High-Dose Thyroid Hormone As a Mood Stabilizer in Bipolar Disorder

Table of Contents

This webpage describes the use of thyroid hormone as a treatment for rapid cycling bipolar disorder. If you did not arrive from my main page on Thyroid Hormone and Bipolar Disorder, you could start there for some basics about this hormone and why it is relevant for this illness.

To summarize the information below: several independent research groups have been studying this treatment approach. The results are not conclusive, primarily because we need several more studies to confirm the initial results. Only one so far is a “randomized trial” comparing results versus a placebo treatment. But the results so far are quite promising, and the risk levels appear to be very low, so this approach is worth at least knowing about. For patients who have tried many other approaches, this represents one more alternative. Because the risks appear to be so low, it might even be worth considering early in the usual progression through standard mood stabilizer medication options. More data on its effectiveness would be nice, of course.

This high-dose approach has been described by a research group which has done most of the investigations, in a review article which provides a nifty summary if you are trying to get references to your doctors. Because otherwise your job will be quite difficult in this respect, I have taken the liberty of giving you a direct link to this document, which follows my summary below. All errors in interpretation are my responsibility. (If anyone objects to my having made this paper available in this fashion, please let me know and I will remove the link.)

The Hardest Evidence: a randomized trial

The only randomized trial of this approach was published in early 2014.Stamm Bottom line: high-dose thyroid was better than placebo, but only in women, not in men. And (very satisfying, I’d been nursing this hunch for years) it worked better in women who started the trial with a high TSH.  This is hard evidence, because of all the work it took to produce this study with no big money behind it. At the rate it took to get these data, we’ll not likely seen another study like this for quite a while.

Skip to the next section, Who Might Benefit,  unless you really want the details of this study.

Here are their results, first for women and men together; then for women only, then men only. Note the remarkable placebo response among the men, wiping out any difference with thyroid. When men’s results are mixed with the women’s, the men’s placebo response limits the “statistical significance” of the results in the overall group (A).




There are plenty of quibbles possible here. First, the sample size is so small, maybe that’s why the results are not more strikingly significant. Imagine what these results might look like if the authors had been able to recruit hundreds of patients, as was done in the (industry-supported) study of quetiapine. Looks to me like the trend here is very strongly positive, just as quetiapine was; the lack of “statistical significance” can be attributed to the small sample size — as long as one presumes that a larger sample would have shown the same pattern of results.

The dosing: 100 mcg daily in week 1, 200 mcg/day in week 2, 300 mcg/day in week three to six. (Really). Patients were all on mood stabilizers, optimized by blood level; over half were also on antidepressants.

Negative effects: of the 31 subjects receiving thyroid, one had mild hyperthyroidism, one got a rash, and one became manic. None had ECG changes, increases in blood pressure, nor weight loss.

Conclusion: at least in women, this strategy has been shown to be superior to placebo in a randomized trial. Based on these new data, but also on some complex thyroid physiology, the authors are much less enthusiastic about the likelihood of response to this treatment in men.

Who might benefit?

Most of the open-trial data on this treatment approach has focused on patients with rapid cycling bipolar disorder. The Berlin randomized trial described above focused on bipolar depression. Treatment-resistant” depression, whether bipolar or unipolar (a very difficult distinction in this group of patients), may also respond to this approach.

What are the risks, and side effects?

Surprisingly, this approach is generally tolerated very well. You would think that people would become hyperthyroid, which would lead to the following symptoms:

  • heat intolerance (hot when others are comfortable, sweating when no one else is, wearing fewer clothes than others)
  • loose stool or diarrhea
  • irregular menses
  • tremor
  • feeling “wired” or agitated
  • increased resting pulse rate

But that does not seem to happen for patients who do well with this approach. Pulse rates to go up, generally about 10-20 beats per minute. Some minor symptoms from the above list were noted by some patients, but not to the point of requiring a treatment change. The researchers note that in patients who do not have bipolar disorder, high-dose thyroid would be expected to produce
substantial levels of these symptoms.

Heart risk

Becoming hyperthyroid on your own, because your thyroid gland is not working properly, has been associated with an increased frequency of an abnormal heart rhythm called “atrial fibrillation”. This is not lethal, but it is uncomfortable. Heart pumping capacity decreases by about half, and people feel quite strange. They often go to the emergency room, where generally this abnormal rhythm can be corrected.

But does taking thyroid hormone cause this same increased frequency of atrial fibrillation? in other words, does it make a difference, when the thyroid hormone comes into your body from the outside, instead of when you make it (due to a hyperthyroid state) from the inside? This is not known. All we can say at this point is that in following the patients who are using this high-dose thyroid approach, the researchers have not seen episodes of atrial fibrillation. Of course, at this point they may not have seen enough patients to catch the few cases which might arise from this treatment, so we cannot yet say that thyroid hormone from the outside is different in this respect.

