessed in Reverse T3: what the value can do.
A lab panel is not a shopping list where more is automatically better. Every value is the answer to a question, and if you do not have the question, you do not need the answer.
The most useful sentence before a blood draw is therefore: what exactly do we want to know, and what changes depending on how the result turns out.
And now you know why a good report is not the one with the most lines.
Where the reference ranges come from
Now comes the part that is surprisingly rarely covered. Almost every guide page lists values and names ranges, hardly any explains how these ranges came about. Yet that is exactly the decisive information, because a reference range is not a law of nature, it is a calculation.
Four steps from the blood sample to the number in the bracket
- Define a reference population. Who counts as healthy? Usually: no known thyroid disease, no medication, no goiter, often additionally without pregnant women and without antibody positive people. Each of these decisions shifts the result.
- Measure. The value is determined in everyone from this group, with a particular test method.
- Take the middle 95 percent. Order all values by size, cut off 2.5 percent at the bottom and 2.5 percent at the top.
- Write down the two edges. These two numbers end up in the bracket on your report.
Schematic illustration, not exact data. The point is the shape: most values crowd to the left, a long thin tail stretches to the right. The entire debate about the upper limit hangs on the question of how to read this tail.
The concrete reference group behind your bracket
For TSH the most important reference comes from an American population survey. It is called NHANES III and ran between 1988 and 1994.
Hollowell and colleagues published the thyroid values of 17,353 people aged twelve and over from the US population in 2002 and from them formed a narrower reference group of 13,344 people, without pregnant women, without hormone preparations, without antibody positive people. The geometric mean was 1.40 mIU/l, with differences by background from 1.57 in white people to 1.18 in black people.
So when somebody tells you your value is normal, translated that means: it lies within the middle 95 percent of this American sample from the early nineties.
Hollowell JG et al. J Clin Endocrinol Metab. 2002;87(2):489-499. DOI: 10.1210/jcem.87.2.8182 · PMID: 11836274 [Cross-sectional, n=17,353]Why the shape of the curve changes everything
For many lab values there is a bell shaped distribution, falling away symmetrically to the left and right of the peak. With TSH that is not the case. The distribution is right skewed, even after mathematical straightening: the peak sits far to the left, and a long thin tail stretches to the right.
And here the disputed question begins: who are the people in that tail?
Surks and Boucai reviewed in 2010 how TSH reference limits come about methodologically. The traditional assumption that the values in the right hand area reflect mild hypothyroidism was never tested, and the distribution additionally shifts with age and background.
The same curve therefore allows two readings: the right hand tail as hidden patients leads to a low upper limit, the right hand tail as biological spread leads to a higher one.
Surks MI, Boucai L. J Clin Endocrinol Metab. 2010;95(2):496-502. DOI: 10.1210/jc.2009-1845 · PMID: 19965925 [Review]The same raw data, four calculation methods, four limits
One study makes this point particularly clear, and it is also cited in the German guideline.
Strich and colleagues removed everyone with any indication of thyroid disease from a laboratory database in 2016 and evaluated the same TSH values with four statistical methods. Depending on the method, the upper limit dropped by up to 43 percent, down to 3.1 mIU/l, while free T3 and free T4 stayed stable across a wide TSH span.
The authors conclude from this that a universal upper cut-off may not exist at all.
Strich D et al. Clin Endocrinol (Oxf). 2016;85(1):110-115. DOI: 10.1111/cen.12970 · PMID: 26529455 [Cross-sectional, method comparison]Age and background shift the curve
There is a second reason why a single number for everyone does not fit well: the distribution moves with age.
Surks and Hollowell evaluated TSH distributions separately by decade of life and by antibody status in 2007. The upper limit rose from 3.56 in 20 to 29 year olds to 7.49 in people over 80, and even in the antibody negative group the curve of the very old was shifted to the right, peak included.
The shift therefore cannot be explained as hidden autoimmune disease, and a TSH of 5.2 means something different in a 28 year old than in an 82 year old.
Surks MI, Hollowell JG. J Clin Endocrinol Metab. 2007;92(12):4575-4582. DOI: 10.1210/jc.2007-1499 · PMID: 17911171 [Cross-sectional, NHANES III]The same picture appeared in a practice population of 22,116 people: the TSH median was 1.54 in white people and 1.18 in black people, and national rather than age and background specific limits wrongly classified 3 to 8 percent of people over 80 as conspicuous. A second finding belongs to both camps: when TSH rose above 4.5, free T4 fell on average by around 7 percent.
And a further analysis of the same data shows how strongly the definition of the reference group alone shifts the limit: if antibody positive people were taken out, the upper limit dropped by around 1.0 mIU/l.
The DEGAM guideline on a raised TSH value puts this remarkably directly. In its own words it says there: "There is no normal range." And: "The TSH ranges used so far are drawn too narrowly." As evidence it cites, among others, exactly that method comparison paper from 2016.
Consistently, it grades its limits by age: above 4.0 mU/l in 18 to 70 year olds, above 5.0 in those over 70 up to 80, above 6.0 in those over 80. That one number does not fit everyone is therefore disputed by nobody here, only the conclusions differ.
