T3 Controversies: Is TSH Reliable when Taking T3?


In 2019, I published a post called Is TSH the Best Test? If you’re a fan of the blog — or even if you hate-read Hormones Demystified — you almost certainly know that I advocated strongly for the use of TSH in the clinical care of hypothyroidism. I also debunked several of alternative medicine’s claims regarding the supposed deficiencies of using TSH to diagnose thyroid hormone deficiency and guide thyroid hormone replacement. And I did this all by using science to show how alt med corrupts actual medical evidence, twisting it in ways that — in the absence of sufficient inspection — appear to bolster alt med’s claims.

One thing I didn’t adequately explore in that post, however, is whether TSH is specifically less reliable in guiding thyroid hormone dosing when the thyroid preparation in question contains a sizable proportion of T3 (such as any desiccated thyroid medication, or pig thyroid, as I prefer to call it). I didn’t realize the extent to which this assertion is promoted in the alt med world, until it was brought to my attention by an astute reader (thanks, Tayshia1).

In Search of the (Pseudo) Science

As y’all are aware, I love diving down alt med rabbit holes to unearth the perversions of real science to be found there.2 Regarding the influence of exogenous T33 on TSH production, however, I found surprisingly little internet fodder that was more substantive than stereotypical alt med word salad. That said, I was able to find one well-researched, intelligently-written, and ultimately misguided piece on the web that cited legitimate articles from legitimate medical journals.

So, in this post, I will endeavor to explain why I think the author of that well-intentioned piece drew the wrong conclusions. As always, if you can point me in the direction of other proposed mechanisms that you’ve found on the internet, I’m happy to vet them.

If you don’t spend much of your time marinating in this stuff, you may want to first re-read just this section of Is TSH the Best Test? which goes over the relevant aspects of how T4 and T3 interact with/feed back to the pituitary. 


The “Evidence” for T3-containing Medication Making TSH Unreliable is Weak — Very, Very Weak

The first argument made by the author derives from an old study in which the researcher gave some people with and without hypothyroidism a few different doses of a combination T4/T3 medication, for 3-4 weeks at a time. Then, the investigator did what’s called a TRH stimulation test. In this test, a synthetic version of a natural hormone produced by the hypothalamus — thyrotropin-releasing hormone — is infused into the subject, and then TSH levels are serially measured at intervals after the infusion. Because TRH’s purpose is to tell the pituitary to release TSH, the hope here is that one can learn something useful by observing the TSH response to an infusion of TRH — more on this in a minute.

In this particular study, the investigator wanted to “quantitate the degree to which small quantities of triiodothyronine (T3) and thyroxine (T4) can inhibit the release of thyrotropin (TSH)….” He seems to be postulating that, if a given dose of T4/T3 medication does not push serum T4 and T3 levels above the normal range, then any subsequent blunting of the expected TSH rise after an infusion of TRH should be interpreted as the TSH being overly sensitive to suppression by even “small” doses of thyroid medication. Taking this line of reasoning a step further, if the TSH can potentially be suppressed by even small doses of thyroid hormone, then TSH might not be the most accurate reflection of one’s thyroid status during thyroid hormone replacement therapy.

I have multiple criticisms of this study’s design, as well as the extrapolation of its findings to suggest that exogenous T3 suppresses TSH out of proportion to what an equivalent amount of exogenous T4 does to TSH.

But I don’t think any of my complaints are more important than the fact that TRH stimulation tests yield results that lack both reliability and reproducibility, they never caught on as a useful clinical tool, TRH hasn’t been available in the US for decades, and the rare clinician who gets her hands on some TRH usually doesn’t understand the test characteristics or what to make of the results.4

Before the advent of third and fourth generation TSH blood tests that were sensitive enough to detect lower levels of TSH, the TRH stimulation test was mainly used to attempt to diagnose central hypothyroidism, a rare condition in which the pituitary does not release enough TSH to stimulate the thyroid to produce adequate amounts of thyroid hormone. TRH testing fell out of favor a long time ago, so we simply don’t have enough cumulative experience with the test to know what to make of the way in which it was used in the particular study I mentioned above.

Next, it is critical to note that the pituitary’s response to an artificially-induced stimulus like exogenous TRH cannot be extrapolated to interpret how the pituitary would respond to a normal physiologic stimulus. What I mean to say is: walloping someone’s pituitary with a big dose of TRH should not be thought to reflect the normal cross-talk between the hypothalamus and pituitary when responding to changes in ambient T4 and T3 levels.

