Question and Answer Session

August 15, 2006 Presentation: Clinical Implications of the New TSH Reference

Welcome to AACC’s Expert Access Live Online program

Clinical Implications of the New TSH Reference Range

This month's expert is Carole Spencer, Ph.D., F.A.C.B.,from University of Southern California, Los Angeles, California. View the presentation and direct your questions to our online experts. AACC would like to thank Bayer HealthCare Diagnostics for making this program possible.

For the authoritative document describing use of thyroid function testing in the clinical laboratory, you can visit this web site (
) to review the NACB Laboratory Medicine Practice Guideline titled "LABORATORY SUPPORT FOR THE DIAGNOSIS AND MONITORING OF THYROID DISEASE." This guideline was edited by today's moderator, Dr. Carole A. Spencer, and her NACB colleague Dr. Laurence Demers, and it is available in English, French and Spanish languages.
Washington, DC

AACC Moderator: In the spirit of full disclosure, this information has been provided by the Expert Access program moderator.

Why is the TSH measurement more clinically sensitive for assessing thyroid status than a free FT4 test?
Paris, France

Carole Spencer Ph.D., FACB.: Because the TSH/free T4 relationship is log/linear, when free T4 moves off the optimal setpoint for that individual this free T4 abnormality is magnified 50 times in terms of an inverse change in TSH. TSH measurement is the only test that can detect an abnormality in free T4 status because we do not typically know where the free T4 setpoint of an individual patient lies relative to the population reference range (~0.8-1.8 ng/ dL).

There seems to be an implication that an individual's "set point" TSH concentration is the optimal value for that individual. Is there any evidence to support that? (My weight "set point" is quite stable, but almost certainly not optimal.)
Seattle, WA

Carole Spencer Ph.D., FACB.: The set point is not in TSH but free T4. This has been established from free T4 studies in mono- and dizygotic twins. If free T4 is being maintained on it's setpoint and the pituitary is not sensing any free T4 deficiency or excess relative to this, the TSH remains within a fairly narrow range (~ 0.75 mIu/L) over long periods of time.

How about the correlation between TSH and Hyperthyroidism and the risk of Cardio-vascular disease?
St.Louis, MO

Carole Spencer Ph.D., FACB.: Thyroid status (as reflected by TSH) is only one of many factors that influence risk for cardiovascular disease (CVD). CVD and atrial fibrillation (AF) are exaccerbated with both thyroid hormone excess and deficiency (Sawin NEJM 331:1249,1994 and Hak Ann Intern Med 132:270). Thyroid-mediated CVD and AF risks increase with age and the degree and duration of thyroid dysfunction. The relationship is one of association rather than correlation. The strength of the relationship will vary depending on the selection criteria used for the patient cohort. For example, thyroid dysfunction would likely show a stronger association with CVD for groups already at risk (i.e. hyperlipidemia or diabetes or a previous history of CVD). This is why the risk of having a TSH outside the empirical range has to be weighed on a case-by-case basis.

When the TSH is between 2.5 and 4 mUI/L Why not to use the TRH test?
Buenos Aires

Carole Spencer Ph.D., FACB.: The TRH response is merely proportional to the basal TSH (Spencer JCEM 76:494–498, 1993) so that when basal TSH is detectable TRH stimulation provides no additional information. Currently TRH testing is only useful to test hypothalamic-pituitary function. Measurement of TPOAb would be far more predictive of underlying thyroid dysfunction than a TRH test.

Which reference interval for TSH do you reccomend for children and young adults?
Porto Alegre, Brazil

Carole Spencer Ph.D., FACB.: Clearly, TSH levels are highest at birth and decline throughout childhood until after the age of about 14 mIU/L (see Table 3 NACB guidelines on Again, a useful adjunctive test would be TPOAb.