Bone risk

To reiterate from my Basics page:

Risk #2, Bone: This only applies if you stay on high doses of thyroid treatment, presumably because you and your doctor have decided that the treatment is working really well. If you try the treatment and conclude in a month or two that it is not working, this risk is not an issue. A risk that has been associated with staying hyperthyroid due to an overactive thyroid gland is osteoporosis loss of calcium in your bones, with a risk of fractures, especially when you get older.

But so far, In a 5-year follow-up of patients being treated with high doses of thyroid hormone for bipolar disorder, this was not a problem. Bone density decreased, because the average age in this study was 50, and we all are going down at that age (so to speak; it’s the voice of experience…). But there was no more bone loss than was seen in people of the same age and gender who were not being treated.Ricken

A massive review of high-dose thyroid treatment and bone density also concluded that the risk is either non-existent or at least lower than that of multiple other common treatments for bipolar disorder.Kelly For more  see details of bone risk in thyroid treatment.


Once in every 15 or 20 patients or so it seems that even low doses of thyroid hormone cause a feeling of agitation and muscle tension and anxiety. This goes away when the dose is lowered or the thyroid is stopped; it takes about 3-4 days to fade away. Very unpleasant. This is a short-term risk I now warn my patients about. I’ve seen it even when the patient’s TSH is quite high, meaning we’re not seeing this agitation because the thyroid dose is “too high”.  For example, one patient has a TSH around 4 and cannot take even one quarter of a 25 mcg T4 pill without getting this agitation thing.  I have several patients like that.

A case reportRao describes a patient who had agitation and depression at the same time, associated with being hyperthyroid (symptoms resolved when her own thyroid hormone production was lowered).   So it seems that in addition to inducing hyperthyroid symptoms, high doses of thyroid hormone could induce a “mixed state”, or something like it. Unfortunately, mixed states are relatively common in people with rapid cycling, the very ones who might theoretically benefit from the high-dose thyroid strategy. Neverthelss one must keep in mind the possibility that if a person is on the high-dose approach and having “mixed state” symptoms, this could be from the thyroid dose.

Tests to do before starting

Several groups of patients would not be suitable for this treatment approach, and so must be sought before treatment begins. These include:

  • High thyroid levels (hyperthyroid)
  • Cardiac history (previous heart attack, severe high blood pressure, previous abnormal rhythms, low cardiac output (“heart failure”))
  • Pregnancy or breast-feeding
  • Already low bone density, or age greater than 70, or dementia

The researchers also recommend some procedures which may require modification for use in the real world, as opposed to a research setting. For example, they recommend a measure of bone density (an expensive test) before treatment begins, whereas I think it makes more sense to limit this to women who have other reasons for concern about bone density, or at least those who do so well on the treatment that it will be continued long-term. An electrocardiogram is recommended for patients with a history of heart rhythm problems, and a consultation with an endocrinologist/internist for patients with a history of thyroid abnormalities.

Doses used for starting

Which form: T3 or T4?

As you know (if you don’t, read my Thyroid Basics page), there are two forms of thyroid hormone, T3 (triiodothyronine, liothyronine/Cytomel) and T4 (levothyroxine). Both of them have been used in this “hypermetabolic” treatment of bipolar depression. The randomized trial referred to above used T4. You could assume that because this has been the form which has been more widely used in research (namely the UCLA group and their spin-offs), that is the form which should be considered for this approach in the hands of most average psychiatrists.

However, a remarkable pioneer, Dr. Tammas Kelly, in Colorado, has gathered a huge number of patients that he has treated with the other form, T3. He is carefully analyzed the results achieved with this approach, and published them a well respected psychiatric journal (Journal of Affective Disorders).Kelly He clearly has one of the largest patient samples ever studied with this approach.

While I have generally used T4, following the UCLA approach, Dr. Kelly is strongly in favor of T3. There has never been a direct comparison of the 2 approaches. T3 has some significant advantages. Indeed, a recent large federally funded study found T3 better than lithium for treatment-resistant major depression: same degree of benefit, with less side effects and need to discontinue on that basis.STAR*D

So, which one to use? at this point, it’s rather a toss-up. If you saw Dr. Kelly, he’d give you T3. If you saw me, I’d give you T4, because that’s the one I have more experience with.