Your personal set point
There is a second layer that appears on no report. Within the group range, every person has their own, fairly narrow position. Technically this is called the set point.
A Danish working group took blood from sixteen healthy men every month over twelve months in 2002. Each person oscillated around their own mean, the individual intervals were only about half as wide as the group interval, and the index of individuality for TSH was 0.49.
The honest limit: sixteen men, one year, and to this day no intervention trial that has tested treatment guided by a personal set point against treatment guided by the reference range.
Andersen S et al. J Clin Endocrinol Metab. 2002;87(3):1068-1072. DOI: 10.1210/jcem.87.3.8165 · PMID: 11889165 [Longitudinal Cohort, n=16]The reference range describes a group of strangers. You are not a group. That is why a value inside the bracket can be unusual for you and a value just outside it can be entirely unremarkable.
The second sentence matters just as much: this is no licence to pick your own target value, it is an argument for the course over time rather than a single value.
And now you know why two laboratories can print two different brackets from the same raw data.
The debate about 2.5: both sides
Perhaps you have already come across the sentence that the TSH upper limit should really be 2.5 and not 4.0.
This sentence circulates in forums, podcasts and practices, often presented as secret knowledge. It is not. It is a peer reviewed professional position from one of the most important endocrinology journals, and the counter position stands in the same issue from September 2005. I am presenting both to you, as fairly as I can.
Should the TSH upper limit be lowered to 2.5?
- According to the laboratory guidelines of the National Academy of Clinical Biochemistry, over 95 percent of normal people lie below 2.5 mU/l.
- Earlier reference groups were mixed with undetected dysfunction, which pulled the mean upwards.
- The remaining higher values are outliers, mostly with Hashimoto thyroiditis.
- Core argument: African Americans with a very low Hashimoto rate have a mean TSH of 1.18, possibly the value of a genuinely healthy population.
- 9.7 percent of the reference group lie between 2.5 and 4.5, extrapolated around 20.6 million US Americans with a new diagnosis.
- Counter finding to the contamination thesis: most people with TPO antibodies lie below 2.5, so antibodies precisely do not explain the upper area.
- TSH varies considerably across the day and across months, a single value close to a limit carries little weight.
- Anyone who sits there and does have a dysfunction probably has a minimal one, without documented disadvantage and without proven treatment benefit.
Why the professional societies have not followed
Most professional societies have stayed at 4.0. That is readily interpreted as stubbornness. But there is a concrete reason in the data, and I find it strong.
Rodondi and colleagues pooled the individual data from eleven prospective cohorts in 2010: 55,287 people, over 542,000 person years. At a TSH of 4.5 to 6.9 the hazard ratio for coronary events was 1.00, at 7.0 to 9.9 it was 1.17 and only at 10 to 19.9 was it 1.89, while overall mortality was not raised.
The threshold of 10 in the guidelines is therefore not an arbitrary round number, it is the point from which something hard shows up in large data at all.
Rodondi N et al. JAMA. 2010;304(12):1365-1374. DOI: 10.1001/jama.2010.1361 · PMID: 20858880 [Individual Participant Data Meta-analysis, k=11, n=55,287]So the guidelines do not draw their limit out of convenience, but because below this threshold no demonstrable harm becomes visible in large datasets. What a functional view asks in addition: between the point at which cardiac risk becomes measurable and the point at which somebody feels different, there could be a gap. It has not been mapped so far, and it is a question, not a claim.
Both camps look at the same curve. They do not argue about the numbers. They argue about what the long tail to the right means.
The core of the reference range debateI deliberately give you no personal target value here. Not 2.5, not 4.0, not 1.8. Anyone who does that online acts as if a question were settled that has been argued over in professional journals for twenty years.
What you can take with you: the limit on your report is a decision that was made, not a constant of nature. Whoever knows that asks different questions in a consultation than only whether they are inside or outside.
And now you know why two well informed doctors can arrive at different assessments of the same value, without either of them working sloppily.
What disturbs the measurement
Before you interpret your value, a completely different question is worth asking. Was the measurement clean in the first place?
That is not mistrust of the laboratory. Modern immunoassays are excellent. But they measure under conditions, and you bring these conditions with you.
The time of day
TSH follows a daily rhythm. The peak lies at night, after which the value falls. Measuring at 7 in the morning measures something different from measuring at 11.
Roelfsema and colleagues took blood at close intervals over 24 hours from 117 healthy adults between 22 and 77 years old. The rhythm was robust and largely independent of age and sex, and the sample at 9 am explained roughly two thirds of the variability across 24 hours.
So you do not need to be afraid of the clock, you only need to keep it constant.
Roelfsema F et al. J Clin Endocrinol Metab. 2014;99(2):570-578. DOI: 10.1210/jc.2013-2858 · PMID: 24276453 [Cross-sectional, n=117]A side note on diligence: online you find the claim that TSH varies by up to 50 percent across the day. I could not find this number documented in any primary source. What is documented is the rhythm, so I leave the percentage out.