Will there be any discussion on Hyperthyroidism?
Albany, GA

Carole Spencer Ph.D., FACB.: There is broad agreement that the lower TSH reference limit lies around 0.3 mIU/L. The risk of cardiovascular problems for patients having a low TSH depends on duration of the low TSH and patient-specific risk factors, especially age (Sawin NEJM 331:1249,1994 and Parle Lancet 258:861,2001). Osteoporotic fractures are also increased in post-menopausal women not taking estrogen replacement with a chronically low TSH in the 0.1 -0.5 mIU/L range (Bauer AIM 134:561,2001). Concerns of inducing iatrogenic subclinical hyperthyroidism is why it is critical to titrate an L-T4 dose to produce a TSH in the 0.5 to 2.0 mIU/L range, and not overtreat. It is difficult to diagnose hyperthyroidism in the first trimester of pregnancy because TSH can be as low as ~0.02 mIU/L (Slide 35) as a result of direct hCG stimulation of the thyroid - a physiological response to ensure adequate maternal thyroxine is available to the developing fetus.

Should we even attempt to establish reference ranges locally?
Bangor, Maine

Carole Spencer Ph.D., FACB.: I don't think it is useful to try to establish "local" reference ranges. There are differences in the reference ranges of different locations (d'Herbomez Clin Chem Lab Med 43:102,2005). There are also practical difficulties as a result of HIPPA, and the cohort the more likely a "ringer" with occult thyroid dysfunction or in the rebound phase of a NTI will be included who will skew the range for the entire cohort. Participants with outlying TSH values relative to the group median would need to be checked after several weeks according to the NACB guidelines ( and in addition to TPOAb measurements ideally all participants should have ultrasound evaluations - which is impractical and expensive. It is much better not to be fixated on establishing "absolute" TSH limits but use the risk stratification approach described in Slides 42 and 43. For example, one would seriously consider L-T4 treatment for a 60 year old hypercholesterolemic woman with a TSH of 3.5 mIU/L but not a 28 year old active male with no TPOAb detected and an identical TSH. It is patients that are treated not a TSH result.

If you had discarted TSHoma or RTH and you suspect an inappropiate TSH syndrome, which diagnostic you consider if the patient shows no sintoms, normal proteinogram and high TSH and T4 L determined by two differents methods? POSGRADO DE ENDOCRINOLOGIA, HOSPITAL NACIONAL DE CLINICAS. CÓRDOBA ARGENTINA.
Córdoba, Argentina

Carole Spencer Ph.D., FACB.: If the TSH were high and a TSH-secreting pituitary tumor and thyroid hormone resistance (RTH) had been ruled out I would consider the following: 1. Obviously the first thing is to confirm the high TSH with a new blood specimen after 4-8 weeks.This was presumably done. 2. If TPOAb were positive this would suggest thyroid hormone deficiency. 3. Even if TSH were high by two different methods you cannot entirely rule out heterophilc antibody (HAMA) interference - no IMA method is immune. I would check with a blocker tube. 4. I would look at where the total and free T4 fell relative to their respective population reference ranges. I would expect T4 status to be in the lower third of the range. 5. TRH testing was presumably done to exclude central hypothyroidism. 6. Ultrasound might show a hypoechoic pattern suggestive of lymphocytic infiltration. 7. If there appeared absolutely no reason for the high TSH I would consider the possibility that the patient had an inactivating TSH receptor polymorphism requiring a higher TSH to maintain euthyroidism (? test family members). 8. Cases of patients secreting a high MW bioinactive TSH have been described. (identified by gel chromatography).

