TSH (Thyroid Stimulating Hormone) levels are used by American endocrinologists to diagnose and treat thyroid problems. The following is from the first paragraph of the American Association of Clinical Endocrinologist Medical Guidelines:
The sensitive thyroid stimulating hormone (TSH or thyrotropin) assay has become the single best screening test for hyperthyroidism and hypothyroidism, and in most outpatient clinical situations, the serum TSH is the most sensitive test for detecting mild thyroid hormone excess or deficiency. 
This is not logical. TSH is a pituitary hormone, not a thyroid hormone, so TSH is an implied measurement of thyroid levels. There are too many instances where TSH levels fall below the reference range (which implies hyperthyroidism), when the patient is actually clinically hypothyroid or normal, but certainly not hyperthyroid. Many have Free and Total T3, and Free and Total T4 levels that are within range with low TSH levels, which contradicts this paradigm. Because of this pseudo-suppression, low TSH levels really should not be used to diagnose hyperthyroidism without confirming the Free T3 and Free and Total T4 levels, nor should low TSH levels alone ever be a reason for decreasing one’s dose. To make dosing decisions based solely on TSH levels is a form of paradigm paralysis, which is a fixation on the current paradigm even in the face of overwhelming proof that the paradigm is flawed. Could this flawed paradigm be why so many with “normal” TSH levels have so many hypothyroid symptoms?
A TSH can sometimes be useful. Low TSH levels (near zero) with hyperthyroid symptoms may be the first indication that someone (who is not on any thyroid medication) has Graves’ disease. Further antibody testing will confirm that diagnosis. When a TSH is greater than 1.0, it may indicate some degree of hypothyroidism for someone already taking thyroid hormone who is still symptomatic. A higher dose may be beneficial if other factors like iron, cortisol, blood sugar, etc. have already been addressed. The current TSH reference range probably goes too high because it includes undiagnosed hypothyroid patients. Data has shown that African-Americans, who have a very low incidence of Hashimoto’s thyroiditis, have an average TSH of 1.18. This may be more representative of the true normal mean TSH of a normal population.  Another study found increased arterial stiffness (which correlates with heart disease) in subjects whose TSH was 2.01-4.0, which is within the normal range. The authors concluded that “it may be proposed that TSH values between 2.01 and 4.0 µU/ml are rather mildly abnormal and not high-normal and that the definition of the “normal” range for TSH values should probably be reconsidered.”  In any case, dosing adjustments should never be based on TSH levels alone, for some of the reasons listed below.
The TSH in people with a dysfunctional hypothalamus or pituitary gland will never accurately reflect their actual thyroid levels.  This is common in people who have had head injuries from sports, car accidents, etc. Has anyone not bumped their head at some point, ever?
Graves’ disease patients who have had radioactive iodine treatment (RAI), a thyroidectomy, or who take anti-thyroid drugs (ATD) and now take replacement thyroid medication, should not be dosed by TSH. Graves’ is caused by TSH Receptor antibodies, and TSH levels can stay suppressed (near zero) for months or longer, even after treatment brings thyroid hormone levels down into the normal range or even below normal.  If they are unable to get adequate thyroid hormone replacement because of their suppressed TSH levels, these patients will suffer from hypothyroid symptoms. They would fare much better by adjusting their dose to Free T3 and Free and/or Total T4 levels, instead of the TSH. [optimal thyroid levels]
Also, thyroid ablation (destruction) does not remove the Graves’ antibodies from the blood, it merely eliminates the organ (thyroid) that was under attack. If the antibodies remain high, TSH levels will continue to be suppressed, no matter how low the actual thyroid levels are, and other organs like the eyes and skin can be attacked. [Graves’ disease]
Hashi’s patients have different antibodies (thyroid peroxidase and thyroglobulin) and often cycle from hyper to hypo, which can cause huge fluctuations in TSH levels. Eventually the weakened thyroid stays in a hypothyroid state, so patients generally feel better when placed on enough thyroid hormone to keep their TSH levels suppressed. The suppressed TSH reduces stimulation to the thyroid, which results in lowered antibodies, and stops the continuous hyper/hypo fluctuations. [5,6]
Any thyroid medication that contains T3, like Cytomel or desiccated thyroid, will suppress TSH levels.  One study (on rats) determined that the relative potency of T3 to T4, in terms of its ability to inhibit TSH, was 100:12.  The direct T3 in the blood is sensed by the hypothalamus/pituitary, the body determines that no additional thyroid is needed, so no TSH is released. A normal thyroid secretes only a minimal amount of T3. The majority of the body’s T3 is converted from T4 as needed throughout the day, whereas someone on T3 or desiccated thyroid takes a concentrated dose of T3 all at once. This may explain the TSH suppression. Surprisingly, patients can have suppressed TSH levels while taking any form of T3, but do not exhibit any hyperthyroid symptoms. Dosing by TSH will usually leave the patient undermedicated with hypothyroid symptoms.