Quite honestly, I feel like I’ve said all I need to say about why I find the citing of this particular study uncompelling. I realize, though, that some of you want more — good on you for being curious. But there’s another reason why I’m not going to dive into the results of the TRH stimulation testing from that study, and that’s because the study was not designed in a way that could even address whether T3 suppresses TSH production more than an equivalent dose of T4. Let me explain.

The author of the study said he wanted to see what “small doses” of T4 and T3 would do to TSH, and the three medication combinations he used were: 15mcg T3 + 60mcg T4, 22.5mcg T3 + 90mcg T4, and 30mcg T3 + 120mcg T4.5 

How can you possibly draw any conclusions about T3 suppressing TSH out of proportion to what an equivalent dose of T4 would do when you’re giving both T3 and T4 in escalating doses? Why not give only T3 to the subject and then do your fancy TRH testing, followed by giving only an equivalent dose of T4, followed by more TRH testing?6 Or how about giving three different combination products that all have the same effective dose of thyroid hormone, just increasing the amount of T3 and decreasing the T4 in products two and three? Even these two suggested study designs wouldn’t be perfect, since T3 and T4 have different pharmacokinetics and dynamics, but they’d be better if you’re trying to prove that the TSH becomes unreliable in the presence of exogenous T3. 

In defense of the clinical investigator, he wasn’t trying to prove that T3 medication acts differently on the pituitary than does T4 medication. But the author of the piece I found on the internet seems to be hanging her hat on this study, as if it’s earth-shattering evidence. I don’t mean to sound overly dismissive of her article, as she does approach the topic with intellectual rigor, but I think the author has completely missed the mark in trying to apply this study to her hypothesis.

And finally, for what it’s worth, not a single subject in the study — control or intervention group — demonstrated suppression of the TSH to below the normal range with any of the medication doses. Frankly, it makes me wonder if some of the subjects were skipping doses or taking the T4 with food or other substances known to interfere with absorption, as I’m surprised that non-hypothyroid people taking 30mcg T3 + 120mcg T4 didn’t exhibit suppressed TSHs. But whatever, all it does is make me even less certain that this study reveals any important secrets about thyroid hormone economy. What this study does do — and all it really does — is show how healthy controls and hypothyroid people react to infusions of TRH while under the influence of different doses of thyroid hormone. Full stop.

More Evidence, Same Problem — Irrelevant and Weak

The author of the internet article then turns her attention to another study by the same clinical investigator. In this study, the researcher looked at what happens to T3, T4, and TSH levels in healthy subjects without hypothyroidism, during overfeeding vs taking exogenous T3. What did he find? In overfeeding, T3 went up a bit, while T4 and TSH remained about the same. While taking T3, T3 went up a bit, while T4 and TSH went down a bit.

I will allow for the possibility that I may be missing something here, but while these findings may have been interesting when the study was published in the 1980s, they’re eliciting barely a yawn from me today.7

The regulation of thyroid hormone economy in healthy people is a complex process that involves but is not limited to: thyroid hormone synthesis and release; transport in the bloodstream; transport into the cell; transport into the nucleus; binding to nuclear receptors; conversion to active and inactive forms by deiodinase enzymes; local regulation of thyroid hormone activity by deiodinase enzymes; macronutrient composition of the diet and fasting; metabolic requirements; etc.8

My point is that comparing what happens to thyroid function tests during a physiologic stimulus like overfeeding and an exogenous therapy like giving T3 pills is apples to oranges. It’s not that we can’t learn interesting things about the hypothalamic-pituitary-thyroid axis by stressing the body in different ways — we totally can. In fact, if you follow the link in the prior paragraph, you’ll get to read about how ketogenic diets produce some interesting derangements of thyroid function testing that appear to have absolutely no clinical significance (to the best of my knowledge).

The fact that giving 10mcg or 20mcg of T3 (doses used in the above study) caused the TSH to go down a bit does not mean that T3 pills have a suppressive effect on TSH secretion that is out of proportion to the effect of endogenous T3 on TSH secretion. Think about it this way: when you give T3 to someone, you are intervening at one of the final steps in the normal pathway of thyroid hormone metabolism. In other words, the body’s deiodinase enzymes normally convert T4 to T3 within the body’s tissues. When you give T3, you’re jumping all the way to that step, fairly late in the process. Because of normal feedback, this T3 is going to have some sort of suppressive effect on TSH secretion, which will in turn lead to lower T4 secretion by the thyroid.