Dr. Spencer, Your excellent documented presentation covers the most important and controversial aspects on the topic. In a recent European Journal of Endocrinology ( 2006) 154 633: 637, Dr. Brabant and colleagues say that there is only one study providing direct data to support an upper normal limit of 2.0-2.5 (Ship-1), which probably still included people with autonomous functioning areas in the thyroid as judged from the fall in median TSH values with age, that probably reflecting the previously moderately severe iodine deficiency in the population studied. Question: Based on this: Do you have any comment on this? Specifically, do you believe that reference range should be geographical area specific and not “universal”, the same way as the threshold for treatment should be patient specific? Thanks Laboratory of Endocrinology. Hospital de Clínicas. Córdoba.Argentina
Cordoba, Argentina, Laboratory of Endocrinology,Hospital de Clínicas

Carole Spencer Ph.D., FACB.: A comparison of the NHANES (USA) and SHIP (Germany) TPOAb negative cohorts across decades of age clearly show that the TSH upper limit increases with NHANES and decreases with SHIP. In the high iodine environment of the USA the underlying pathology is an increasing prevalence of Hashimotos' thyroiditis and hypothyroidism with aging. In contrast, previous iodine deficiency in the German cohort has left an underlying pathology of nodular goiter and thyroid autonomy that increases with age. This is responsible for an apparent increasing TSH with age in NHANES and decreasing TSH with age in SHIP. It is clear that it is impossible to completely remove the effects of the underlying pathology of a population by TPOAb or even ultrasound exclusions. The underlying pathology always "contaminates" the mathematically calculated TSH reference limits. Thus, because it is impossible to establish a fixed TSH upper limit the risk stratification approach described in slides 42 and 43 is more appropriate.

I presume all the data presented here is based on an adult population. Are these recommendations applicable to pediatrics?
Memphis, TN

Carole Spencer Ph.D., FACB.: Yes, I was discussing only the TSH reference range for adults. TSH levels are highest at birth and decline throughout childhood until after the age of about 14 mIU/L (see Table 3 NACB guidelines on

do you recommend the TSH-TRH test in the patients with TSH basal levels in the upper limmits?
buenos Aires-Argentina

Carole Spencer Ph.D., FACB.: This has been an active question. The TRH response is merely proportional to the basal TSH (Spencer JCEM 76:494–498, 1993) so that when basal TSH is detectable TRH stimulation provides no additional information. Currently TRH testing is only useful to test hypothalamic-pituitary function. There are also issues regarding how a reference range for the TRH test is established. It is necessary to standardize for time of day because the TRH response is much higher in the morning than the afternoon - again amplification of the physiologic TSH diurnal variation. Also, it is important to consider the CV of repetitive TRH testing (at the same time of day, allowing a week at least between tests to allow thyroid hormones to return to baseline) i.e the biologic variability of the TRH response. In my view, I do not think TRH testing adds any value to basal TSH. Furthermore, TRH stimulated values are assay dependent because different TSH IMAs have been shown to detect the more biologically active, newly released TSH to differing degrees (Clin Chem 39:2167,1993).

Should we discourage the use of Total T4 in favor of the Free T4 assay. I still have physicians that order it and would like discontinue the test. Very low volume. Thanks
Waconia, MN

Carole Spencer Ph.D., FACB.: It is critical to retain total T4 as an accessible lab test. Unfortunately, current free T4 immunoassays are merely ESTIMATE tests and do not measure free T4 directly. They are still all binding protein dependant and prone to give false values when binding proteins are very abnormal, especially in patients with nonthyroidal illness (NTI) (Sapin Clin Chem 46:418,2000). Also, it is apparent that the non-pregnant FT4 reference ranges cannot be used to manage pregnant patients. FT4 reference ranges are both trimester and method-specfic and there are currently no appropriate ranges established for pregnancy. Current guidelines (NACB) suggest the continued use of total T4 (TT4) and FT4 indexes to assess the thyroid status of hospitalized patients. In this setting the total T4 has to be interpreted relative to the severity of illness. For example, the low T4 state of NTI only occurs with severe illness (TSH is not elevated >10 mIU/ L). If TT4 is low and TSH not elevated and the patient is not severely sick, a diagnosis of central hypothyroidism should be excluded. New guidelines for pregnancy (Mandel Thyroid 15:44,2005) also recommend the use of the total T4 and free T4 index or total T4 if the reference limits are adjusts by multiplying by 1.5 to account for the increased TBG.

