|Values are valid only on day of printing.|
Click CC to turn on closed captioning.
Published: April 2009Print Record of Viewing
Accurate assessment of thyroid function relies on thyroid-stimulating hormone (TSH) and free thyronine (FT4). Dr. Grebe reviews the strengths and weaknesses of the currently available thyroid hormone assays.
Presenter: Stefan K. Grebe, MD, PhD
Welcome to Mayo Medical Laboratories' Hot Topics. These presentations provide short discussion of current topics and may be helpful to you in your practice.
Our presenter for this program is Dr. Stephan Grebe, Chair of the Division of Clinical Biochemistry and Immunology at Mayo Clinic. Accurate assessment of thyroid function relies on thyroid-stimulating hormone (TSH) and free thyronine (FT4). Dr. Grebe reviews the strengths and weaknesses of the currently available thyroid hormone assays.
Now, the topic of today is thyroid function testing. The thyroid gland can have both diseases of its structures, as well as its function. What we are going to look on today is simply the functions, so it’s just on the right side of this equation.
Just to remind ourselves briefly about thyroid function and its regulation, it’s a typical and classical negative-feedback endocrine regulation system. The pituitary gland secretes thyroid-stimulating hormone, or TSH for short, which in turn interacts with the thyroid gland and makes it produce increased amounts of thyroxin, also known as T4 for short, and triiodothyronine, also known as T3 for short. Both of those provide negative feedback on the hypothalamus and the pituitary gland, and assay levels increase the levels of TSH downregulated, a classical thermostat.
Consequently, all of those 3: TSH, T4, and T3, are the prime hormones for assessment of thyroid function. Let’s cover them one by one. Thyroid-stimulating hormone, TSH, has many reliable assays. They are all of them, today, very reliable with a wide dynamic range. It provides an integrated measure of thyroid hormone action, being the target of the feedback inhibition; it really gives the measure of the tissue affect of thyroid hormone. Furthermore, and we will enlarge this further down in this presentation a little, the response of TSH to changing thyroid hormone levels is greatly amplified, making it exquisitely sensitive to changes in thyroid hormone levels. Finally, another plus is, it’s not affected by any binding protein changes. There are, of course, also a few minuses. First of all, it’s obviously an indirect measure of thyroid function, as it measures the response to rising or falling thyroid hormone levels. So, there is some delay in responses to acute changes in peripheral thyroid hormone levels, which in some clinical situations could be a disadvantage. Finally, in a very rare case, this is <1% of cases, where the patient might have pituitary or hypothalamic disease, or thyroid hormone resistance, or some other exotic disorder, the TSH measurement alone can be misleading.
Total thyroxin has several advantages of its own. First of all, the assays, like those for TSH, are pretty reliable and they are a good measure of thyroid gland output. Moreover, the changes occur fast when thyroid gland activity changes, which makes it very suitable for investigating acute changes in thyroid function. However, the concentration of total thyroxin is highly variable, and highly dependent on the variable thyroid hormone-binding globulin concentrations. TBG, which is short for thyroid hormone-binding globulin, is the main binding protein for thyroid hormones and its concentrations can vary widely depending on a number of exogenous factors such as sex-hormone concentrations or liver function. Only a small fraction of the total thyroxin is free and that’s actually the biologically active fraction. This small fraction, in addition, is a pro-hormone, the active hormone being actually T3. And, finally, the assays themselves, while reliable, have limitation of being all competitive assays and suffering from a limited dynamic range.
Total T3 has also got reliable assays, and it represents the active thyroid hormone. Again, changes occur fast with changes in thyroid gland activity and, over the other analytes mentioned, it can have occasionally the advantage of being selectively overproduced in thyrotoxicosis, particularly early on. However, similar to total T4, the concentration is linked to the highly variable thyroid hormone-binding globulin concentrations, and only a small free fraction of T3, again, is biologically active. Furthermore, the circulating levels of T3 may not be representative of the levels found at the actual tissue sites, as the majority of T3 is actually not secreted from the thyroid glands, but manufactured in peripheral tissues on demand. Finally, all T3 assays are competitive assays similar to T4 assays and, hence, suffer from the same problems of a limited dynamic range.