A study of 832 hypothyroid patients showed that 24 hour urine Free T3 had the highest inverse correlation with clinical symptoms. In other words, the higher the urine Free T3, the lower the symptoms and vice versa. Urine Free T3 is not influenced by binding globulins and correlated well with the severity of eight clinical hypothyroid symptoms: fatigue, depression, coldness, headache, muscle cramps, constipation, arthritis, and Achilles tendon reflex. Serum T4, Free T4, and TSH often had no correlation to these symptoms. 
Suppressed TSH levels are frowned upon by doctors because they feel it can lead to pituitary atrophy. But a study where thyroid hormone was discontinued after long-term use and TSH suppression showed a return to normal levels within two to five weeks. Serum T4 also returned to normal at least four weeks after hormone withdrawal. 
One report claims that only metabolic measurements such as basal temperature can be used to determine an optimum dose, because those show the actual metabolic effects of the thyroid hormone at the cellular level. Serum T3 and T4 levels are still an inference because of thyroid hormone resistance at the cellular level. TSH is actually a second level of inference! There are some interesting charts in the report that depict the inverse relationship of TSH to exogenous T4 and T3 intake, and how the desired level of T3 and/or T4 is only attained when TSH falls below the reference range. 
Thyroid cancer patients on TSH suppression therapy did not show hyperthyroid signs and symptoms unless their serum Free T4 was also high. There was no correlation with TSH. The abstract states: “Although the degree of TSH suppression can now be exactly monitored with new third generation TSH assays, hyperthyroidism cannot be defined using TSH concentration . . . ” 
Because TSH is made by the pituitary in the brain, and the brain and the body are in two different compartments, TSH represents only the brain’s need for thyroid hormone. Problems in the nervous system, endocrine system, immune system, metabolism, and nutritional status can all affect thyroid signaling. This is known as non-thyroidal illness; there is nothing wrong with the thyroid gland itself, but the TSH signal is affected. 
Pregnancy, diabetes, trauma, renal disease, liver disease, sepsis, and cardiac conditions (heart transplant or bypass) all impact TSH and make TSH levels a poor diagnostic tool. [13,14,15] And this list is by no means complete!
TSH displays a seasonal variation in healthy people, with TSH levels lowest in the spring. The variation from the mean TSH (average of 12 monthly TSH values) was 29.1%. 
TSH decreases when fasting.  Most patients do their lab tests in a fasting state, because other labs like glucose and cholesterol require it. But this may result in an artificially low TSH that does not reflect true thyroid levels. In fact, TSH has a circadian rhythm, with a peak around midnight (with much variability between individuals), and a low in the afternoon; fluctuations are normal. The change in TSH from peak to trough is approximately 72%. Free T3 levels also show a similar circadian rhythm (with a smaller amplitude) with a time lag of approximately 90 minutes behind the TSH curve. The Free T4 curve did not follow the TSH curve at all.  Labs drawn in the morning could be significantly different from labs drawn in the afternoon after lunch, with one TSH in “normal” range and the other in the hyperthyroid or hypothyroid range. How can TSH be a valid diagnostic measure when it varies so much?
A study compared early morning fasting serum TSH levels to late morning non-fasting serum TSH levels in the same patients on the same day. In 97 of 100 subjects, the late morning non-fasting TSH tests declined by an average of 26.39% when compared to early morning, fasting, TSH test results. This meant that 6% of patients who were earlier diagnosed as subclinical hypothyroid were now reclassified as “normal.” Since time of day and fasting status of the patient can significantly affect serum TSH test results, a diagnosis based solely on TSH, a value that fluctuates, is questionable. 
Thyroid hormone resistance syndrome is a condition where the body’s tissues are resistant to the effects of thyroid hormone. In Generalized Resistance to Thyroid Hormone (GRTH), someone can have elevated serum thyroid hormone levels but normal or elevated TSH levels, when a suppressed TSH would be expected. These people are usually clinically euthyroid and require no treatment.  This is just another example where the TSH cannot be used to make a diagnosis.