Does T3 or Ft3 has any diagnostic importance as on date??

Carole Spencer Ph.D., FACB.: T3 is a tertiary test and only needed for investigating unusual presentations of hyperthyroidism or factitious hyperthyroidism which can be induced by certain over the counter (? weight loss) preparations containing T3 or Triac. Current automated T3 assays are set up to measure high T3 levels. Total T3 assays are far more robust and preferred over current free T3 immunoassays which offer little benefit. By the time a T3 status is needed for the diagnosis binding protein abnormalities are not usually an issue. NACB guideline #14 lists the uses for T3 measurements.

I have been approached about refining local strategies to evaluate thyroid status in pregnancy (a topic a considerable controversy in past decades). (1) Please comment on the utility of TSH in first trimester...New guidelines suggest maintaining TSH < 2.5 mIU/L, however some groups claim First trimester-Free T4 is a better predictor of fetal outcome (and First trimester-TSH has poor predictive value). Could Free T4 be a better predictor? (this assay is difficult to standardize across manufacturers). Could these claims that Free T4 is a better predictor be a consequence of confounding by albumin or other factors? Are all current commercial TSH assays NOW up to the challenge of accurate TSH measurement in the presence of high hCG concentrations? Have previous recomendations by the NACB emphasize the need for manufacturer and trimester-specific reference ranges to interpret Free T4.
Calgary, Alberta

Carole Spencer Ph.D., FACB.: This was an issues discussed in detail at the CDC conference on

Carole Spencer Ph.D., FACB.: This was an issues discussed in detail at the CDC conference on thyroid function and maternal fetal health in April 2004 (Proceeding in the January issue of Thyroid 2005). The relative merits of TSH vs Free T4 caused some discussion. Conclusions were that whereas there was some value in free T4 measurement in areas of severe iodine deficiency (not the USA!) TSH was the most accurate way to assess thyroid status. However, free T4 ranges were both trimester-specific and method-specific whereas TSH ranges was both trimester specific and geographically (?iodine) sensitive. Because of the log/linear relationship between TSH and free T4, TSH is always likely to be the more sensitive test for detecting thyroid dysfunction in pregnant patients in the same way as it is in non- pregnants provided that appropriate reference limits are used.

Do the new TSH reference ranges have any implications for the diagnosis of TSH-secreting tumours?
Leeds, U.K.

Carole Spencer Ph.D., FACB.: I do not think the setting of the TSH reference limits relates to the diagnosis of TSH secreting pituitary tumor which are detected in patients with biochemically high thyroid hormone levels and paradoxically DETECTABLE TSH. The TSH might only be in the 0.05-0.1 mIu/L range and be paradoxically high if the patient had very high thyroid hormone levels and/or was exhibiting symptoms of thyroid hormone excess. The differential diagnosis in this case rests more with excluding heterophilic antibody (HAMA) as the cause of the paradoxically high TSH (by blocker tube and testing TSH by different manufacturers assays) before embarking on expensive sella imaging.

Why is the assumption made (on page 24) that the TSH distribution should be simple Gaussian (thereby justifying cutting off the upper tail of the observed TSH distribution)? Log-normal distributions are not uncommon in endocrinology, particulary for trophic hormones that participate in negative feedback loops where the relationship between the trophic hormone and the target organ hormone is not a simple linear inverse relationship.
San Diego, CA

Carole Spencer Ph.D., FACB.: This is a good question and the assumption that TSH should be Gaussian is not in itself a reason for lowering the TSH reference limit. However, it is clear that the degree of deviation from Gaussian is influenced by the dominant pathology. The SHIP population (Volzke Thyroid 15:279,2005) that excludes both TPOAb+ and on the basis of ultrasound does appear Gaussian, although I do not have the data to perform the appropriate statistics.