The free hormone measurements can obviously address some of the shortcomings of the total hormone measurements. They also, at least for free T4, are reasonable reliable, give a good measure of the thyroid gland hormone output, change fast with thyroid gland activity and, most importantly, are independent of TBG concentrations and they represent the biologically active fraction of T4. However, for the free T4, the same goes as for the total T4, it is a pro-hormone—the active hormone is T3, and levels can occasionally fluctuate with nonthyroidal illness, when binding protein levels change very rapidly, and it takes awhile for new equilibrium to be established. In those situations, free T4 measurements can be misleading. Standard assays again have a very limited dynamic range and at some ranges of binding protein concentrations, they may be unreliable.
Free T3 has the same advantages, more or less, as free T4 and very similar disadvantages. One of the major disadvantages here is that all current free T3 assays have serious shortcomings and generally are not recommended. They are too unreliable in medical clinical situations. The limitations are caused, by and large, the very large low concentrations of the analyte which is being measured. Free thyroxine concentrations are already only a few percent of the total thyroxin concentrations. Free T3 concentrations, in turn, are only a few percent of the total T3 concentrations, and the total T3 concentrations are never more then 20% of the total thyroxin concentration. So, we are talking about low picomolar concentrations, which causes a lot of analytical problems.
So, how do we balance it up? What tests do we use in what situations and when? Well, there are several scenarios we have to consider. First of all, the initial diagnosis. For initial diagnosis, the general recommendation of all professional bodies and of all laboratories is to measure TSH. It is equally useful for hypo- and hyperthyroidism, has the highest sensitivity and specificity for initial diagnosis, and is least likely to be disturbed by nonthyroidal illness or drugs. The reason it is so exquisitely sensitive for changes in thyroid hormone function is that its response is magnified. When we look at this graph on the right, you can see that a doubling or halving of the free thyroxin concentrations will result in a 100-fold increase or decrement in the TSH concentration. So, the response of TSH to changing thyroid hormone levels is greatly exaggerated, which facilitates the rapid feedback. This is very useful for analytical measurements as it will show even minor changes in thyroid functions very graphically.
The downside is that, in some situations, such an exaggeration might not give us a good measure of the overall severity, because even a slightly severe thyrotoxicosis or hypothyroidism can give massive change in TSH. So, free T4 measurements are often used when either the TSH alone is not clearly diagnostic (ie, borderline measurements) or when there is some need to gauge the severity of hypo- or hyperthyroidism. Finally, there is the odd situation where total T3 measurement can also be helpful, and this is really those cases where both TSH and FT4 measurements are on the fence, or not clearly diagnostic.
These recommendations are summarized in this algorithm, which Mayo Medical Laboratories generally recommends for initial workup of nonhospitalized patients with suspected thyroid disease. We can clearly see it’s primarily driven by the initial TSH measurement. If this is below the reference range, then hyperthyroidism can be suspected and can be further confirmed by the free thyroxin measurement. In rare cases, where that is only borderline elevated, or not elevated at all, it could be that this is early Grave’s disease, for example, where primarily total triiodothyronine would be elevated and, hence, a total T3 measurement might be indicated. In situations of borderline thyrotoxicosis the algorithm is very similar. Finally, in the normal range, no further testing is generally justified. Lastly, for those patients with elevated TSH, the diagnosis of hypothyroidism can be suspected if the elevation is minor. This is typically called subclinical hypothyroidism and further supporting factor would be a normal T4. If the T4 or the free T4, in particular, is also slightly depressed and the TSH is >10, which we don’t have in this algorithm, then generally a clinically diagnosis and biochemical diagnosis of manifest hypothyroidism is made. If the etiology of that is in question, then autoantibodies to thyroid tissues can be measured. This can also be helpful in cases of subclinical or borderline hypothyroidism to determine what the risk of the patient is to progress to full-blown hypothyroidism.