A TSH near zero can’t be healthy, say most doctors, because it means the patient is hyperthyroid. Healthy people do not have a TSH that low. There are actually TSH receptors throughout the body , not just on the thyroid, so intuitively, TSH must have a purpose. TSH’s role may be to stimulate T4 to T3 conversion at the cellular level, since each organ has its own T3 requirement, which may be higher or lower than other tissues. In a lab experiment on isolated rat liver and kidneys, adding TSH had a positive effect on T3 levels. The release of T3, tissue T3 production, net T3 production, and the conversion rate of T4 to T3 in both the rat liver or kidney (perfused with 250 μU/ml TSH) was significantly higher than those of the controls that did not have the additional TSH. [25,26]
A statistical analysis that compared correlation slopes of log TSH vs. FT3 and FT4 between untreated patients, and patients taking T4-only, showed marked differences between the two groups. In patients taking T4-only, higher doses of T4 were needed to bring T3 levels up to the reference range. Apparently, T4-only does not mimic normal thyroid physiology because there is a disconnect between FT4-TSH feedback and T3 production. 
If I had to prioritize the three hormones (T3, T4, TSH), I would say T3 is absolutely essential, and one could not live without it. In fact, it is preferable to give someone T3 to bring them out of myxedema coma (near-death hypothyroid state), rather than T4. In one study, 50% of the patients presenting with myxedema coma died while given T4.  In another study, all patients survived, and they were given T3 first, followed by T4.  One could hypothesize that these people had poor T4 to T3 conversion to start with, which is what caused their condition, or that T4 is too slow-acting to quickly bring levels up in such a critical state.
While T4 has its own essential properties, it is possible to live without it by compensating with extra T3, but there are some serious side effects to the T3-only protocol. [T3-only side effects]
If the primary role of TSH is to aid in T4 to T3 conversion at the cellular level, then someone with low TSH may have insufficient T3 levels. In fact, many patients initially report a low TSH with low T3 levels, but T4 levels close to mid-range. On desiccated thyroid, a suppressed TSH is a common side effect, because the T3 content in desiccated suppresses stimulation from both the hypothalamus and pituitary. So anyone taking desiccated thyroid may have limited T4 to T3 conversion due to a lack of TSH. But desiccated thyroid contains T3, so the loss of conversion is compensated for by the same T3 that is suppressing the TSH. In any case, it is possible to live with a suppressed TSH, and many on desiccated thyroid do just that, because they find that keeping their TSH in range means they’ll still have hypothyroid symptoms.
In an ideal world, all our lab values would be somewhere within the reference range, and we’d feel great. My observation is that most people on T4 do not have suppressed TSH levels, but most people on desiccated do, probably because of the T3. But anecdotally, most who have been on both types of medications feel better with desiccated. Maybe the best of both worlds could be achieved by combining the two medications. By lowering the desiccated dose and splitting it up throughout the day, the TSH might rise into range. And since desiccated has too high a ratio of T3 to T4 for some people, combining the two would correct that problem too. This obviously would not work for people with conversion or other issues, but it might work for some.
In my case, I have to make a choice, because on desiccated thyroid I have asthma if my TSH is kept in range. Raising my dose by only ¼ grain suppresses my TSH, but completely relieves the asthma. My free and total T3 and T4 levels are not over range with this suppressed TSH. But because breathing is something I do 24/7, I have chosen to overlook the TSH. I’ve recently switched to a combination of T4 + desiccated, because my labs are very lopsided on desiccated alone, with upper range Free T3 and lower range Free T4. Improvements so far are more hair!
Getting your thyroid levels tested
If you’d like to have your thyroid levels tested, please ask for these thyroid tests, and note where your levels are in the thyroid lab ranges compared to healthy people. If you do not ask for these specific tests, your doctor will most likely just run a TSH test, which as described above, does not catch many cases of hypothyroidism.
Further discussion about the flaws of dosing by TSH is in chapter 25 of my book, Tired Thyroid: From Hyper to Hypo to Healing–Breaking the TSH Rule. In addition, five case studies illustrate how the TSH Rule can lead to poor treatment.