Breakfast
An Indian working group measured TSH and free T4 in 57 people while fasting and two hours after eating in 2014. TSH was lower after eating in everyone, and 15 of 20 people were classified as having latent hypothyroidism on the basis of their fasting value, although their value after eating lay inside the reference range.
The study is small and not randomized, the lesson still stands: the question is not fasting or not, but always the same way rather than sometimes one way and sometimes the other.
Nair R et al. Indian J Endocrinol Metab. 2014;18(5):705-707. DOI: 10.4103/2230-8210.139237 · PMID: 25285290 [Interventional Study, n=57]Biotin, the most avoidable error of all
I consider this section the practically most useful of the whole article. Because biotin sits in over the counter hair and nail supplements, and many people take it without thinking of it as a medication.
Li and colleagues gave six healthy adults 10 mg biotin daily for seven days in JAMA in 2017, an amount common in supplements. Nine of 23 tests using biotin streptavidin technology then showed interference and none of the 14 tests without this technology, with competitive assays delivering falsely high and sandwich assays falsely low values.
Because TSH is mostly measured as a sandwich assay and the free hormones often competitively, the false picture looks like hyperthyroidism, without one being present.
Li D et al. JAMA. 2017;318(12):1150-1160. DOI: 10.1001/jama.2017.13705 · PMID: 28973622 [Crossover Study, n=6]If you take biotin, say so before the blood draw. Not because you did something wrong, but because the laboratory knows which technology its test uses.
And the sentence that belongs to every section: nothing about your thyroid medication changes because of this. Changes to dose or intake belong exclusively in medical care, never in your own hands after reading a blog article.
Acute illness
When your body is busy with something bigger, the thyroid situation in the blood shifts. There is a name of its own for this: non-thyroidal illness syndrome, formerly also called low T3 syndrome.
Fliers and Boelen summarized the state of knowledge on non-thyroidal illness syndrome in 2021. In practically every severe illness, T3 in serum falls, while concentrations in tissue do not necessarily mirror this drop and, depending on the organ, can fall, stay the same or rise.
That disciplines both sides: during an infection you should not interpret values, and a low T3 is not automatically a thyroid problem.
Fliers E, Boelen A. J Endocrinol Invest. 2021;44(8):1597-1607. DOI: 10.1007/s40618-020-01482-4 · PMID: 33320308 [Mechanism Review]The German guideline draws a clear practical consequence from this: after discharge from hospital, except after thyroid surgery, a check should only take place four weeks later.
When the test itself is deceived
There is one constellation that can follow people for years: TSH is clearly raised, free T4 and free T3 are entirely normal, there are no symptoms, and the value does not move for years.
Hattori and colleagues looked at 1,794 women trying to conceive in 2018 and examined conspicuous samples with special methods. In 24 people TSH was falsely raised by human anti mouse antibodies, genuine macro TSH was found in three people, and all three commercial tests examined recognized it and therefore reported raised values.
Very rare, then, but without clarification it leads to years of unnecessary worry, and the way there is a query to the laboratory.
Hattori N et al. Thyroid. 2018;28(10):1252-1260. DOI: 10.1089/thy.2017.0624 · PMID: 29943675 [Cross-sectional, n=1,794]Medication
A whole range of medications can influence thyroid values, through hormone production, through the binding proteins in the blood or through absorption in the gut. Substances with instructions attached I deliberately do not list. The sensible approach is the same anyway: bring a complete list of all your preparations to the appointment and change none of them yourself.
Pregnancy and trying to conceive
This section is short, and that is intentional. In pregnancy, thyroid values have their own assessment standards and their own framework of care. The American Thyroid Association published a dedicated and very extensive guideline for this in 2017, which treats diagnostics, iodine supply, antibodies, conception, breastfeeding and screening separately.
Target values for pregnancy I deliberately do not name here. Not because they are unimportant, but because they belong in the setting where they can also be checked.
One number shows how dynamic the situation is: according to the German guideline, 70 to 80 percent of previously treated pregnant women temporarily need a higher dose, around 20 to 30 percent more, checked at least once per trimester. That is not something you steer yourself, it belongs in close medical care. On the connection with the cycle: The thyroid and female hormones.
A lab value is not a photograph of your thyroid. It is a photograph of thyroid plus time of day plus breakfast plus infection plus supplements plus test method.
That does not devalue the measurement, it makes it readable. Whoever knows which factors are in the picture can keep them constant next time. And that is exactly how two single values become a course over time.
And now you know why the question about time of day and supplements is not a side issue but half the interpretation.
What antibodies mean and what they do not
The report says TPO-Ab, behind it a three digit number and an arrow pointing up. Suddenly there is a word in the room that was not there before: autoimmune.
Many people experience this moment as a turning point. Until a minute ago it was fatigue, now it is a diagnosis that sounds like forever. So it is worth sorting out precisely what antibodies say and what they do not.
What they say: a probability across decades
The Whickham Survey followed 2,779 adults from northern England over twenty years, with a completeness of over 97 percent. Those with positive antibodies alone had an odds ratio of 8 in women and 25 in men for later hypothyroidism, those with both, antibodies and a raised TSH, were at 38 and 173 respectively.