The next scenario is follow-up of patients. There are many more patients with established thyroid disease then with newly diagnosed thyroid disease, and many of them require life-long or, at least, long-term follow up. For those conditions which are acute or subacute, for example, Graves’ disease or subacute thyroiditis, it is important to measure FT4 because the changes in thyroid hormones can occur, particularly under treatment intervention, very rapidly and the TSH levels may lag a few days or even weeks behind. This can be observed particularly if the thyrotoxicosis has been long standing. In those cases, there is actual atrophy in the pituitary gland of thyrotrope cells, and it can take some weeks for TSH secretion to recover. In most cases, the FT4 measurement is, therefore, the best guide to therapy and resolution of Graves’ disease or subacute thyroiditis. But, it is usually also supplemented with TSH measurements and we have already mentioned that in those rare cases where primarily T3 is secreted, a total T3 measurement can be useful. Chronic or slowly progressive conditions are a little different. Those primarily consist of permanent hyperthyroidism or toxic nodular goiter, where there is either no progression in the disease or very slow progression and, here, the TSH measurement is the main stay, which is occasionally supplemented by FT4 measurements and, rarely, by total T3 measurements.
Finally, there is some auxiliary testing which is used to establish etiology or when things are murky. Antithyroperoxidase autoantibodies and antithyroglobulin autoantibodies can be useful to access the risk of progression of borderline hypothyroidism to complete hypothyroidism. If those autoantibodies are detectable, then there is a larger risk that a patient with borderline hypothyroidism will progress eventually to full-blown hypothyroidism.
Anti-TSH-receptor autoantibodies, measured either by a binding assay or a bioassay, can also be useful in some select situations. Those would include the differential diagnosis of Graves’ disease versus a thyrotoxic phase of subacute thyroiditis. Or, for example, also Graves’ disease on top of a toxic goiter. Many patients with a toxic goiter may also develop Graves’ disease. The main applications, however, are in pregnancy. First of all, a differential diagnosis between excessive vomiting of pregnancy versus first trimester Graves’ disease. Sometimes, Graves’ disease can manifest for the first time in the first trimester, this is also the time when excessive vomiting and nausea can occur and, those by themselves, through mechanisms which are not entirely clear, but may involve very high levels of HCG cross-reacting on the TSH receptor, can lead to apparent thyrotoxicosis. So here, detecting or not detecting anti-TSH receptor autoantibodies can give a definitive diagnosis. A second pregnancy-related scenario centers on the risk of a baby of a women who had Graves’ disease in the past, to be afflicted by neonatal Graves’ disease. Neonatal Graves’ disease is caused by autoantibodies that the mother still has in her system passing through the placenta to the baby causing transient, but sometimes severe, thyrotoxicosis. As most of those mothers either would have had their Graves’ disease treated, or the Graves’ disease would naturally go into seeming remission in the second trimester, it is very difficult to use a mother’s thyroid-function testing to judge whether there is circulating autoantibodies, which could affect the baby. Hence, measurement of TSH-receptor autoantibodies can be very helpful. If those are significantly elevated, either by a binding or activity assay, then this means the baby is at risk of neonatal thyrotoxicosis and expert obstetric care is required.
Thyroid hormone-binding protein measurements are also useful in some situations, particularly if there are inconsistent thyroid hormone test results between the 3 tests we’ve mentioned before. They can occur in some situations where they are inherited or acquired abnormalities in binding proteins, sometimes related to drug treatment and sometimes related to congenital genetic conditions. The hallmark is that there is large discrepancies between free and total hormone levels, and those type of results should really tip one off that one might want to do a thyroid hormone-binding protein measurements, either TBG concentration measurements or TBG-binding capacity measurements. Finally, there are very, very rare cases where molecular analysis of the thyroid hormone receptors are useful. Those would be scenarios where free T4, total T4, and total T3 are elevated, and TSH level is normal, without there being an acute illness or interfering drug treatment, which could cause this. Those patients may have inherited disorders of the thyroid hormone receptor, so-called thyroid hormone resistance, or some of these patients may have a TSH-secreting pituitary tumor.