What are the American Thyroid Association's guidelines on performing TSH and FT4 simultaneously? If so, what would be the rationale for requesting and performing the assays together? Thanks.
Durham, North Carolina

Carole Spencer Ph.D., FACB.: I am not aware that there are strict ATA guidelines covering the use of simultaneous TSH and FT4. Whether you need FT4 in addition to TSH depends on the clinical situation. For example: 1. TSH alone is recommended for case finding. In this case TPOAb is a more useful second-line test than FT4 because FT4 is not expected to be abnormal when TSH is between 0.05 and about 10 mIU/L because of the log/linear relationship (slide 8). 2. When thyroid status is unstable, in the early phase of treating hyper-or hypothyroidism T4 (and possible T3) are more useful than TSH which will exhibit a delay in pituitary reset. 3. When patients are suspected of being non-compliant with L-T4 medication the combination of tests can be useful because the patient can ingest a large number of missed doses on the morning of the clinic visit and have a FT4 within the reference range and yet an elevated TSH, reflecting the previous non- compliance because the TSH takes time (days/weeks) to fully reset. 4. In patients being monitored on high doses of L-T4 for TSH suppression of thyroid cancer it is often useful to have both tests because often TSH is suppressed below detection and the FT4 is a way to gauge the degree of iatrogenic hyperthyroidism produced. So, combined TSH+FT4 - it all depends!

Dr. Spencer, Is there any specific thyroid function test to differentiate Grave's disease from postpartum thyroiditis? There was a 22-year old lady, her TSH was severly suppressed, and FT4 and FT3 were highly elevated. From biochemical parameters, it is a status of hyperthyroidism. These tests were done 4 months after delivery. Correct diagnosis will determine the treatment. thanks, Qing Meng
Saskatoon, SK

Carole Spencer Ph.D., FACB.: In the post partum period there can be a flare up of thyroid autoimmunity which can be transient. This can be a flare up of Graves' hyperthyroidism or thyroiditis. Typically, thyroiditis will have a T3/T4 ratio of <20:1, indicative of a "dumping" of preformed hormone whereas Graves' will have a T3/T4 ratio >20:1 indicating active stimulation caused by TSH receptor antibodies. The other way to distinguish is a radioiodine uptake which will be suppressed in thyroiditis and high in Graves'. If the FT3 was highly elevated I would guess that your patient has Graves' hyperthyroidism.

why is 3rd gen TSH more important than second gen
Durham, NC

Carole Spencer Ph.D., FACB.: Reliable TSH measurement in the 0.01-0.1 mIu/L range can only be guaranteed using 3rd. generation assays. It is critical to be able to detect the difference between a fully suppressed TSH <0.01 mIU/L and a partially suppressed TSH in many situations: 1. In hospitalized patients with nonthyroidal illness (NTI) a TSH < 0.01mIU/L suggests hyperthyroidism whereas a TSH of ~0.04 mIU/L would suggest a transient effect of NTI. Treatment or non-treatment strategy would differ. 2. In patients receiving high doses of L-T4 to suppress TSH for thyroid cancer it is essential to be able to accurately titrate the degree of TSH suppression. Studies have shown that thyroglobulin levels respond to changes in TSH in the range below 0.1 mIU/L. 3. Most thyroid testing is performed in outpatients being monitored for L-T4 replacement therapy. It is critical to titrate L-T4 dose to give a TSH in the range 0.5-2.0 mIU/L. There is always concern for overtreating and causing iatrogenic hyperthyroidism with the attendant risk of atrial fibrillation (slide 19). The assay needs to be really solid in the range below 0.1 mIu/L to reliably detect this. 4. Assays always exhibit 'noise" around their sensitivity level. This noise is always positive - i.e. falsely high TSH values. There is no rationale for using a second generation method and "reflex" low values to a 3rd. generation because low values cannot be reliably detected (NACB guideline 21).