So antibodies say something about probability in the coming decades, nothing about how you are doing today.
Vanderpump MP et al. Clin Endocrinol (Oxf). 1995;43(1):55-68. DOI: 10.1111/j.1365-2265.1995.tb01894.x · PMID: 7641412 [Cohort, n=2,779]What they do not say
One particularly persistent assumption goes: whoever has antibodies should be treated earlier, because they benefit more. This assumption has been tested.
Lyko and colleagues combined the individual data from two randomized trials in 2022: 660 people aged 65 and over, 188 of them with positive TPO antibodies. The adjusted between group difference in symptoms was minus 2.07 in the antibody positive group and 0.89 in the negative group, with a p value for the interaction of 0.31, so no robust difference.
That applied to people aged 65 and over and cannot be transferred to younger people, and prognostically the antibodies remain informative.
Lyko C et al. J Intern Med. 2022;292(6):892-903. DOI: 10.1111/joim.13544 · PMID: 35894851 [Pooled RCT Data, n=660]And one more common finding belongs here. When only the thyroglobulin antibodies are positive but the TPO antibodies are negative, then in NHANES III, the largest available population sample, that was not significantly associated with dysfunction. For TPO antibodies it very much was.
The DEGAM guideline is unusually clear here. TPO antibodies can be measured once to clarify a suspicion of Hashimoto thyroiditis. A repeat measurement is not recommended, and with a TSH inside the normal range the measurement is not recommended at all.
The reason: the titer does not run parallel to the state of the thyroid, it is no good as a temperature chart. Whoever measures it every three months anyway mainly buys an emotional rollercoaster, usually without any additional information coming out of it.
Antibodies are a weather forecast, not a weather report. They say something about the probability of rain in the coming years, nothing about whether it is raining today.
And a forecast is not a verdict, it is information you can plan ahead with, for example the intervals between checks.
Further reading: why the immune system attacks the thyroid at all is in Understanding Hashimoto. What moves behind the titer belongs in Lowering Hashimoto antibodies, and the role of environmental exposure in Heavy metals and Hashimoto.
And now you know why an antibody titer is a prognosis and not a description of your current state.
The grey zone between 4 and 10
Let us come to the constellation that affects the most people and triggers the most discussion.
TSH above the limit, free T4 inside the reference range. This is called latent or subclinical hypothyroidism. In NHANES III it affected 4.3 percent of the population.
First: these values are often not stable
Somwaru and colleagues measured 3,996 people aged 65 and over again after two and after four years, 459 of them with latent hypothyroidism and no medication. After four years it was still present in 56 percent, with a starting TSH of 4.5 to 6.9 the value was back inside the reference range in 46 percent after just two years, with a starting TSH of 10 and above in only 7 percent.
A single measurement in the grey zone is therefore not a diagnosis, it is a reason to look again after two to three months.
Somwaru LL et al. J Clin Endocrinol Metab. 2012;97(6):1962-1969. DOI: 10.1210/jc.2011-3047 · PMID: 22438233 [Cohort, n=3,996]Second: what happened when the value was actually lowered
This is the question it really comes down to. Not whether the value can be moved, but whether that changes anything.
The TRUST trial randomized 737 people aged 65 and over with persistent latent hypothyroidism double blind to levothyroxine or placebo. After one year TSH had fallen from 6.40 to 3.63, the difference in hypothyroid symptoms was 0.0 points and in tiredness 0.4 points, against a clinical relevance threshold of 9 points.
So lab value and how people felt decoupled, and the limits belong with it: people aged 65 and over, mean TSH 6.4, one year of observation.
Stott DJ et al. N Engl J Med. 2017;376(26):2534-2544. DOI: 10.1056/NEJMoa1603825 · PMID: 28402245 [RCT, n=737]Feller and colleagues summarized all randomized trials on this question in JAMA in 2018: 21 trials, 2,192 non pregnant adults. The standardized mean difference was minus 0.11 for quality of life and 0.01 for thyroid related symptoms, with a low risk of bias and GRADE moderate to high.
A narrow interval around zero is not a shrug, it is a robust result.
Feller M et al. JAMA. 2018;320(13):1349-1359. DOI: 10.1001/jama.2018.13770 · PMID: 30285179 [Meta-analysis, k=21, n=2,192]Third: the opposite direction belongs here too
In the functional discourse the other side is often missing: the risk of doing too much.
Taylor and colleagues evaluated 52,298 people with a first levothyroxine prescription between 2001 and 2009 from a British primary care database. The mean TSH at the start of therapy fell from 8.7 to 7.9, and after five years 5.8 percent had a TSH below 0.1, particularly people who had been treated for fatigue or depression.
Whoever is treated because of symptoms and feels no improvement is easily titrated further upwards, and that is exactly why medical follow-up checks are needed.
Taylor PN et al. JAMA Intern Med. 2014;174(1):32-39. DOI: 10.1001/jamainternmed.2013.11312 · PMID: 24100714 [Cohort, n=52,298]Nothing in this section is a reason to change anything about ongoing thyroid medication, neither dose nor timing nor preparation. A TSH that is too low under therapy is a reason for a conversation and a check, not for a decision of your own.