So, to close this section off, we should mention some thyroid hormone tests which are either, these days, unnecessary or probably belong in the realm of voodoo. First of all there is a free thyroxin index. This test is certainly no voodoo, but it is an indirect measure of free thyroxin and requires 2 analytical measurements, a total T4 measurement and a so-called resin T3-uptake test. What those tests do, is they assist the total thyroid hormone level, plus with the resin uptake test, an indirect measurement of thyroxin-binding globulin concentration. The 2 results are then multiplied and an index is formed. This has really no advantages over FT4 measurement, while at the same time, introducing additional testing, additional costs, and, of course, additional analytical errors as the analytical errors of the 2 methods are not applicative. Another one in this category is reverse T3, which more tends towards the voodoo sides of things. It was once in vogue for distinguishing nonthyroidal illness from thyroidal illness, but was clinically really not found to be very useful. Mostly the clinical context and measurement of TSH, free T4, free T4 by dialysis, and total T4 and total T3 will give a sufficient answer, or comprehensive answer, and reverse T3 does not add much. It does, however, have occasional niche application for suspected inherited or acquired abnormalities and deiodinase activity. These are the enzymes, which in peripheral tissues convert T4, the pro-hormone, either to the active hormone T3 or the inactive hormone reverse T3. Such applications are really rare and generally most endocrinologists, let alone general internists and general practitioners, will never see patients in their lifetime who will need a reverse T3 test.
Finally, a few key features of the actual assays listed here in an appendix, and we will concentrate on the 2 man assays we talked about, TSH and FT4. Those are the 2 tests which, if I had a medical practice on a desert island I would like to have to assess thyroid function; beyond my fingers, of course.
TSH assays are all immunometric. The desirable analytical performance is a functional sensitivity of ≤0.01 mIU/L and all these assays are also known as enhanced third-generation assays. Other useful characteristics are that there is a difference in susceptibility of different TSH assays to either autoantibody interferences or heterophile-antibody interferences. Both of those are not common phenomenon, but they can cause significant headaches in the laboratory and in the clinical practice. Unfortunately, manufacturers don’t exactly advertise that their assay is particularly susceptible, so this type of information travels word-of-mouth amongst laboratorians and clinicians, and if you have any questions, you should consult with your laboratorians and clinicians. Free
T4 assays come in 2 flavors. The most commonly used ones are immunoassay. They are all competitive assays, which means they have a limited dynamic range, as we mentioned before. When you have a choice or the manufacturer gives you that information, then you should select the assay with the widest dynamic range, because you want to cover both low Free T4 level, as well as substantially high Free T4 level, because at very high levels, many laboratories have an alert or alarm levels for critical values. So if your test hasn’t got enough dynamic range to cover both low levels and all the way up to the alarm levels, the critical value levels, then you end up diluting a lot of samples. For competitive assays, in general, and for free hormone assays, in particularly, dilutions are always a dicey thing in terms of accuracy. So, if you have a choice, select an assay with the widest dynamic range. Also, these assays do differ whether they are affected by TBG concentrations or not. The claim to fame, of course, is that they aren’t affected, and the way they work is really that the assay uses an analog to thyroxin as a competitor, and that analog is meant to not bind at all to TBG. But, of course, life is never perfect and they do bind to TBG and, if you know that information, then you should select the assay which has the lowest TBG binding. Again, this can be very difficult to come by. Anyway, those assays which do have the lowest TBG binding for the analog competitor do correlate most closely with the reference method, dialysis. Which brings us to the reference methodology, free thyroxin measurement by either dialysis or ultrafiltration (ie, by physical separation of the TBG-bound T4 from the free T4). This method is a little cumbersome and a little slower then the immunoassays, but in some situations, where there is highly abnormal binding protein or binding proteins that have changed very quickly, or binding proteins that are immature (this would be in very small children for example), the reference methodology is indispensible to get accurate free T4 measurements.