But when you overfeed someone, you are sending all sorts of signals to the body: there’s an overabundance of available energy, there’s an acute need to increase metabolic rate, etc. Because the hypothalamic-pituitary-thyroid axis is involved in this sort of stuff, you will see a response by the hypothalamus and pituitary that will ultimately be manifested by some changes in thyroid hormone synthesis and action — some of which can be measured and many of which can’t. Note that the overfeeding stimulus is proximal to the whole pathway of thyroid hormone synthesis. It’s a totally different kind of intervention, so it simply can’t be used to tell you how the pituitary should respond to a dose of exogenous T3.


Given that I couldn’t find a single study that was actually designed to determine whether exogenous T3 has an outsized suppressive effect on TSH when compared to an equivalent dose of T4, I have to say that we simply don’t know the answer to this question. And given that the alternative medicine community promotes the idea that pig thyroid should be titrated to relief of symptoms, even if that means suppressing the TSH to very low or undetectable levels, I would like to know the answer to the question. Unfortunately, from what I’ve seen so far on the internet, no one has offered a compelling argument in support of the concept that hypothyroid people on pig thyroid are more likely to have suppressed TSHs while being adequately replaced, when compared to their counterparts on T4 monotherapy.

If you have a scientific reference that you believe supports this concept, please note it in the Comments section below.


By using this site and interacting in the Comments, you agree to abide by my Disclaimer.  Please keep your comments respectful and refrain from ad hominem attacks.  In the past, I’ve been permissive — no longer.  The social and political discourse in our country has become so toxic that I cannot, in good conscience, allow my blog to provide a space for those who simply want to express outrage.  If you disagree with something I’ve written, or something written by a fellow reader, fabulous.  Make your argument in as dispassionate a way as you can, and we’ll all get along just fine.


Image credit: Photo by Volkan Kaçmaz on Unsplash

  1. Not her real name, but bonus points to the first person who gets this reference and notes it in the Comments. []
  2. IMHO one of my greatest feats of debunkery is Everything You Never Needed to Know About Reverse T3. For an extra chuckle, go to that post and read Samantha’s comment. It was low-hanging fruit, but it was just too delicious to not take a bite. []
  3. In this circumstance, exogenous refers to T3 hormone that you ingest, as opposed to endogenous T3 that your body produces on its own. []
  4. I will allow for the possibility that there may be physicians in other countries who have more experience with TRH testing, but given the lack of well-standardized guidelines for interpreting the results, I would still question the utility of the test. []
  5. The first medication combination is close in equivalency to a 90mg dose of Armour thyroid, for your reference. Let’s table the fact that I wouldn’t really call any of these medication combinations “small doses,” especially if giving them to control subjects without hypothyroidism. []
  6. While different sources will give different estimates for how to convert a T3 dose to a T4 dose, one rule of thumb is a roughly 1:4 conversion. Therefore, 15mcg of T3 should have similar strength to about 60mcg of T4, meaning that the first medication combination should be equivalent to roughly 120mcg of levothyroxine (T4). I can tell you, however, that in real life, these conversions often/usually don’t play out consistently. Sometimes I see more of a 1:3 relationship between T3:T4, sometimes 1:5. It depends a lot on dose preparations, dose timing in relation to meals, number of doses per day, etc. []
  7. Again, I don’t mean to denigrate the author of the internet article; rather, I’m trying to point out that the research she has cited cannot be used to support her thesis. []
  8. In the clinical investigator’s defense, the discovery of deiodinase enzymes came after he published this study, so there was a whole lot he didn’t know at the time of his research. []

8 Replies to “T3 Controversies: Is TSH Reliable when Taking T3?”

  1. Hi HD, good to see another post 🙂

    For papers, the only ones I know about this are actually in the opposite direction. Apparently, there are some studies in mice and cell cultures suggesting that T4 might be more potent than T3 at suppressing TSH. Something to do with how the D2 deiodinase is regulated in the hypothalamus & pituitary.

    [1] Atypical Expression of Type 2 Iodothyronine Deiodinase in Thyrotrophs Explains the Thyroxine-Mediated Pituitary Thyrotropin Feedback Mechanism , by Christoffolete et al. https://doi.org/10.1210/en.2005-1300

    [2] Differences in hypothalamic type 2 deiodinase ubiquitination explain localized sensitivity to thyroxine, by de Castro et al. https://doi.org/10.1172/JCI77588

    I stumbled across these in a presentation at the Thyroid Cancer Survivor’s Association, talking about TSH suppression. Which reminds me… If T3 pills really were a risk-free way to suppress TSH, one would expect that thyroid cancer treatment would have been one of the first applications.