Stopping or reducing abruptly can trigger symptoms and create dangerous situations. If something about your therapy is on your mind, the next step is an appointment, not the pill box.
Fourth: the guideline view that hardly anyone knows
Now comes the part that surprised me most in this whole debate. Because the European guideline is considerably more differentiated than the discourse suggests.
The European Thyroid Association distinguishes two categories: mildly raised TSH from 4.0 to 10.0 and clearly raised TSH above 10. Every first finding should be confirmed with a repeat measurement of TSH and free T4 plus TPO antibodies, preferably after two to three months.
And then there is a sentence many people do not know: in younger people with a TSH below 10 and symptoms that fit hypothyroidism, a trial of therapy should be considered. With a clear condition: the response is checked three to four months after a TSH inside the reference range has been reached, and if the symptoms do not improve, the therapy should as a rule be stopped again.
As a target it names the lower half of the reference range, that is 0.4 to 2.5. In the very old, over 80 to 85 years, it conversely advises a watchful approach.
So that no misunderstanding arises here: this is the description of a guideline procedure, not an instruction to you. A trial of therapy is started medically, checked medically and stopped medically. Neither starting nor stopping nor a dose change belongs in your own hands.
I find that remarkable. The guideline says neither a blanket yes nor a blanket no, it describes a structured trial with a defined stopping point. That is exactly the stance a functional view would also take.
Important for perspective: TRUST included exclusively people aged 65 and over with a mean TSH of 6.4 over one year. Neither the German nor the European guideline transfers that to people under 30, to pregnant women or to women trying to conceive. Whoever concludes from it that latent hypothyroidism should never be treated overstretches the study.
How cleanly good guideline work distinguishes is shown by the JAMA consensus paper from 2004: whether a latent hypothyroidism progresses into a manifest one, the data on that are good, whether a treatment prevents this progression, the data on that are insufficient. This distinction between an established association and an established treatment benefit is the central figure of thought in this topic. Pregnancy counted as a special case even back then.
When something belongs in medical hands, independent of any debate
- TSH above 10. Here you leave the grey zone, and the data look different.
- Palpitations, racing heart, rhythm disturbances. Especially with a very low TSH, have this looked at promptly.
- Signs of hyperthyroidism: unintended weight loss, inner restlessness, tremor, feeling hot. More in Graves disease and hyperthyroidism.
- Nodules, rapidly growing swelling, difficulty swallowing or hoarseness. That needs to be examined, see Nodules and goiter, properly investigated.
- Thyrotoxic crisis or myxedema are emergencies. Rare, but they exist, and they belong in hospital immediately.
- Trying to conceive and pregnancy. Own standards, own framework of care, raise it early.
The question is not: treat or do not treat. It is: what exactly should get better, how would we notice it, and by when.
A therapy with a defined goal and a defined point of review is something different from one that simply keeps running. You can raise this difference in a consultation.
And now you know why the grey zone between 4 and 10 is contested not because someone is thinking sloppily, but because the data there really are ambiguous.
How to read your report without falling into number panic
That leaves the question of what you can do with all this at the kitchen table. Three thoughts that I clinically observe to take the most pressure out of it.
One value is a photo, not a film
A single measurement tells you where you stood on one morning. Two measurements spaced apart tell you a direction, and that is almost always the more useful information. Almost half of the values in the lower grey zone were back inside the reference range on their own after two years. Whoever measures again first instead of acting immediately is not hesitant, they are careful.
The range describes a group, not you
The bracket on your report comes from a sample of people you have never met, measured with a particular method and calculated with a particular approach. That is not an argument against reference ranges, without them we would have no orientation at all. It is an argument for reading them as what they are: a good first approximation, not a personal verdict.
Why there is no meaningful thyroid value calculator
I am regularly asked whether this could not be automated: enter the numbers, get an assessment out. After everything you have read, you can see for yourself where that fails. A calculator does not know your time of day, your breakfast, your hair supplement, your infection three weeks ago and certainly not your personal set point.
What you can bring to the appointment
- The time of the blood draw. Choose the same one every time if possible.
- Fasting or not. More important than the answer is that it is the same next time.
- All preparations, really all of them. Including hair and nail products and multivitamins. Biotin is the most common silent disturbing factor.
- Acute illnesses in recent weeks. An infection changes the situation temporarily.
- Old reports. A course over years is worth more than the most precise single value.
- Your actual question. Not only the values, but what you hope the result will give you.
When the values are fine and the symptoms stay
That is a question of its own. Two sentences on it: the fact that all values lie inside the range does not end the search, it only moves it. And the thyroid is rarely the first link in a chain. Read on in Normal values, symptoms anyway and in Functional hypothyroidism.
Anthroposophic medicine looks at the thyroid not only as a hormone gland, but as an organ at the border between head and trunk, where breath, voice and swallowing come together. That is a way of looking and a clinical tradition, not a study finding, and it stands next to physiology, not in its place.
What I find useful about it is that it reminds us that a person is more than their measured value. If you would like to lay it alongside once, you will find it in The thyroid seen anthroposophically.