What resource do you recommend for pediatric ranges? How does lowering the TSH range affect pediatric ranges
Springfield, MO.

Carole Spencer Ph.D., FACB.: My presentation relates to the adult TSH range. Clearly, TSH levels are highest at birth and decline throughout childhood until after the age of about 14 mIU/L (see Table 3 NACB guidelines on

Carole Spencer Ph.D., FACB.: My presentation relates to the adult TSH range. Clearly, TSH levels are highest at birth and decline throughout childhood until after the age of about 14 mIU/L (see Table 3 NACB guidelines on

If the patient's TSH is 3.0 mIU/L or greater, should a aTPO test follow?
New York, NY

Carole Spencer Ph.D., FACB.: Yes, that would be a good strategy. A positive TPOAb result elevates the risk category for current or risk for developing hypothyroidism in the future (Whickham study, slide 32). It the patient is a young woman desiring pregnancy there would be concern that thyroid stress in the first trimester might result in thyroxine insufficiency for the developing fetus. Guidelines suggest prepregnancy TSH should be below 2.5 mIU/L. A positive TPOAb in the first trimester conveys a 30% risk for post- partum thyroiditis, if TPOAb is positive in the 3rd. trimester the risk is 50%. A TPOAb-positive patient with a TSH of 3.0 mIU/L would also be at risk for developing thyroid dysfunction if started on medications such as interferon, interleukin, lithium or amiodarone (NACB guideline #34).

At the bottom of page 33 it says, "New guidelines now..." and doesn't complete the sentence or thought. Would you please finish the sentence? Thank you!
Terre Haute, IN

Carole Spencer Ph.D., FACB.: My apologies. New guidelines now recommend that optimally, TSH should be maintained below 2.5 mIU/L during pregnancy.

Any comment on the low free T4 levels during pregnancy despite increases in L-T4 dose that consistently decrease TSH?

Carole Spencer Ph.D., FACB.: I do not think we know why FT4 (even when measured by equilibrium dialysis) progressively decreases with gestation. It is probably related to binding protein changes (? albumin status) maybe T4 clearance. The general consensus (CDC conference) that TSH is the better marker for thyroid status, however, TSH does not respond instantaneously to changes in thyroxine or thyroxine dosing so perhaps pregnancy is one situation to order both TSH and a free T4 index or total T4 (not a free T4 immunoassay for the reasons previously covered).

Do you see a shift in normal ranges in TSH in the same way you see a shift in FT4 in pregnancy?
Birmingham, Alabama

Carole Spencer Ph.D., FACB.: Usually the TSH is lower in the first trimester because of the high hCG stimulation and then gradually increases to pre-pregnancy levels by mid- gestation. The FT4 shifts (progressive lowering as pregnancy progresses) is not seen with TSH. Also if iodine intake is normal TSH does not increase as gestation progresses. There is in essence a disconnect between the TSH and FT4 patterns seen during gestation and TSH is considered the more reliable indicator of thyroid status.

Do you believe the disparity betwwen TSH assays is a major problem and is there any resolution to it?

Carole Spencer Ph.D., FACB.: The disparity between TSH assays is only a problem if you are trying to establish a rigid TSH reference range that covers all groups. The disparity between assays is likely to reflect TSH heterogeneity recognized by the monoclonal antibodies - it's not fixable. 2.5-3.0 mIU/L measured by one assay will always likely to be 3.0-3.5 mIU/L in another. It is better for us to abandon this concept of a fixed TSH range. It cannot be rationalized on either a methodology, biologic or practical basis.

I agree that the upper TSH limit should be at 3.0 mIU/L. It seems to me though that most primary physicians treat TSH like any other chemistry test. When the results come back from the reference lab,and your TSH is within the noted "normal" range, they do absolutely nothing. They do not track the individual' normal range or worry that a 3.0 may be subclinical hypothyroidism. What would do you recommend as the best approach ( by KOLs,labs or assay manufacturers ) to having physicians understand the idiosyncracies of TSH and the individual patient?