  2. This is an interesting discussion. From what I’ve seen discussed online, it’s not uncommon for people to have suppressed or very low TSH’s on NDT (and they would be under-dosed on higher TSH’s). I am of the opinion that each form of thyroid hormone replacement option has different ways of interpreting lab results for optimal replacement (T4 only, T4 and T3, NDT, T3 only). I believe that when on T4 only, TSH generally works well (under 2.5 being a goal), but I still like to see that FT4 and FT3 look fairly normal (not a high FT4 and low FT3). I think most people feel poor because they are under-dosed and have high “in range” TSH labs. People on thyroid medication should have a lower range in my opinion than what many labs have listed. Getting my TSH below 2.5 was a key to me becoming normal again. But in general I trust TSH on T4 only despite the alt med angst against TSH. But on the other types of hormone replacement it seems like the numbers change a bit, especially for TSH. Does that mean TSH is useless as people often claim, I don’t think so. Even if TSH is lower on NDT, maybe instead of having a 1.5 on T4 only, it would be best at .3 on NDT. Honestly I don’t think there’s been enough study on dosing NDT, T4/T3 synthetic, T3 only (which seems like a last resort option). Maybe that’s because a large amount of providers are anti T3 all together? There should be many many more official studies looking at combination therapy labs and dosing. Even the Guidelines for the Treatment of Hypothyroidism paper said there should be a lot more studies done on these items, but I don’t see enough being done on this. Where this hurts patients is they are left to rely on alt med dosing and lab guidance which as you know has plenty of issues.

    1. Just to put a fine point on this, although many people online may say that they are adequately replaced (not over-replaced) on NDT with a low TSH, there don’t seem to be any good data to support that assertion. I agree that it would be very helpful to see a study designed to look at this issue. IMO it should be a crossover design in which each subject serves as his or her own control. This would get around the problem of many thyroid studies showing no overall differences between groups on different treatments. We know that individuals seem to vary a fair amount regarding at what TSH they feel optimal, so I think the crossover design makes sense.

      1. I get the impression there’s been a lot of issues with NDT quality control over the past few years. I’d rather do synthetic T4 and T3 at this point if I were in the market for trying T3 (which I am not currently). Seems like there needs to be more studies on T3 dosing/types as was suggested in the guidelines from 2014. But I have not seen many studies done despite the call for them. It’d be a lot better for patients to see some science backed research rather than relying on random people in a facebook group full of conflicting advice (LOTS of conflicting advice).

  3. I’m trying to learn all the time and so have some questions. I’m a post RAI Graves patient on combo therapy. My labs on 75mcg of T4 with 7.5mcg of T3 show a TSH of .83, and yet a dose of 88mcg of T4 with 5mcg of T3 make it 1.15. Oddly, the ‘total dose’ is lower if we are using a replacement ratio of T3 to T4 of 1:3 in the 75/7.5mcg dosing. I see here you’ve noted,

    “Next, it is critical to note that the pituitary’s response to an artificially-induced stimulus like exogenous TRH cannot be extrapolated to interpret how the pituitary would respond to a normal physiologic stimulus. What I mean to say is: walloping someone’s pituitary with a big dose of TRH should not be thought to reflect the normal cross-talk between the hypothalamus and pituitary when responding to changes in ambient T4 and T3 levels.”

    Is it possible that the pituitary is also not used to having its thyroid hormone served up in a non ” normal physiologic stimulus”? For example, having say, 75mcg of T4 released into the system all at once, or 7.5mcg of T3 released into the system all at once could be considered an abnormal physiologic stimulus, could it not?

    I understand that thyroid hormones in a normal individual are slowly secreted from gland throughout the day, and that would be a normal physiologic stimulus, correct? How do we know the pituitary doesn’t find a full dose of thyroid replacement hormone a ‘non physiologic stimulus’?

    thank you so much for your answer, if you do take the time/have the time to answer.

    Darcey Lyn

    1. I think I understand what you’re asking, but please let me know if I miss the boat in my response. Note that I’ll be responding to the gist of your question, rather than addressing your individual medication regimen and lab results.