And now you know why the calmest reaction to a borderline result is almost always the best one: measure again, under the same conditions, and then talk it through in peace.
Frequently asked questions about thyroid values
Which thyroid values are there in the first place?
The basis is three values: TSH from the pituitary gland, free T4 as the storage form and free T3 as the active form. Added to that are TPO antibodies, thyroglobulin antibodies and, when Graves disease is suspected, TRAb. Thyroglobulin itself is a follow-up parameter and not a screening test. As accompanying values, ferritin, selenium, zinc and vitamin D are often measured alongside.
Are thyroid values part of a normal blood count?
No. The small and the large blood count measure blood cells. TSH, free T4, free T3 and the antibodies are hormone and immune values and have to be requested specifically. If you were told your blood count was fine, that does not automatically mean the thyroid was examined too.
Do I have to fast for thyroid values?
What matters is less whether you fast than that you handle it the same way every time. In a small study of 57 people, TSH was lower after eating than when fasting, and 15 of 20 people were classified differently depending on their fasting status (Nair 2014). Agree on the approach with your practice and then stick to it.
At what time of day should thyroid values be measured?
TSH follows a daily rhythm with a peak at night. In a study of 117 healthy adults, the 9 am sample explained roughly two thirds of the variation across 24 hours (Roelfsema 2014). For everyday life this mainly means: always at roughly the same time.
Why is the TSH value alone often not enough?
TSH is produced in the pituitary gland, not in the thyroid. It describes how loudly the brain is asking, which makes it a very sensitive screening test but an indirect measure of how well the tissue is supplied. This is why the American professional society explicitly names better markers as an open research question.
Where do the numbers in the reference range come from, and who set them?
A reference range is a statistical construct: you measure a group defined as healthy and write down the middle 95 percent. For TSH the most important reference comes from NHANES III with 17,353 people measured, 13,344 of them in the cleaned group (Hollowell 2002). By definition, 2.5 percent of healthy people lie above it.
Why do some say the TSH upper limit should be 2.5 instead of 4.0?
This position was formulated in 2005 by Wartofsky and Dickey in the journal JCEM. Their argument: over 95 percent of healthy people would lie below 2.5, and older reference groups were mixed with undetected disease. As evidence they point out that African Americans with a very low Hashimoto rate have a mean TSH of 1.18.
Why have the professional societies mostly not followed this lowering?
In the same issue Surks, Goswami and Daniels replied: 9.7 percent of the reference group lie between 2.5 and 4.5, which would mean around 20.6 million people with a new diagnosis without proven treatment benefit. On top of that: in pooled data from 55,287 people the risk of coronary events at a TSH of 4.5 to 6.9 was not raised (Rodondi 2010).
My TSH is raised, fT4 and fT3 are normal. What does that mean?
That is the constellation described as latent or subclinical hypothyroidism. It is common and often not stable: in a cohort of 3,996 people aged 65 and over, the value was back inside the reference range after two years in almost half of them (Somwaru 2012). This is why guidelines schedule a follow-up measurement after two to three months.
Can biotin distort my thyroid values?
Yes, and it is a common and avoidable measurement error. In a study in JAMA, six adults took 10 mg biotin daily for seven days, an amount common in hair supplements. Nine of 23 tests using biotin streptavidin technology then gave deviating results (Li 2017). The false picture looks like hyperthyroidism.
What do raised TPO antibodies mean when TSH is normal?
They show that the immune system is occupied with the thyroid. About today they say little, about the coming decades quite a lot: in the Whickham Survey the odds ratio for later hypothyroidism with antibody positivity alone was 8 for women and 25 for men (Vanderpump 1995). That is a statement of probability, not a diagnosis.
Do thyroid antibodies need to be checked regularly?
The German DEGAM guideline explicitly says no. TPO antibodies can be measured once to clarify a suspicion of Hashimoto thyroiditis. A repeat measurement is not recommended, and with a TSH inside the normal range the measurement is not recommended at all. The titer is not a follow-up parameter.
How often should thyroid values be checked?
That depends on the situation and belongs in the hands of the practice treating you. The DEGAM guideline gives as orientation: with latent hypothyroidism without symptoms and a TSH up to 10, a check after 6 to 12 months. Under ongoing therapy at the earliest 8 weeks after a dose change. After a hospital stay, wait four weeks.
Should I have blood drawn before or after the tablet?
What matters is that you handle it the same way every time and that the practice knows how you handled it. Note the time of the dose and of the blood draw and bring that with you. Please do not change anything about how you take your thyroid medication on your own. How it should run in your case is something you discuss with your doctor beforehand.
Why was my value different three months ago even though I changed nothing?
There are two explanations. First, every person has their own position inside the reference range, and a single TSH value describes it only roughly: in 16 healthy men the individual variation interval was about half as wide as the group interval (Andersen 2002). Second, values in the grey zone often change on their own.
Where this leads from here
Lab values never stand on their own. They hang on iron, on stress, on inflammation and on the question of what a reference range describes.