T4 doses (the UCLA review description)

  Starting Dose(mcg per day) Dose Increase(mcg/day)
Normal thyroid(TSH in the normal range) 100 100/week
“Subclinical” hypothyroidism”(TSH elevated) 25-50 Slow*
Overt hypothyroidism(TSH elevated, T4/T3 low) 25-50 Slow*

*reach euthyroid status in 6-8 weeks

Lab tests during follow-up

In this high-dose approach, think about what will happen to TSH. If you followed the story about TSH lab result interpretation (How is Thyroid Measured?, here), you will understand that high dose thyroid treatment drives TSH down to very low levels, zero or close to zero. When starting this treatment, everyone should be prepared to see the TSH go that low. This should not be a surprise, and it should be accepted as part of the plan.

What about the other lab tests, like the T4 level, or T3 level? In the article linked below, the researchers with the most experience using this approach recommend allowing an increase in T4 up to 150% of the starting T4 level. In a personal communication, one of the researchers talked in terms of “150% of the upper limit of normal“, which is obviously likely a little higher number. As for T3, if we had our good reliable measure of this one hormone, especially one that was cheap, this might be the best marker of all, as it is closer to what might be the important physiological endpoint. The researchers point out that if this one is not elevated, perhaps the patient is not really hyperthyroid?

To summarize this issue: there is no accepted laboratory marker indicating the upper limit of this process if the patient him or herself does not have signs or symptoms of hyperthyroidism. This is still being worked out in research. Obviously what matters is to prevent bad outcomes that might be associated with “too high a dose”. But how can we say what it is too high, except based on bad outcomes or on some understanding of the physiology? (as you’ll see below, not much of the relevant physiology is understood)

Think about it this way: if the patient does not have signs and symptoms of hyperthyroidism, but the laboratory studies are in the “hyperthyroid” range, who is correct — the lab results, or the patient? It is such a person really “hyperthyroid”? Is she or he really at increased risk of atrial fibrillation and decreased bone density?

Remember, these risks were originally recognized in patients who were “naturally hyperthyroid”: they were made hyperthyroid by their own gland, not by someone giving them thyroid hormone. We do not yet know if high-dose thyroid creates the same risks, presuming that the patient is not symptomatic. For now, I think we can summarize by saying that there is no consensus on how to judge how much thyroid hormone is too much. Going to 300 mcg, as used in the randomized trial in Berlin, seems a reasonable upper limit for now. In cases where many other approaches have been tried, going to 400 mcg or even a little higher seems justifiable. I personally will be using “T4 up to 150% of the upper limit of normal”.

If this seems radical, or daring, or even “Western cowboy style” medicine, think about the kinds of risks to which we expose patients when we use a medication that has just been approved by the FDA. How shall we define the upper limits of acceptable treatment? What are the risks of going to that level, or beyond? how much risk are we taking just getting to levels that manufacture has already deemed acceptable? When a medication is new, these are almost completely unknown — perhaps even more unknown than the risks of the thyroid approach discussed here.

How does this work?

Seems like a strange idea, using these high doses. Does anyone know how this ends up having a mood stabilizer effect? Basically, the answer is no. Remember, T3 is the active version of thyroid hormone. The researchers point out that when someone becomes spontaneously hyperthyroid, from their own thyroid gland overproduction, they have increased levels of T3 as well as T4. When T4 thyroid is given as a pill, from the outside, even at high doses, T3 levels are not abnormal. In this respect, the patient is not “hyperthyroid”, at least in the same sense as if TSH went to zero from his or her own thyroid production. Thyroid hormone control is extremely complicated. Using high doses of thyroid is simply bumping a system we don’t understand very well. Unfortunately, this is what similar to the situation we face using nearly any medication for bipolar disorder (although our understanding of the causes of bipolar disorder is improving rapidly).

How to Explain This to Your Doctors

Most primary care physicians and endocrinologists have never heard anything about this high-dose approach. But some physicians have been using thyroid hormone for patients with mood problems for years, without a clear rationale. In their opinion, it just seemed to work and not cause too much trouble. Because there was no research behind this approach, it was thought to be almost irresponsible, at least by some endocrinologists. Therefore, if you are considering high-dose thyroid, you are very likely to run into resistance to the idea from other doctors. You can send them to this webpage, if they will go, but for many doctors you might need to walk in with a review article from a source with very good credentials on this issue. Because you could not easily put your hand on this otherwise, and you are so likely to need it, I offer you a link below to a PDF you can print. (Again, if anyone objects to my having made this available, please let me know and I will take it down and post an explanation of the objection.)

The Article You Will Need to Take to Your Doctors

(updated 11/2014)

Want to know more? Try Dr. Kelly’s 2018 book, The Art and Science of Thyroid Supplementation for the Treatment of Bipolar Depression.

Do Benzos Treat Depression?

Around 50 controlled trials have tested benzodiazepines in depression, and the results are surprising. In this episode we look at the controversies surrounding this research and what it means for practice.

Read More »

Get Smarter About Mental Health

Our Brain Bulletin decodes mental health updates for you.

It’s free.