      If you’re asking whether the pituitary may react differently to exogenous T4 and T3, as opposed to endogenous T4 and T3, the answer should be a qualified yes. For example, if someone took a single 5mcg dose of T3 in the form of a pill, and you were able to insert a catheter that would continuously sample the venous blood draining from the pituitary, I would guess that we would see a small decline in TSH levels for some period of time, after the dose had been taken. Contrast this to the normal thyroidal production of 5mcg of T3 over the course of a day, and yeah, I think it’s reasonable to assume that the TSH might look a little different for some portion of the day, between these two situations. The problem is that the evidence proving this, as well as the evidence showing that the change in TSH would be clinically relevant, is lacking.

      As for exogenous T4, I don’t worry quite as much about that being non-physiologic, as the half-life of the medication is pretty long, creating an available pool of T4 that can be converted to T3 as needed, based on the conditions in the body at any given time.

      Again, if I missed the mark in my answer, please let me know.

      1. Thank you so much for taking the time to answer my question. I appreciate it immensely because this is a difficult world to navigate without a charter. Combination therapy (not counting the historical use of NDT) is relatively new for a large portion of thyroid patients. Personally, I have been on combination therapy for many years, but it hasn’t been done properly.

        There doesn’t seem to be a consensus on how to do it effectively. And what I can’t find anywhere, no matter what study I read, is when to test the TSH. During T4 only therapy it’s relatively simple. You take your T4 in the morning for most people and you test your TSH the next day in the morning at baseline or trough. With T3 dosing added to the mix, many people are breaking their doses up throughout the day because that’s what’s frequently recommended, even for doses of 5 µg.

        As you have indicated, that could inappropriately suppress TSH soon after the dose because it is not the normal physiological way the pituitary would be exposed to thyroid hormone (T3).

        I know you can’t comment on my own dosing or medication regime and I don’t expect you to, but I definitely have a question about when to test TSH. And that I think you can help me with.

        I have been doing my thyroid labs 12 hours after the last dose of T3 is taken. I am not on a high dose of T3, and I think that this 12 hours of timeframe is making my TSH look much lower than it would ordinarily be from the combined amount of medication that I’m taking (using a 1:3 T4 to T3 ratio) if it were T4 only and if I were to test in the standard way the next morning.

        Do you recommend that the whole dose of thyroid medication be taken in the morning the day before testing, and the labs run the following day after no thyroid medication has been taken for 24 hours? Is that the standard that I should be following?

        Honestly, I have looked at so much information trying to find the answer to this very simple question and I would be thrilled if you could provide me with the answer.

        1. I have an answer to the general question of when the TSH should be tested, but it’s not a totally satisfying one. In people with mild hypothyroidism (generally not people who have had I-131 ablation or total thyroidectomy), the circadian rhythm of TSH secretion may be preserved. This means that those folks, who are probably on smallish doses of T4 +/- T3, may exhibit a slightly higher TSH in the early morning and slightly lower in the late afternoon. Whether the delta between the morning and late-afternoon TSH is clinically relevant is up for debate.

          In people with more complete hypothyroidism (I-131 ablation or total thyroidectomy), that diurnal rhythm most likely will be lost, so the time of day itself becomes a little less important.

          But you’re asking mainly about when to check TSH in relation to when the meds are taken, and that’s where the answer becomes kind of unsatisfying. There are some studies looking at what happens to TSH in the 12-24 hours after giving T3 to a hypothyroid person. The data of which I am aware used doses of T3 that I normally wouldn’t give in one dose (10-50 mcg), so it’s kind of hard to extrapolate the data to someone taking, say, 2.5mcg twice or 3x daily. Anyway, some studies found no change in TSH in the 12 hours or so after a big dose of T3, and at least one found a slight decrease. As for T4, given the steady state concentration of that long half-life med, I wouldn’t worry so much about it.

          So, what to do about checking TSH in someone on a reasonable dose of T3 (which I usually consider to be a total daily dose of less than 10mcg, divided into 2-3 doses)? I would just be consistent (e.g. always check in the morning, fasting, before meds; or always check in the late afternoon after all morning meds have already been taken; etc). If someone is really into self-experimentation, they could try checking the TSH at different times of day, on the same medication regimen, just to see if there are meaningful differences.

          But I think the bottom line is that consistency is probably the best idea. And I guess I should also clarify something else: When I mentioned in my previous response that a cannulated vein near the pituitary might show a subtle decrement in TSH production after a dose of T3, I was speaking theoretically and without much to back it up. I really am not sure that there would be a clinically meaningful drop in serum TSH after a single T3 dose, even in the setting of a somewhat larger T3 dose.

          Hope this helps, though I understand the water is probably still kind of muddy.

Leave a Reply

Your email address will not be published. Required fields are marked *

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