Thyroid values
Panel, reference ranges, disturbing factors
this articleFerritin: what is normal
The same question asked of another lab value, and why the bracket is especially wide there
Cofactors and energy
Which building blocks metabolism needs before hormones can work at all
Cortisol and the HPA axis
The second axis from the same control centre in the head
And if you are interested in why a low ferritin and an inflammation can hide each other, Iron, inflammation and the hepcidin block is worth reading.
Scientific sources
- Hollowell JG, Staehling NW, Flanders WD et al. Serum TSH, T4, and thyroid antibodies in the United States population (1988 to 1994): NHANES III. J Clin Endocrinol Metab. 2002;87(2):489-499. PMID: 11836274 · DOI: 10.1210/jcem.87.2.8182 [Cross-sectional, n=17,353]
- Surks MI, Boucai L. Age- and race-based serum thyrotropin reference limits. J Clin Endocrinol Metab. 2010;95(2):496-502. PMID: 19965925 · DOI: 10.1210/jc.2009-1845 [Review]
- Strich D, Karavani G, Levin S, Edri S, Gillis D. Normal limits for serum thyrotropin vary greatly depending on method. Clin Endocrinol (Oxf). 2016;85(1):110-115. PMID: 26529455 · DOI: 10.1111/cen.12970 [Cross-sectional, method comparison]
- Surks MI, Hollowell JG. Age-specific distribution of serum thyrotropin and antithyroid antibodies in the US population. J Clin Endocrinol Metab. 2007;92(12):4575-4582. PMID: 17911171 · DOI: 10.1210/jc.2007-1499 [Cross-sectional, NHANES III]
- Boucai L, Surks MI. Reference limits of serum TSH and free T4 are significantly influenced by race and age in an urban outpatient medical practice. Clin Endocrinol (Oxf). 2009;70(5):788-793. PMID: 18727705 · DOI: 10.1111/j.1365-2265.2008.03390.x [Cross-sectional, n=22,116]
- Boucai L, Hollowell JG, Surks MI. An approach for development of age-, gender-, and ethnicity-specific thyrotropin reference limits. Thyroid. 2011;21(1):5-11. PMID: 21058882 · DOI: 10.1089/thy.2010.0092 [Cross-sectional, NHANES III]
- Wartofsky L, Dickey RA. The evidence for a narrower thyrotropin reference range is compelling. J Clin Endocrinol Metab. 2005;90(9):5483-5488. PMID: 16148345 · DOI: 10.1210/jc.2005-0455 [Review]
- Surks MI, Goswami G, Daniels GH. The thyrotropin reference range should remain unchanged. J Clin Endocrinol Metab. 2005;90(9):5489-5496. PMID: 16148346 · DOI: 10.1210/jc.2005-0170 [Review]
- Laurberg P, Andersen S, Carlé A, Karmisholt J, Knudsen N, Pedersen IB. The TSH upper reference limit: where are we at? Nat Rev Endocrinol. 2011;7(4):232-239. PMID: 21301488 · DOI: 10.1038/nrendo.2011.13 [Review]
- Andersen S, Pedersen KM, Bruun NH, Laurberg P. Narrow individual variations in serum T4 and T3 in normal subjects. J Clin Endocrinol Metab. 2002;87(3):1068-1072. PMID: 11889165 · DOI: 10.1210/jcem.87.3.8165 [Longitudinal Cohort, n=16]
- Hoermann R, Midgley JEM, Larisch R, Dietrich JW. Homeostatic Control of the Thyroid-Pituitary Axis: Perspectives for Diagnosis and Treatment. Front Endocrinol (Lausanne). 2015;6:177. PMID: 26635726 · DOI: 10.3389/fendo.2015.00177 [Mechanism Review]
- Stott DJ, Rodondi N, Kearney PM et al. Thyroid Hormone Therapy for Older Adults with Subclinical Hypothyroidism (TRUST). N Engl J Med. 2017;376(26):2534-2544. PMID: 28402245 · DOI: 10.1056/NEJMoa1603825 [RCT, n=737]
- Feller M, Snel M, Moutzouri E et al. Association of Thyroid Hormone Therapy With Quality of Life and Thyroid-Related Symptoms in Patients With Subclinical Hypothyroidism. JAMA. 2018;320(13):1349-1359. PMID: 30285179 · DOI: 10.1001/jama.2018.13770 [Meta-analysis, k=21, n=2,192]
- Lyko C, Blum MR, Abolhassani N et al. Thyroid antibodies and levothyroxine effects in subclinical hypothyroidism: a pooled analysis of two randomized controlled trials. J Intern Med. 2022;292(6):892-903. PMID: 35894851 · DOI: 10.1111/joim.13544 [Pooled RCT Data, n=660]
- Rodondi N, den Elzen WPJ, Bauer DC et al. Subclinical hypothyroidism and the risk of coronary heart disease and mortality. JAMA. 2010;304(12):1365-1374. PMID: 20858880 · DOI: 10.1001/jama.2010.1361 [Individual Participant Data Meta-analysis, k=11, n=55,287]