- American Association Of Clinical Endocrinologists Medical Guidelines For Clinical Practice For The Evaluation And Treatment Of Hyperthyroidism And Hypothyroidism. Endocrine Practice. Vol 8, No. 6, November/December 2002, 457-469. http://www.aace.com/files/hypo-hyper.pdf
- Leonard Wartofsky and Richard A Dickey. The Evidence for a Narrower Thyrotropin Reference Range Is Compelling. The Journal of Clinical Endocrinology & Metabolism Vol. 90, No. 9 5483-5488. 2005. http://jcem.endojournals.org/cgi/content/abstract/90/9/5483
- Steven A. Lieberman, Asra L. Oberoi, Charles R. Gilkison, Brent E. Masel and Randall J. Urban. Prevalence of Neuroendocrine Dysfunction in Patients Recovering from Traumatic Brain Injury. The Journal of Clinical Endocrinology & Metabolism. Vol. 86, No. 6, 2752-2756, 2001. http://jcem.endojournals.org/cgi/content/abstract/86/6/2752
- Chung YJ, Lee BW, Kim J-Y, Jung JH, Min Y-K, Lee M-S, Lee M-K, Kim K-W, Chung JH. Continued suppression of serum TSH level may be attributed to TSH receptor antibody activity as well as the severity of thyrotoxicosis and the time to recovery of thyroid hormone in treated euthyroid Graves’ patients. Thyroid 16: 1251-1257, 2006. http://www.liebertonline.com/doi/abs/10.1089/thy.2006.16.1251
- Aksoy DY, Kerimoglu U, Okur H, Canpinar H, Karaağaoğlu E, Yetgin S, Kansu E, Gedik O. Effects of prophylactic thyroid hormone replacement in euthyroid Hashimoto’s thyroiditis. Endocr J. 2005 Jun;52(3):337-43. http://www.ncbi.nlm.nih.gov/pubmed/16006728
- Max Rieu, Alain Richard, Myriam Rosilio, Sophie Laplanche, Veronique Ropion, Jean-Pierre Fombeur, Jean-Louis Berrod.Effects of thyroid status on thyroid autoimmunity expression in euthyroid and hypothyroid patients with Hashimoto’s thyroiditis. Clinical Endocrinology. Volume 40, Issue 4, pages 529–535, April 1994. http://onlinelibrary.wiley.com/doi/10.1111/j.1365-2265.1994.tb02494.x/abstract
- Appelhof BC, Fliers E, Wekking EM, Schene AH, Huyser J, Tijssen JG, Endert E, van Weert HC, Wiersinga WM. Combined therapy with levothyroxine and liothyronine in two ratios, compared with levothyroxine monotherapy in primary hypothyroidism: a double-blind, randomized, controlled clinical trial. J Clin Endocrinol Metab. 2005 May;90(5):2666-74. Epub 2005 Feb 10. http://www.ncbi.nlm.nih.gov/pubmed/15705921
- Chopra IJ, Carlson HE, Solomon DH. Comparison of inhibitory effects of 3,5,3′-triiodothyronine (T3), thyroxine (T4), 3,3,’,5′-triiodothyronine (rT3), and 3,3′-diiodothyronine (T2) on thyrotropin-releasing hormone-induced release of thyrotropin in the rat in vitro. Endocrinology. 1978 Aug;103(2):393-402. http://www.ncbi.nlm.nih.gov/pubmed/105890
- Apostolos G. Vagenakis, Lewis E. Braverman, Fereidoun Azizi, Gary I. Portnay, and Sidney H. Ingbar. Recovery of Pituitary Thyrotropic Function after Withdrawal of Prolonged Thyroid-Suppression Therapy. N Engl J Med 1975; 293:681-684 http://www.nejm.org/action/showImage?doi=10.1056%2FNEJM197510022931402&iid=f002&
- Peter Warmingham. Effect of Exogenous Thyroid Hormone Intake on the Interpretation of Serum TSH Test Results. Thyroid Science 5(7):1-6, 2010. http://www.thyroidscience.com/hypotheses/warmingham.2010/warmingham.7.18.10.pdf
- Taimela E, Koskinen P, Nuutila P, Nikkanen V, Saraste M, Taimela S, Irjala K. Free thyroid hormones and a third-generation TSH assay in the detection of hyperthyroidism during long-term thyroxine treatment in thyroid carcinoma patients. Scand J Clin Lab Invest. 1995 Apr;55(2):181-6. http://www.ncbi.nlm.nih.gov/pubmed/7667611
- Romeo Mariano. The Usefulness of TSH. DefinitiveMind.com. July 24, 2010. http://www.definitivemind.com/2010/07/24/the-usefulness-of-tsh/
- Stephen LaFranchi. Thyroid hormone in hypopituitarism, Graves’ disease, congenital hypothyroidism, and maternal thyroid disease during pregnancy. Growth Hormone & IGF Research, Volume 16, Supplement 1, July 2006, Pages 20-24. http://www.ncbi.nlm.nih.gov/pubmed/16707271