Carole Spencer Ph.D., FACB.: Physicians need education. it is to easy for them to "treat the number" and not the patient. I do think that labs should post a separate range for pregnancy (~0.02-2.5 mIU/L). Many physicians are chasing a diagnosis of hyperthyroidism when they see a low TSH in the first trimester.

in light of these new recommendations how should TSH measurements be used in monitoring patients on thyroxin replacement therapy for hypothyroidism
New Haven, CT

Carole Spencer Ph.D., FACB.: The guidelines (NACB, AACE and Endo Soc) all recommend that a TSH between 0.5 and 2.5-3.0 mIU/L is the recommended target for L-T4 replacement dose adjustment. Thyroxine has a long half-life and if the same dose is taken daily (without food etc) the TSH on long term replacement therapy remains remarkably constant (within the expected 0.75 mIU/L - NACB guideline 8).

TRH stimulated TSH can serve as a peripheral marker of thyroid hormone action at the pituitary?

Carole Spencer Ph.D., FACB.: I do not believe that TRH conveys any additional information over TSH as indicated in previous questions.

By lowering TSH lower reference range to 0.3, do you not increase risk of including some with subclinical hyperthyroid?

Carole Spencer Ph.D., FACB.: Young people with iatrogenic hyperthyroidism (thyroid cancer patients receiving high L-T4 for suppression do not have problems with marginally low TSH. A TSH of 0.3 mIu/L is not an absolute. TSH has an episodic and diurnal variability such that a 0.3 in the afternoon could well be 0.6 mIu/L at an early morning appointment. So setting the lower limit to 0.3 or 0.4 mIu/L does not impact the diagnosis of subclinical hyperthyroidism. For example, if you had an elderly patient receiving L-T4 replacement you would try to titre the dose to give a TSH in the 1.0-1.5 mIu/L range rather than down around the lower limit.

In regards to TRH challange related to hypothyroid my understanding is: Primary is due to the thyroid and a TRH challenge leads to TSH increasing 2-3x normal. Secondary is due to the pituitary and a TRH challenge leads to low/normal TSH (ie no response). Tertiary is due to the hypothalamus and a TRH challenge leads to increased TSH (after a delay). Is this correct? Another question, what is the actual new TSH value, is it 2 or 2.5 and how many labs are using it? Thanks
Fort McMurray, Alberta

Carole Spencer Ph.D., FACB.: You are correct in your comments regarding the type of TRH responses. The relationship between basal and fold TRH response tends to be biphasic (Spencer JCEM 76:494,1993). I do not know how many labs are lowering their upper limit. We use an upper limit of 3.0 mIU/L in accord with the AACE guidelines. However, the increasing concerns regarding pregnancy suggest that perhaps we should lower it to 2.5 mIU/L. We are using the Roche Elecsys. Note in slide 23 this has a trend for higher values than the Access 2. If we changed methods we might consider reducing from 3.0 - 2.5 mIU/L for technical reasons.

Are TPOAb and TSH a valuable test in considering reasons behind a patient's multiple reoccurring spontaneous abortions?
Birmingham, AL

Carole Spencer Ph.D., FACB.: Most definitely. This is well established (NACB guideline #34).

Would you recommend using Anti-TPO or free T4 as the next test for an abnormal (high) TSH result.
Peterborough, ON

Carole Spencer Ph.D., FACB.: Because of the log/linear TSH/free T4 relationship you do not expect an abnormal FT4 unless TSH were <0.05 or >10 mIU/L (slide 8). TPOAb is by far the more important test to use to investigate whether a mild TSH abnormality might have an autoimmune basis.


Carole Spencer Ph.D., FACB.: Table 3 of the NACB guidelines gives some data regarding neonatal age and TSH. However this data was only obtained with one method (? Nichols).