- Somwaru LL, Rariy CM, Arnold AM, Cappola AR. The natural history of subclinical hypothyroidism in the elderly: the cardiovascular health study. J Clin Endocrinol Metab. 2012;97(6):1962-1969. PMID: 22438233 · DOI: 10.1210/jc.2011-3047 [Cohort, n=3,996]
- Taylor PN, Iqbal A, Minassian C et al. Falling threshold for treatment of borderline elevated thyrotropin levels. JAMA Intern Med. 2014;174(1):32-39. PMID: 24100714 · DOI: 10.1001/jamainternmed.2013.11312 [Cohort, n=52,298]
- Roelfsema F, Pijl H, Kok P et al. Thyrotropin secretion in healthy subjects is robust and independent of age and gender. J Clin Endocrinol Metab. 2014;99(2):570-578. PMID: 24276453 · DOI: 10.1210/jc.2013-2858 [Cross-sectional, n=117]
- Nair R, Mahadevan S, Muralidharan RS, Madhavan S. Does fasting or postprandial state affect thyroid function testing? Indian J Endocrinol Metab. 2014;18(5):705-707. PMID: 25285290 · DOI: 10.4103/2230-8210.139237 [Interventional Study, n=57]
- Li D, Radulescu A, Shrestha RT et al. Association of Biotin Ingestion With Performance of Hormone and Nonhormone Assays in Healthy Adults. JAMA. 2017;318(12):1150-1160. PMID: 28973622 · DOI: 10.1001/jama.2017.13705 [Crossover Study, n=6]
- Hattori N, Aisaka K, Chihara K, Shimatsu A. Current Thyrotropin Immunoassays Recognize Macro-Thyrotropin Leading to Hyperthyrotropinemia in Females of Reproductive Age. Thyroid. 2018;28(10):1252-1260. PMID: 29943675 · DOI: 10.1089/thy.2017.0624 [Cross-sectional, n=1,794]
- Fliers E, Boelen A. An update on non-thyroidal illness syndrome. J Endocrinol Invest. 2021;44(8):1597-1607. PMID: 33320308 · DOI: 10.1007/s40618-020-01482-4 [Mechanism Review]
- Vanderpump MP, Tunbridge WM, French JM et al. The incidence of thyroid disorders in the community: a twenty-year follow-up of the Whickham Survey. Clin Endocrinol (Oxf). 1995;43(1):55-68. PMID: 7641412 · DOI: 10.1111/j.1365-2265.1995.tb01894.x [Cohort, n=2,779]
- Teng X, Shan Z, Li C et al. Iron Deficiency May Predict Greater Risk for Hypothyroxinemia: A Retrospective Cohort Study of Pregnant Women in China. Thyroid. 2018;28(8):968-975. PMID: 29968513 · DOI: 10.1089/thy.2017.0491 [Cohort, n=723]
- Wichman J, Winther KH, Bonnema SJ, Hegedüs L. Selenium Supplementation Significantly Reduces Thyroid Autoantibody Levels in Patients with Chronic Autoimmune Thyroiditis. Thyroid. 2016;26(12):1681-1692. PMID: 27702392 · DOI: 10.1089/thy.2016.0256 [Meta-analysis, k=16]
- Wang J, Lv S, Chen G et al. Meta-analysis of the association between vitamin D and autoimmune thyroid disease. Nutrients. 2015;7(4):2485-2498. PMID: 25854833 · DOI: 10.3390/nu7042485 [Meta-analysis, k=20]
- Surks MI, Ortiz E, Daniels GH et al. Subclinical thyroid disease: scientific review and guidelines for diagnosis and management. JAMA. 2004;291(2):228-238. PMID: 14722150 · DOI: 10.1001/jama.291.2.228 [Systematic Review]
- Pearce SHS, Brabant G, Duntas LH et al. 2013 ETA Guideline: Management of Subclinical Hypothyroidism. Eur Thyroid J. 2013;2(4):215-228. PMID: 24783053 · DOI: 10.1159/000356507 [Guideline]
- Jonklaas J, Bianco AC, Bauer AJ et al. Guidelines for the treatment of hypothyroidism: prepared by the American Thyroid Association task force on thyroid hormone replacement. Thyroid. 2014;24(12):1670-1751. PMID: 25266247 · DOI: 10.1089/thy.2014.0028 [Guideline]
- Alexander EK, Pearce EN, Brent GA et al. 2017 Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease During Pregnancy and the Postpartum. Thyroid. 2017;27(3):315-389. PMID: 28056690 · DOI: 10.1089/thy.2016.0457 [Guideline]
- Schübel J, Voigt K, Uebel T. DEGAM-Leitlinie Nr. 18: Erhöhter TSH-Wert in der Hausarztpraxis. German S2k guideline, AWMF register no. 053-046, update 2023, valid until May 2027. AWMF register [Guideline]
- National Academy of Clinical Biochemistry. Laboratory Medicine Practice Guidelines: Laboratory Support for the Diagnosis and Monitoring of Thyroid Disease. Cited after Wartofsky & Dickey 2005, PMID: 16148345 · DOI: 10.1210/jc.2005-0455 [Guideline Document, cited indirectly]