- Celani MF, Bonati ME, Stucci N. Prevalence of abnormal thyrotropin concentrations measured by a sensitive assay in patients with type 2 diabetes mellitus. Diabetes Res. 1994;27(1):15-25. http://www.ncbi.nlm.nih.gov/pubmed/7648793
- MF Bayer, JA Macoviak and IR McDougall. Diagnostic performance of sensitive measurements of serum thyrotropin during severe nonthyroidal illness: their role in the diagnosis of hyperthyroidism. Clinical Chemistry 33: 2178-2184, 1987. http://www.clinchem.org/cgi/content/abstract/33/12/2178
- Maes, M., Mommen, K., Hendrickx, D., Peeters, D., D’Hondt, P., Ranjan, R., De Meyer, F. and Scharpe, S. Components of biological variation, including seasonality, in blood concentrations of TSH, TT3, FT4, PRL, cortisol and testosterone in healthy volunteers. Clinical Endocrinology. 1997. May;46(5):587-98. http://www.ncbi.nlm.nih.gov/pubmed/9231055
- J. A. Romijn, R. Adriaanse, G. Brabant, K. Prank, E. Endert, And W. M. Wiersinga. Pulsatile Secretion of Thyrotropin during Fasting: A Decrease of Thyrotropin Pulse Amplitude. J. Clin. Endocrinol. & Metab., Jun 1990; 70: 1631 – 1636. http://jcem.endojournals.org/cgi/content/abstract/70/6/1631
- W. Russell, R. F. Harrison, N. Smith, K. Darzy, S. Shalet, A. P. Weetman and R. J. Ross. Free Triiodothyronine Has a Distinct Circadian Rhythm That Is Delayed but Parallels Thyrotropin Levels. The Journal of Clinical Endocrinology & Metabolism Vol. 93, No. 6 2300-2306, 2008. http://jcem.endojournals.org/cgi/content/full/93/6/2300
- Scobbo RR, VonDohlen TW, Hassan M, Islam S. Serum TSH variability in normal individuals: the influence of time of sample collection. W V Med J. 2004 Jul-Aug;100(4):138-42. http://www.ncbi.nlm.nih.gov/pubmed/15471172
- Michael T. McDermott, E.Chester Ridgway. Thyroid hormone resistance syndromes. The American Journal of Medicine. Volume 94, Issue 4 , Pages 424-432, April 1993. http://www.amjmed.com/article/0002-9343%2893%2990155-I/abstract
- Terry Davies, Russell Marians, Rauf Latif. The TSH receptor reveals itself. J Clin Invest. 2002;110(2):161–164. http://www.jci.org/articles/view/16234/table/1
- Dutta P, Bhansali A, Masoodi SR, Bhadada S, Sharma N, Rajput R. Predictors of outcome in myxoedema coma: a study from a tertiary care centre. Crit Care. 2008;12(1):R1. Epub 2008 Jan 3. http://www.ncbi.nlm.nih.gov/pubmed/18173846
- Pereira VG, Haron ES, Lima-Neto N, Medeiros-Neto GA. Management of myxedema coma: report on three successfully treated cases with nasogastric or intravenous administration of triiodothyronine. J Endocrinol Invest. 1982 Sep-Oct;5(5):331-4. http://www.ncbi.nlm.nih.gov/pubmed/7153477
- W. V. Baisier, J. Hertoghe and W. Eeckhaut. Thyroid Insufficiency. Is TSH Measurement the Only Diagnostic Tool? Journal of Nutritional & Environmental Medicine (2000) 10, 105–113. http://informahealthcare.com/doi/abs/10.1080/13590840050043521
- Ikeda T., Honda M., Murakami I. Effect of TSH on conversion of T4 to T3 in perfused rat kidney (1985) Metabolism: Clinical and Experimental, 34 (11), pp. 1057-1060. http://www.sciencedirect.com/science/article/pii/0026049585900794
- Tadasu Ikeda, Tatsuo Takeuchi, Yasuo Ito, Isao Murakami, Osamu Mokuda, Masato Tominaga, Hiroto Mashiba. Effect of thyrotropin on conversion of T4 to T3 in perfused rat liver. (1986) Life Sciences, 38 (20), pp. 1801-1806. http://www.sciencedirect.com/science/article/pii/0024320586901335
- Anna G. Dagre, John P. Lekakis, Theodore G. Papaioannou, Christos M. Papamichael,Demetrios A. Koutras, Stamatios F. Stamatelopoulos, Maria Alevizaki. Arterial stiffness is increased in subjects with hypothyroidism. International Journal of Cardiology 103 (2005) 1 – 6. http://www.heartsmarttesting.com/assets/files/pdf/research_articles/thyroid.pdf
- Hoermann, R., Midgley, J. E., Larisch, R., & Dietrich, J. W. (2012). Is Pituitary Thyrotropin an Adequate Measure Of Thyroid Hormone-Controlled Homeostasis During Thyroxine Treatment?. European Journal of Endocrinology. http://www.eje-online.org/content/early/2012/11/26/EJE-12-0819.short