Since there is currently no standardization of TSH assays, how can one reference interval be defined for all assays?
Memphis, TN

Carole Spencer Ph.D., FACB.: I believe an empiric reference range of 0.3-3.0 mIu/L is the best generalization we can make to cover all assays FOR NON-PREGNANT PATIENTS. The pregnancy reference range needs to be lower as discussed earlier.

A question of Tg assay. In patients with anti-Tg antibody, is the Tg level useful? What is the difference between immunoassay Tg and RIA Tg? Can the Tg level by RIA be used as a marker to monitor the recurrence of the cancer?
Chantilly, Virginia

Carole Spencer Ph.D., FACB.: I recommend the Tg/TgAb method comparison paper Spencer JCEM 90:5566,2005. Clearly all Tg IMAs are prone to produce falsely low or undetectable values in the presence of TgAb and not all TgAb methods detect all cases of interfering TgAb. The absolute TgAb concentrations reported by different methods vary widely and the same method needs to be used for serial monitoring. Serial TgAb can be used as a surrogate tumor marker test. Tg recovery tests are useless and the NACB guidelines state these should be discouraged and abandoned (NACB guideline 46). Tg RIA methods are clearly less prone to TgAb (or HAMA) interferences than IMAs. The best approach is to first assess the TgAb status of the specimen (by sensitive immunoassay) and if positive measure the Tg by RIA. TgAb positive patients can be monitored by serial TgAb and Tg RIA measurements. If RIA is not available monitor serial TgAb as a surrogate tumor marker (using the same method).

For subclinical hypothyroidism (TSH 3-10), how important are symptoms like tiredness, weight gain etc in deciding on replacement therapy?
Saskatoon, Canada

Carole Spencer Ph.D., FACB.: Clinical symptoms are weak indicators but because it is patients that are treated and not a number one cannot ignore them. Furthermore it is not unrealistic to assume that some patients will be more susceptible to slight degrees of thyroid hormone insufficiency (? different efficiency of the intracellular deiodinases for example could be postulated). If the patient has symptoms, a clinical trial of L-T4 treatment may be helpful.

When would you test for Anti-TG? Would it only be in patients receiving high doses of L-T4 to suppress TSH post-thyroidectomy for thyroid cancer? When would you test for anti-TSH receptor antibodies?
Los Angels, California

Carole Spencer Ph.D., FACB.: Ordering TgAb in addition to TPOAb is not useful to make the diagnosis of thyroid autoimmunity (NHANES study). The main use of TgAb is as an adjunctive test to Tg in patients with differentiated thyroid cancers in whom TgAb prevalence is twice the general population (~20%). It is rare to need TSH receptor antibody (TRAb). Most Graves' hyperthyroidism is self evident. Also, TRAb is not helpful stratifying patients for antithyroid drug treatment (until remission) versus radioiodine ablation. The primary use of TRAb is in pregnant patients with Graves' disease or a past history of RAI ablation for Graves' in whom there is risk of neonatal thyroid dysfunction caused by transplacental passage of the TRAb (blocking or stimulating).

You stated that 3rd gen TSH assays are important for hospitalized patients since they can detect levels <0.01 mIU/L. When looking at 3rd gen TSH assays, what is the lowest detectable concentration that is useful clinically? Is 0.0020 mIU/L any more useful than 0.004 ?
New York, NY

Carole Spencer Ph.D., FACB.: Current assays cannot reliably measure below 0.01 mIU/L. If a 4th. generation assay were available, the degree of suppression below 0.01 mIU/L might correlate with the chronicity of the thyroid hormone excess. For example, a patient entering the hospital with chronic undiagnosed hyperthyroidism might have a lower TSH <0.001 mIU/L than a patient with a mere transitory suppression perhaps due to high endogenous glucocorticoids of NTI.

Page Access: