TOTAL THYROID HORMONE MEASUREMENTS (TT4 AND TT3)
Thyroxine (T4) circulates 99.97% bound to the plasma proteins, primarily TBG (60-75%) but also Transthyretin TTR/TBPA (15-30%) and Albumin (~10%)(Table 2) . In contrast, approximately 99.7% of Triiodothyronine (T3) is protein-bound, primarily to TBG [34,35,75]. Total (free + protein-bound) concentrations of thyroid hormones (TT4 and TT3) circulate at nanomolar concentrations and are considerably easier to measure than the free hormone moieties (FT4 and FT3) that circulate in the picomolar range. Serum TT4 measurement has evolved over the past four decades from the protein-bound iodine and competitive protein binding tests [1,76] to non-isotopic immunometric assays  and LC-MS/MS methods [13,78-80].
Table 2: Conditions that Influence Thyroid Hormone Binding Proteins
Serum TT4 measurement has evolved over the past four decades from the protein-bound iodine and competitive protein binding tests [1,76] to non-isotopic immunometric assays  and LC-MS/MS methods [13,78-80]. Total hormone methods require the inclusion of inhibitors, such as 8-anilino-1-napthalene-sulphonic acid, to block hormone binding to serum proteins in order to facilitate binding to the antibody reagent . Methodology for TT4 measurement has changed over the decades and been paralleled by changes in TT3 methodology. However TT3 measurement presents a greater sensitivity and precision challenge, because TT3 concentrations are ten-fold lower than TT4 [13,82-86]. Most laboratories currently measure TT4 and TT3 concentrations by non-competitive immunometric assays performed on automated platforms using enzymes, fluorescence or chemiluminescent molecules as signals [25,75,87]. A recent IFCC C-STFT study compared eleven TT4 and twelve TT3 immunoassays marketed by eight diagnostic companies . TT4 and TT3 measurements were made in sera from healthy individuals using the various immunoassays and compared with values reported by isotope dilution tandem mass spectrometry (ID-LC-MS/MS) – the reference measurement procedure (RMP) based on using primary T4 and T3 standards for calibration [80,88]. Although most methods fell short of the optimal 5 percent goal established by the C-STFT, 4/11 TT4 assays agreed within 10 percent of the reference, whereas most TT3 assays exhibited a positive bias that would necessitate re-standardization [80, 88] (Figure 1). Thus, as would be expected, TT4 assays are more reliable than TT3 although assay variability persists, likely as a result of matrix differences between calibrators and patient sera, the efficiency of the blocking agent employed by different manufacturers and lot-to-lot variability [53,56,89,90].
Figure 1- Between-method TT4 and TT3 Variability
Figure 1. (A), (TT4); (D) (TT3): assay means (1-sided 95% CIs) vs the mean by the RMPs.The x axis gives the codes of the different assays, the dotted lines represent the mean of the RMP _10%. For the assays differing >10% from the mean of the RMP, the numerical value of the mean is listed. (B), (TT4); (E), (TT3): scatter plot (x = mean of the RMP, y = mean of singlicate results per assay) with indication of the line of equality (dotted) and the most extreme Deming regression lines/equations. The results for the most deviating assays are indicated by circles and triangles; all other assays are indicated with the same symbol, X. (C), (TT4); (F), (TT3): percent-difference plot with indication of the strongest negatively (circles) and positively (triangles) biased assays. Note that (B), (C), (E), and (F) are extended to show the complete range (10–221 nmol/L for TT4, 0.6 –1.9 nmol/L for TT3) .
Clinical Utility of TT4 and TT3 Measurements
The diagnostic accuracy of total hormone measurements would be equivalent to that of free hormone tests if all patients had similar binding protein concentrations [35,75]. In fact, a recent study has reported that a screening cord blood TT4 < 7.6 μg/dL (< 98 nmol/L) can be used as a screening test for congenital hypothyroidism . Unfortunately, many conditions are associated with TBG abnormalities that distort the relationship between total and free thyroid hormones (Table 1). Additionally, some patients have abnormal thyroid hormone binding albumins (dysalbuminemias) [92-94], thyroid hormone autoantibodies [95-98], or are taking drugs [25,99-101] that render total hormone measurements diagnostically unreliable [Table 1]. Consequently, TT4 and TT3 measurements are rarely used as stand-alone tests, but are typically employed in conjunction with a direct TBG measurement or an estimate of binding proteins [i.e. a thyroid hormone binding ratio test, THBR, that can be used to calculate a free hormone index (FT4I or FT3I). This index approach effectively corrects for the most common thyroid hormone binding protein abnormalities that distort total hormone measurements [ [102-104]. Because free hormone immunoassays are more technically challenging than total hormone measurements [49,86] total hormone tests can useful confirmatory when a free hormone immunoassay result appears questionable, especially in pregnancy and critical illness where changes in binding protein concentrations and affinity for thyroid hormones can occur [22,104-106]. Suboptimal FT3 assay sensitivity limits reliable FT3 measurements to the high (hyperthyroid) range . However, since T3 is typically only a 3rd-line test of thyroid status used for diagnosing unusual cases of hyperthyroidism, TT3 measurement can usually suffice in preference to FT3, especially when TT3 is used as a ratio with TT4 to eliminate binding protein effects . In fact, in Graves’ hyperthyroidism preferential thyroidal T3 secretion resulting from increased deiodinase activity secondary to thyroidal stimulation by TSH receptor antibodies (TRAb)  such that a high serum TT3/TT4 or FT3/FT4 ratio that can be used to differentiate Graves’ from other causes of hyperthyroidism [107,109,110].
TT4 and TT3 Reference Ranges
Total T4 reference ranges have approximated 58 to 160 nmol/L (4.5-12.5 µg/dL) for more than four decades, although some between-method differences and sample-related variability remains [80, 104]. The IFCC C-STFT found that most TT4 methods report values within 10 percent of the ID-LC-MS/MS RMP (Figure 1) . In euthyroid pregnant subjects the major influence on TT4 is the TBG concentration that rises approximately two-fold by mid-gestation. As a consequence, TT4 steadily increases from the first trimester to plateau at approximately 1.5-fold pre-pregnancy levels by mid-gestation [104,106,111-114]. Thus the non-pregnant TT4 reference range, adjusted by a factor of 1.5 can be used to assess thyroid status in the latter half of gestation [66,67,104,106,115,116].
TT3 reference ranges generally approximate 1.2 – 2.7 nmol/L (80 –180 ng/dL) . However, TT3 methods display far more between-method variability than TT4, and most display more than a 10 percent bias relative to the reference method [79,80,86]. The IFCC C-STFT continues to work with manufacturers to the reduce variability and improve the calibration of TT3 methods against the RMP.
Free Thyroid Hormone Tests (FT4 and FT3)
In accord with the free hormone hypothesis, it is the free fraction of the thyroid hormones (0.02% of TT4 and 0.2% of TT3) that exerts biologic activity at the cellular level , whereas protein-bound hormone is considered as biologically inactive. Since binding-protein abnormalities are highly prevalent (Table 1) , free hormone measurement is considered preferable to total hormone testing [22,118]. However, free hormone measurement that is independent of thyroid hormone binding proteins remains challenging [22,118-120]. Free hormone methods fall into two categories – direct methods, that employ a physical separation of the free from protein-bound hormone, and estimate tests, that either calculate a free hormone “index” from a measurement of total hormone corrected for binding proteins with either a TBG measurement or a binding-protein estimate, or immunoassays that employing an antibody to sequester a small amount of the total hormone that is purportedly proportional to the free hormone concentration [22,75,118]. All free hormone tests are subject to limitations. Both index tests (FT4I and FT3I) and FT4 and FT3 immunoassays are typically protein-dependent to some extent, and may under- or overestimate free hormone, when binding proteins are abnormal [52,92,118-128]. Even direct methods that employ equilibrium dialysis or ultrafiltration to separate free from protein-bound hormone are not immune from technical problems relating to dilution, adsorption, membrane defects, temperature, the influence of endogenous binding protein inhibitors, fatty acid formation and sample-related effects [22,128-133]. The IFCC C-STFT has now established a reference measurement procedure (RMP) for free thyroid hormones that is based on equilibrium dialysis-dilution-mass spectrometry (ED-ID-MS) and primary calibrators [15,51,54,134]. An evaluation of current FT4 immunoassays has revealed major between-method variability and significant biases relative to the RMP that are far in excess of FT4 biological variation [50,53]. Recalibrating methods against the RMP was shown to significantly reduce biases that currently preclude implementing universal reference intervals that would apply across methods. The C-STFT is actively working with the in vitro diagnostic industry to re-standardize free hormone methods against the RMP to reduce current biases.
Direct FT4 and FT3 Methods
Direct free hormone methods have employed equilibrium dialysis [51,54,135-137], ultrafiltration [14,17,18,23,131,138-142] or gel filtration  to separate free hormone from the dominant protein-bound moiety. These separation techniques can be prone to inaccuracies causing under- or overestimate of free hormone due factors relating to dilution, adsorption, membrane defects, temperature, pH, the influence of endogenous binding protein inhibitors, fatty acid formation and sample-related effects [22,118,128,130-133,141,142,144-146]. The IFCC C-STFT has now established the RMP for FT4 as ED ID-LC-MS/MS. Specifically, equilibrium dialysis of serum is performed under defined conditions before FT4 is measured in the dialysate by isotope-dilution-liquid chromatography/tandem mass spectrometry [15,51,54]. Manufacturers are recommended to use this RMP to recalibrate their FT4 immunoassay tests [52-54,134]. Because direct free hormone methods are technically demanding, inconvenient and expensive, they are typically only readily available in reference laboratories. Most FT4 and FT3 testing is made using estimate tests – either the two-test “index” approach or an immunoassay “sequestration” method . However, all current FT4 and FT3 estimate tests are binding-protein dependent to some extent [118,147-150], and a direct free hormone test can be especially useful for evaluating thyroid status when immunoassay values appear discordant with the clinical presentation and/or the TSH measurement .
Early equilibrium dialysis methods used I-131 and later I-125 labeled T4 tracers to measure the free T4 fraction, that when multiplied by a total hormone measurement gave an estimate of the free hormone concentration . Subsequently, symmetric dialysis in which serum was dialyzed without dilution (or employing a near-physiologic medium) was used to overcome dilution effects . By the early 1970s higher affinity T4 antibodies (>1×1011 L/mol) and high specific activity T4-I125 tracers were used to develop sensitive RIA methods that could to directly measure FT4 and FT3 in dialyzates and ultrafiltrates [82,136-138,142,151-154]. Subsequent improvements have involved employing more physiologic buffer diluents and improving the dialysis cell design [132,137]. More recently, isotope-dilution liquid chromatography/tandem mass spectrometry (ID-LC-MS/MS)  has been used to measure FT4 in ultrafiltrates [14,156,157] and dialyzates [27,50,51,134]. The FT4 RMP recently established by the IFCC C-STFT is based on ED followed by ID-LC-MS/MS [15,51].
Figure 2. FT4 and FT3 Immunoassay Method Comparison
Figure 2. Between Assay Comparison of FT4 and FT3 Measurements in Healthy Euthyroid Subjects. A=FT4 and D=DT3: assay means versus the mean by the RMPs. Different assays are coded A-O on the x axis, manufacturer codes used to designate assays were different for FT4 and FT3 assays. The dotted lines represent mean +/- 10% of the RMP ED-ID-MS). B=FT4 and E=FT3: scatter plot (x=mean of the RMP vs. y= mean of 6 singlicate results per assay. Line of equality indicated by dotted line. The results for the most deviating assays are indicated by circles and triangles; all other assays are indicated with the same symbol, X. C=FT4 and F=FT3: percent-difference plot indicating the strongest negatively (circles) and positively (triangles) biased assays .
A number of studies have used ultrafiltration to remove protein-bound T4 prior to LC-MS/MS measurement of FT4 in the ultrafiltrate [14,17,18,23,55,131,138-142]. Direct FT4 measurements employing ultrafiltration are sometimes higher than those made by equilibrium dialysis, because ultrafiltration avoids dilution effects . Furthermore, ultrafiltration is not influenced by dialyzable inhibitors of T4-protein binding that can be present in conditions such as non-thyroidal illness (NTI) . However, ultrafiltration can be prone to errors when there is a failure to completely exclude protein-bound hormone and/or adsorption of hormone onto the filters, glassware and tubing . In addition, ultrafiltration is temperature sensitive and ultrafiltration performed at ambient temperature (25°C) will report FT4 results that are 67 percent lower than ultrafiltration performed at 37°C [133,158]. However, FT4 concentrations measured by ID-LC-MS/MS following either ultrafiltration at 37°C or equilibrium dialysis usually correlate .
Gel Absorption Methods.
Some early direct FT4 methods used Sephadex LH-20 columns to separate free from bound hormone before eluting the free T4 from the column for measurement by a sensitive RIA. However, because of a variety of technical issues, assays based on this methodologic approach are not currently used .
Indirect FT4 and FT3 Estimate Tests
Two-Test Index Methods (FT4I and FT3I)
Free hormone indexes (FT4I and FT3I) are unitless mathematical calculations made by correcting the total hormone test result for the binding protein, primarily TBG, concentration. These indexes require two separate tests and have been used to estimate free hormone concentrations for more than 40 years . The first test involves the measurement of total hormone (TT4 or TT3) ,whereas the second test assesses the binding protein concentration using either (i) a direct TBG immunoassay, (ii) a Thyroid Hormone Binding Ratio (THBR) or “Uptake” test or (iii) an isotopic determination of the free hormone fraction [118,160].
There is conflicting data concerning whether indexes employing THBR in preference to direct TBG are diagnostically superior . Free hormone indexes calculated using direct TBG measurement (TT4/TBG) may offer improved diagnostic accuracy over THBR when the total hormone concentration is abnormally high (i.e. hyperthyroidism), or when drug therapies interfere with THBR tests [101,162-165]. Regardless, the TT4/TBG index is not totally independent of the TBG concentration, nor does it correct for Albumin or Transthyretin binding protein abnormalities (Table 1) .
Thyroid Hormone Binding Ratio (THBR) / “Uptake” Tests
The first “T3 uptake” tests developed in the 1950s employed the partitioning of T3-I131 tracer between the plasma proteins in the specimen and an inert scavenger (red cell membranes, talc, charcoal, ion-exchange resin or antibody) [119,166,167]. The “uptake” of T3 tracer onto the scavenger provided an indirect, reciprocal estimate of the TBG concentration of the specimen. Initially, T3 uptake tests were reported as percent uptakes (free/total tracer). Typically, sera with normal TBG concentrations had approximately 30 percent of the T3 tracer taken up by the scavenger. During the 1970s methods were refined by replacing I131-T3 tracers by I125-T3, calculating uptakes based on the ratio between absorbent and total minus absorbent counts, and expressing results expressed as a ratio with normal sera having an assigned value of 1.00 [160,167]. Historically, the use of T3 as opposed to T4 tracer was made for practical reasons relating to the ten-fold lower the affinity of TBG for T3 versus T4, facilitating a higher percentage of T3 tracer being taken up by the scavenger and allowing lower isotopic counting times. Because current methods use non-isotopic proprietary T4 or T3 “analogs”, counting time is no longer an issue and current tests may use a “T4 uptake” approach – which may be more appropriate for correcting for T4-binding protein effects. Differences between T3 and T4 “uptakes” have not been extensively studied . Although all THBR tests are to some degree TBG dependent, the calculated FT4I and FT3I usually provides an adequate correction for mild TBG abnormalities (i.e. pregnancy and estrogen therapy) [104,122,169-171], although they may fail to correct for grossly abnormal binding proteins  in euthyroid patients with congenital TBG extremes [120,122,172], Familial Dysalbuminemic Hyperthyroxinemia (FDH) [75,92,173-176], thyroid hormone autoantibodies [95,97,177,178], non-thyroidal illness (NTI) [120,128,179,180] or medications that directly or indirectly influence thyroid hormone binding to plasma proteins [75,99,120,164,181,182].
Isotopic Index Methods
The first free hormone tests developed in the 1960s were indexes calculated from the product of the free hormone fraction, measured isotopically by dialysis, and TT4 measured by PBI and later RIA [135,183,184]. These early isotopic detection systems were technically demanding and included paper chromatography, electrophoresis, magnesium chloride precipitation and column chromatography [135,153,185-187]. The free fraction index approach was later extended to ultrafiltration and symmetric dialysis, the latter measuring the rate of transfer of isotopically-labeled hormone across a membrane separating two chambers containing the same undiluted specimen [92,138,140,184,188-190]. Ultrafiltration and symmetric dialysis had the advantage of eliminating dilution effects that influenced tracer dialysis values [129,191]. However, free hormone indexes calculated using an isotopic free fraction were not completely independent of the TBG concentration and furthermore were influenced by tracer purity and the buffer matrix employed [137,192].
Clinical Utility of Two-Test Index Methods (FT4I and FT3I
Some favored the two-test FT4I approach for evaluating the thyroid status of patients with abnormal binding protein states like pregnancy and NTI [104,193]. Continued use of the FT4I remains controversial . However, until FT4 immunoassays are re-standardized to remove biases [50,52,53], FT4I remains a useful confirmatory test when binding proteins are abnormal and when diagnosing central hypothyroidism .
Free Thyroid Hormone Immunoassay Methods (FT4 and FT3)
Most free hormone testing is made using FT4 and FT3 immunoassays . These immunoassays are based on “one-step”, “labeled antibody” or “two-step” principles, as described below [75,118,196]. For more than twenty years controversy has surrounded the standardization and diagnostic accuracy of these methods, especially in pathophysiologic conditions associated with the binding protein abnormalities such as pregnancy [22,104], or due to polymorphisms, drug interactions, high free fatty acid (FFA) levels or thyroid binding inhibitors such as those present in NTI [25,53,75,92,119,120, 126-128,130,147,150,196-200]. Studies showing correlations between FT4 immunoassay values and both TBG and albumin concentrations, as well as weak inverse FT4/TSH log/linear relationships [17,18,23,126], have emphasized the need to evaluate each method with clinical specimens containing abnormal binding proteins. Currently, most FT4 and FT3 immunoassays display significant negative or positive biases that exceed the intra-individual biological variability (Figure 2) [50,52,53]. The IFCC C-STFT is actively working with the IVD industry to recalibrate their free hormone immunoassays against the RMP [15,50,53,60]. However, although recalibration to the RMP has been shown to greatly reduce between-method biases [50,52,53], implementation of a global re-calibration effort has been delayed by practical, educational and regulatory complexity.
One-Step, FT4 and FT3 Methods
The “one-step” approach uses a proprietary labeled hormone analog, designed for minimal interaction with thyroid hormone binding proteins, that competes with hormone in the specimen for a solid-phase anti-hormone antibody in a classic competitive immunoassay format [22,75,118,119,201,202]. After washing away unbound constituents, the free hormone concentration should be inversely proportional to the labeled analog bound to the solid support. Although conceptually attractive, the diagnostic utility of the one-step approach has been shown to be critically dependent on the degree that the analog is “inert” with respect to binding protein abnormalities [17,18,23,118,119,147,180,200,203-208].
Labeled Antibody FT4 and FT3 Methods
Labeled antibody methods are “one-step” methods that use labeled-antibody in preference to a labeled hormone analog. The free hormone in the specimen competes with solid-phase hormone for the labeled antibody and is quantified as a function of the fractional occupancy of hormone-antibody binding sites in the reaction mixture [22,75,118,120,202,209]. The labeled antibody approach is used as the basis for a number of automated immunoassay platforms because it is easy to automate and considered less binding-protein dependent than the labeled analog approach, because the solid phase hormone does not compete with endogenous free hormone for hormone binding proteins [22,87,118,210,211].
Two-Step, Back-Titration FT4 and FT3 Methods
The two-step approach was first developed by Ekins and colleagues in the late 1970s [75,119,128,202]. Two-step methods typically employ immobilized T4 or T3 antibody (for FT4 and FT3 immunoassays, respectively) to sequester a small proportion of total hormone from a diluted serum specimen without disturbing the original free to protein-bound equilibrium [75,118]. After removing unbound serum constituents by washing, a labeled probe (125-I T4, or more recently a macromolecular T4 conjugate) is added to quantify unoccupied antibody-binding sites that are inversely related to the free hormone concentration – a procedure that has been referred to as “back-titration .
Clinical Utility of FT4 and FT3 Measurements
Most FT4 methods give diagnostically reliable results when binding proteins are near-normal, provided that a method-specific reference range is employed . However, both TT3 and FT3 immunoassay methods tend to be inaccurate in the low range [86,212] and have no value for diagnosing or monitoring treatment for hypothyroidism [70,213], although T3 measurement can be useful for diagnosing or confirming unusual cases of hyperthyroidism.
Free hormone tests (FT4 or FT3) are used in preference to total hormone (TT4 or TT3) measurements in order to improve diagnostic accuracy for detecting hypo- and hyperthyroidism in patients with abnormal thyroid hormone binding proteins (Table 1). FT4 is typically employed as a second-line test for confirming primary thyroid dysfunction detected by an abnormal TSH ,but is the first-line test when thyroid status is unstable (early phase of treating hypo- or hyperthyroidism), in the presence of pituitary/hypothalamic disease when TSH is unreliable, or when patients are taking drugs such as dopamine or glucocorticoids that are known to affect TSH secretion [24,100,101,165,214-219].
Mild “subclinical” thyroid dysfunction is characterized by TSH/FT4 discordances (abnormal TSH/normal FT4). This reflects the intrinsic complex nature of the inverse log/linear TSH/FT4 relationship [24,220,226] – a relationship that is modified by age and gender [227,228]. Thus, small changes in FT4, even within normal limits, are expected to produce a mild degree of TSH abnormality – between 0.05 and 0.3 mIU/L (for subclinical hyperthyroidism) and 5 and 10 mIU/L (for subclinical hypothyroidism). An unexpected TSH/FT4 discordance, if confirmed, should prompt an investigation for interference with FT4, TSH or both tests [229,230]. FT4 interference can result from severe binding protein abnormalities such as congenital TBG excess or deficiency [75,94,122,159,231,232], dysalbuminemias [92,233-236], thyroid hormone autoantibodies [95,97,98,177,178,230,237] or drug interferences [75,99,120].
Hospitalized Patients with Nonthyroidal Illnesses (NTI)
The diagnostic performance of current FT4 methods has not been evaluated in hospitalized patients with NTI where binding protein inhibitors and drug therapies can negatively impact the reliability of both thyroid hormone and TSH testing [24,75,126,130,180,218,238,239]. Three categories of hospitalized patients deserve special attention: a) patients with NTI without known thyroid dysfunction who have a high or low T4 status; b) patients with primary hypothyroidism and concurrent NTI and, c) patients with hyperthyroidism and concurrent NTI [238,240,241]. Because the diagnostic reliability of FT4 testing is still questionable in sick hospitalized patients, a combination of both T4 (FT4 or TT4) and TSH may be needed to assess thyroid status in this setting [24,53,180,242]. In most clinical situations where FT4 and TSH results are discordant, the TSH test is the most diagnostically reliable, provided that the patient does not have pituitary failure or is receiving medications such as glucocorticoids and dopamine that directly inhibit TSH secretion [101,165,218]. Repetitive TSH testing may be helpful in resolving the cause of an abnormal FT4, because the TSH abnormalities of NTI are typically transient whereas the TSH abnormality will persist if due to underlying thyroid dysfunction [243-246]. It may be useful to test for TPOAb as a marker for underlying thyroid autoimmunity
FT4 and FT3 reference ranges
Current reference ranges for FT4 and FT3 immunoassays are method-dependent because of calibration biases [50,52,53] (Figure 2). This calibration problem negatively impacts the clinical utility of FT3 and FT4 tests because it precludes establishing universal reference ranges that would apply across methods.
Pediatric FT4 and FT3 Reference Ranges
The determination of normal reference limits for pediatric age-groups is especially challenging, given the limited number of studies involving sufficient numbers of healthy children [247-249]. Most studies report that serum TSH peaks after birth and steadily declines throughout childhood to reach adult levels at puberty. Likewise, FT3 declines across the pediatric age groups during childhood and approaches the adult range at puberty, whereas FT4 levels for infants less than a year old are higher than for children 1 to 18 years old who have FT4 similar to that observed for adults [247-252].
Pregnancy FT4 Reference Ranges
As with non-pregnant patients, TSH is the first-line test to use for assessing thyroid status during pregnancy . However, FT4 measurement is needed for monitoring anti-thyroid drug treatment of hyperthyroid pregnant patients who have undetectable TSH. The question whether an isolated low FT4 during pregnancy is a maternal or fetal risk factor, remains controversial [254-259]. However, a number of studies suggest that low FT4 may be a risk factor for gestational diabetes and fetal complications [260-264]. Non-pregnant FT4 reference ranges do not apply to pregnancy since FT4 progressively declines as gestation progresses, necessitating the use of trimester-specific reference ranges [104,113,265-271]. Currently it is not possible to propose universal trimester-specific FT4 reference ranges given current between-method differences [50,53,271] (Figure 2) compounded by differences related to the ethnicity [193,270,272-275], iodine intake [276-278], smoking  and BMI [269,270,280-283] between study cohorts. Establishing institution-specific trimester-specific reference ranges from the 2.5 to 97.5 percentiles of least 400 pregnant patients from each trimester  is not practical for most institutions. The feasibility of establishing universal trimester-specific reference ranges will improve after the proposed re-standardization of FT4 methods against the RMP . However, binding protein effects will remain and population-specific factors will still have to be considered.
Interferences with Total and Free Thyroid Hormone Tests
Only the physician can suspect interference with a test result and request that the laboratory perform interference checks! This is because the hallmark of interference is discordance between the test result and the clinical presentation of the patient. Failure to recognize interferences can have adverse clinical consequences [229,284-289].
Laboratory checks for interference include showing discordance between different manufacturers methods [290-293], re-measurement of analyte after adding blocking agents [293-297] and performing linearity studies or precipitating immunoglobulin with polyethylene glycol (PEG) [229,290,291,293,294,298-300]. A change in analyte concentration in response to one of these maneuvers suggests interference, but a lack of effect does not rule out interference. Interferences can be classified as either (a) non-analyte-specific or (b) analyte-specific [301,302].
Immunoassays can be affected by interferences from both paraproteins [303-305] and abnormal immunoglobulins [306,307].
Congenital TBG excess or deficiency.
Free hormone immunoassays and free T4 index tests may be susceptible to interference from grossly abnormal TBG concentrations, such as those seen in congenital TBG excess or deficiency states [75,94,122,159,231,232].
Estrogen stimulation causes TBG concentrations to progressively rise to plateau 2.5-fold higher than pre-pregnancy values by mid-gestation [193,308,309]. As a consequence, both TT4 and TT3 increase to approximately 1.5-fold of pre-pregnancy values by mid-gestation [113,310]. Despite the rise in total hormone, both FT4 and FT3 decline to a method-related degree during gestation [104,265-269]. It should be noted that lower FT4 levels would be expected during pregnancy from a consideration of the law of mass action as applied to T4-binding protein interactions . However, the degree of FT4 decline during pregnancy is variable and method-dependent due to standardization differences (Figure 2) and in some cases method sensitivity to the declining albumin concentrations typical of late gestation [18,193,311].
Familial Dysalbuminemic and Transthyretin-Associated Hyperthyroxinemias.
Autosomal dominant mutations in the Albumin or Transthyretin (prealbumin)  gene can result in altered protein structures with enhanced affinity for thyroxine and/or triiodothyronine. These abnormal proteins can interfere with FT4 and/or FT3 measurements and result in inappropriately high FT4 and/or FT3 immunoassay values [92,173,237,312]. Familial Dysalbuminemic Hyperthyroxinemia (FDH) is a rare condition with a prevalence of ~1.8 % in the Hispanic population . It arises from a number of genetic variants, with the R218H being the most common, some variants result in extremely high TT4, whereas other mutations (i.e. L66P) affect mainly T3 . Affected individuals are euthyroid and have normal TSH and FT4 when measured by direct techniques such as equilibrium dialysis . Unfortunately, most FT4 estimate tests (immunoassays and indexes) report falsely high values for FDH patients that may prompt inappropriate treatment for presumed hyperthyroidism if the condition is not recognized .
Heterophile Antibodies (HAbs)
Heterophile antibodies (HAb) are human poly-specific antibodies targeted against animal antigens, the most common being human anti-mouse antibodies (HAMA) [293,302,314,315]. Alternatively, HAb can target human antigens  such as rheumatoid factor (RF), an immunoglobulin commonly associated with autoimmune conditions that is widely considered a heterophile antibody . RF has been shown to interfere with free and total thyroid hormone tests  as well as TSH  and Tg . HAbs have a prevalence of 30-40 percent [319-321] and have the potential to interfere with a broad range of methods that use IMA principles [290,300,306,322]. In recent years assay manufacturers have increased the immunoglobulin blocker reagents added to their tests and this has reduced interference from 2 to 5 percent [290,297,323]. However, interference is still seen in approximately one percent of patients who have high enough HAb concentrations to overcome the assay blocker [296,298,322,324]. HAMA interference mostly affects non-competitive immunometric assays (IMA) that employ monoclonal antibodies of murine origin . Assays based on the competitive format that employ high affinity anti-antigen polyclonal antibody reagents, are rarely affected [296,319]. HAb has the potential to interfere with both free [178,321,326-328] and total [178,326,327] thyroid hormone tests, as well as THBR , TSH [289,294,300,328-330] and Thyroglobulin (Tg) [295,296,323,324,331,332], TgAb  and calcitonin (CT) [300,334-337] methods. Interference from HAb or HAMA typically causes falsely high results for one or more analytes. Less commonly falsely low test results may be seen . The test marketed by one manufacturer can be severely affected, whereas the test from a different manufacturer may appear unaffected. This is why the first step for investigating for interference is re-measurement of the analyte in a different manufacturers method. It should be noted that patients receiving recent vaccines, blood transfusions or monoclonal antibodies (given for treatment or scintigraphy), as well as veterinarians and those coming into contact with animals, are especially prone to test interferences caused by induced HAb and HAMA [298,338].
Interference can be caused by antibodies against assay reagents. For example, there are a number of reports of anti-Rhuthenium antibodies interfering with TSH, FT4 and FT3 by [339-343]. In Streptavidin-Biotin based assays interference can result from antibodies targeting either Streptavidin  or biotin reagents . Alternatively, high dose biotin ingestion has been known to produce interference with thyroid and other tests in an analyte-specific, platform-specific manner [346-350].
Analyte-specific interferences typically result from autoantibodies targeting the analyte. Depending on the analyte and test formulation, autoantibody interferences typically cause falsely-high test results, but can cause falsely-low test results, as in the case of Tg autoantibodies . It should be noted that transplacental passage both heterophile antibodies or anti-analyte autoantibodies (i.e. TSHAb or T4Ab) have the potential to interfere with neonatal screening tests [351-354]. Specifically, maternal TSH autoantibodies can cross the placenta and may cause a falsely high TSH screening test in the newborn mimicking congenital hypothyroidism, whereas maternal T4 autoantibodies could cause falsely high neonatal T4 masking the presence of congenital hypothyroidism [230,353].
T4 and T3 Autoantibodies (T4Ab/T3Ab)
T4 and T3 autoantibodies can falsely elevate total hormone, free hormone or THBR measurements depending on the method employed [95,97,98,177,178,230,237]. The prevalence of thyroid hormone autoantibodies approximates 2 percent in the general population but may be as much as 30 percent in patients with autoimmune thyroid disease or other autoimmune conditions [316,355-358]. However, despite their high prevalence, significant interference caused by thyroid autoantibodies is not common and depends on the qualitative characteristics of the autoantibody present (i.e. its affinity for the test reagents). Further, different methods exhibit such interferences to a greater or lesser degree [95,97]. Because autoantibody interference is difficult for the laboratory to detect proactively, it is the physician who should first suspect interference characterized by unexpected discordance between the clinical presentation of the patient and the test result(s) [96, 178].
SERUM TSH (THYROID STIMULATING HORMONE/THYROTROPIN) MEASUREMENT
Over the last four decades the dramatic improvements in TSH assay sensitivity and specificity have revolutionized thyroid testing and firmly established TSH as the first-line test for ambulatory patients not receiving drugs known to alter TSH secretion [24,70,71,120,216,218,359]. Serum TSH has become the therapeutic target for levothyroxine (L-T4) replacement therapy for hypothyroidism and suppression therapy for differentiated thyroid cancer . The diagnostic superiority of TSH versus FT4 measurement arises from the inverse, predominantly log/linear, TSH/FT4 relationship, that is modified to some extent by factors such as age, sex, active smoking and TPOAb status [7,24,221-228].
TSH assay “quality” has historically been defined by clinical sensitivity – the ability to discriminate between hyperthyroid and euthyroid TSH values [24,360-364]. The first generation of RIA methods had a detection limit approximating 1.0 mIU/L [365-367] that limited their clinical utility to diagnosing primary hypothyroidism [368-370] and necessitated the use of TRH stimulation to diagnose hyperthyroidism that was characterized by an absent TRH-stimulated TSH response [371-376]. With the advent of immunometric assay (IMA) methodology that uses a combination of poly- and/or monoclonal antibodies targeting different TSH epitope(s) in a “sandwich” format [377-379], a ten-fold improvement in TSH assay sensitivity (~ 0.1 mIU/L) was achieved when using isotopic (I125) signals . This level of sensitivity facilitated the determination of the lower TSH reference limit (as 0.3-0.4 mIU/L), and the detection of overt hyperthyroidism without the need for TRH stimulation [7,374-376,380-386], but was still insufficient for distinguishing between differing degrees of hyperthyroidism (i.e. subclinical versus overt). Sensitization continued until a third-generation of TSH IMAs, using non-isotopic signals, were developed that could achieve a sensitivity of 0.01 mIU/L [7,8,374,387-389]. Initially different non-isotopic signals were used that gave rise to a lexicon of terminology to distinguish between assays: immunoenzymometric assays (IEMA) used enzyme signals; immunofluorometric assays (IFMA) used fluorophors as signals, immunochemiluminometric assays (ICMA) used chemiluminescent molecules as signals and immunobioluminometric assays (IBMA) used bioluminescent signal molecules [8,390]. Current TSH methods are automated ICMAs  that all achieve third-generation functional sensitivity (FS = ≤0.01 mIU/L) – a sensitivity the FS level that has subsequently become the standard of care [7,8,52,53,388,391-396].
Functional Sensitivity (FS) – determines the lowest reportable assay limit
During the period of active TSH assay improvement, different non-isotopic IMAs made competing claims for sensitivity. Methods were described as: “sensitive”, “highly sensitive”, “ultrasensitive” or “supersensitive” – marketing terms that had no scientific definition. This confusion led to a debate concerning what was the most clinically relevant parameter to use to determine the lowest reliable reportable TSH value for clinical practice [8,397-403]. Functional sensitivity (FS), defined as the lowest analyte concentration measured with 20 percent coefficient of variation  is now recognized as the parameter that best represents the between-run precision for measuring low analyte concentrations in clinical practice [24,395,404]. FS is used to define the lower clinical reporting limit for not only for TSH assays, but also Tg and TgAb measurements, for which assay sensitivity is critical [8,24,397,404,405]. Protocols used for establishing FS specify that precision be determined in human serum, not quality control materials based on artificial protein matrices, since immunoassays tend to be matrix-sensitive [406,407]. The time-span used for determining precision is also analyte-specific and should reflect the frequency of testing employed in clinical practice – 6 to 8 weeks for TSH, but 6 to 12 months for the Tg and TgAb assays when used as tumor markers for monitoring differentiated thyroid cancer (DTC). This time-span is important because low-end, between-run assay precision erodes over time as a result of a myriad of variables, reagent lot-to-lot variability being a key variable [9,408-410]. Note that the FS parameter is more stringent than other biochemical sensitivity parameters such as limit of detection (LOD – a within-run parameter) and limit of quantitation (LOQ – a between-run parameter without stipulations regarding matrix and time-span for determining precision) [404,411]. A ten-fold difference in FS has been used to define each more sensitive “generation” of TSH  or Tg [32,404,412,413] method. Thus, TSH RIA methods with FS approximating 1.0 mIU/L were designated “first generation”, TSH IMA methods with functional sensitivity approximating 0.1 mIU/L were designated “second generation”, and TSH IMAs with FS approximating 0.01 mIU/L are designated “third generation” assays [8,57,395,397,405,414]. Analogous to TSH, Tg assays [Section 6A] with FS approximating 1 μg/L are designated “first generation”, whereas Tg IMAs with FS approximating 0.10 μg/L meet the criteria for a “second generation” method [32,58,296,395,404,413,415,416].
TSH Biologic Variability
As compared with between-person variability, TSH intra-individual variability is relatively narrow (20-25 percent) in both non-pregnant and pregnant subjects, as compared with between-person variability [29,222,417,418]. In fact, the serum TSH of euthyroid volunteers was found to vary only ~0.5 mIU/L when tested every month over a span of one year . Twin studies suggest that there are genetic factors that determine hypothalamic-pituitary-thyroid setpoints [419-421]. These studies report that the inheritable contribution to the serum TSH level approximates 65 percent [420,422]. This genetic influence appears, in part, to involve single nucleotide polymorphisms in thyroid hormone pathway genes such as the phosphodiesterase gene (PDE8B) [423-425], polymorphisms causing gain [426-433] or loss [434-436] of function TSH receptors [423,437,438] and the type II deiodinase enzyme polymorphisms [423,439]. Undoubtedly, such polymorphisms account for some of the euthyroid outliers that skew TSH reference range calculations [423,434,440].
Figure 3. TSH Between-Method Variability
Figure 3. A. Geometric mean of the TSH results for the range 0.5– 6.6 mIU/L, (x axis, different assays; dotted lines, overall mean and 10% error). In the plots, the 1-sided 95% CIs of the means are shown (note: the wide interval of assay O is due to results from only 2 runs with a high between-run variation and df = 1 by the Satterthwaite approximation). For the assays outside the 10% limit, the mean value is listed. B. Plot showing the %-difference between TSH methods. The most discrepant assays are shown by triangles and circles. Other assays are shown with the same symbol (x) [29,52].
The narrow TSH within-person variability and low (< 0.6) index of individuality (IoI) [222,417, 418,441-443] limits the clinical utility of using the TSH population-based reference range to detect thyroid dysfunction in an individual patient [222,418,443,444]. When evaluating patients with marginally (confirmed) low (0.1–0.4 mIU/L) or high (4–10 mIU/L) TSH abnormalities, it is more important to consider the degree of TSH abnormality relative to patient-specific risk factors for cardiovascular disease rather than the degree of the abnormality relative to the TSH reference range [69,445,446].
The TSH Population Reference Range
The complex log/linear TSH/FT4 relationship [7,24,221-228] dictates that TSH will be the first abnormality to appear with the development of mild (subclinical) hypo- or hyperthyroidism. It follows that the setting of the TSH reference limits critically influences the frequency of diagnosing subclinical thyroid disease [69,445,448,454].
Guidelines recommend that “TSH reference intervals should be established from the 95 percent confidence limits of the log-transformed values of at least 120 rigorously screened normal euthyroid volunteers who have: (a) no detectable thyroid autoantibodies, TPOAb or TgAb (measured by sensitive immunoassay); (b) no personal or family history of thyroid dysfunction; (c) no visible or palpable goiter and, (c) who are taking no medications except estrogen” [24,450].
Multiple factors influence population TSH reference limits, especially the upper (97.5th percentile) limit. Different methods report different ranges for the same population as a result of between-methods biases (Figure 3) [396,448,451,455]. A key factor affecting the upper limit is the stringency used for eliminating individuals with thyroid autoimmunity (thyroid autoantibody positive ) from the population [452,456-461]. Other factors relate to population demographics such as sex , ethnicity [452,462-464], iodine intake , BMI [466-477] and smoking status [462,478,479]. Age is a major factor the influences the TSH upper limit [460,463,480-482] leading to the suggestion that age-specific TSH reference limits should be used (Figure 4) [69,451,480]. However, the relationship between TSH and age is complex. Most studies in iodine sufficient populations have shown an increase in TSH with age [440,452,460,483], whereas other studies have reported no change or a decreased TSH with aging [457,484,485]. This conflicting data could merely represent population differences – with a rising TSH with age reflecting an increasing prevalence of thyroid autoimmunity in iodine-sufficient populations , whereas in iodine deficient populations, increasing autonomy of nodular goiter can result in decreased TSH with aging [486-488]. Some studies have reported that a mild TSH elevation in elderly individuals may convey a survival benefit [481,489-492], whereas other studies dispute this [493,494]. However, TSH is a labile hormone and studies cannot assume that a TSH abnormality found in a single determination is representative of thyroid status in the long-term [495,496].
Figure 4.Suggested management algorithm from reference # 69 Initial management of persistent subclinical hypothyroidism in non-pregnant adults: persistent subclinical hypothyroidism describes patients with elevated serum TSH and within reference range serum FT 4 on two occasions separated by at least 3 months. This algorithm is meant as a guide and clinicians are expected to use their discretion and judgment in interpreting the age threshold around 70 years. * Depending on circumstances, individuals with goiter, dyslipidaemia, and diabetes may also be considered for treatment, along with those with planning pregnancy in the near future.
TSH is a heterogeneous glycoprotein [497,498], and TRH-mediated changes in TSH glycosylation  have the potential to influence immunoactivity [500,501]. A number of pathophysiologic circumstances are known to alter TSH glycosylation [498,500,502-504]. The demonstration that harmonization of TSH methods successfully mininizes between-method differences [52,53] suggests that under normal conditions current TSH IMAs appear to be “glycosylation blind”, and detect different TSH glycoforms in an equimolar fashion [52,53,501]. However, future studies need to include sera from conditions where TRH dysregulation may lead to abnormal TSH glycosylation and bioactivity, such as pituitary dysfunction, NTI and aging [215,239,246,498,505-509].
Pediatric TSH Reference Ranges
The adult TSH population reference range does not apply to neonates or children. Serum TSH values are generally higher in neonates and then gradually decline until the adult range is reached after puberty [250-252, 485, 510-514]. This necessitates using age-specific TSH reference ranges for diagnosing thyroid dysfunction in these different pediatric age groups.
Subclinical Thyroid Dysfunction
Subclinical Hyperthyroidism (SCHY).
The lower (2.5th percentile) TSH reference limit approximates 0.3-0.4 mIU/L, and is fairly independent of the method used [445,452,484,485,515-520]. Subclinical hyperthyroidism (SCHY), is defined as a low but detectable TSH (0.01 –-0.3 mIU/L range) without a FT4 abnormality. The prevalence of endogenous SCHY is low (0.7%) in iodine-sufficient populations , but is higher in patients reporting thyroid disease as an iatrogenic consequence of L-T4 replacement therapy [521-523]. SCHY is a risk factor for osteoporosis and increased fracture risk [474,524-526] as well as atrial fibrillation and cardiovascular disease [445,474,527], especially in older patient patients.
Subclinical Hypothyroidism (SCHO).
Subclinical hypothyroidism is defined as a TSH above the upper (97.5th percentile) TSH reference limit without a FT4 abnormality [69,448,454,460,516,528-530]. However, since the setting of the TSH upper limit remains controversial, the prevalence of SCHO is highly variable – 4 to 8.5 % [452,521], rising to 15% in older populations [446,456]. In most cases, SCHO is associated with TPOAb positivity, indicative of an autoimmune etiology [452,456]. The clinical consequences of SCHO relate to the degree of TSH elevation . Most guidelines recommend L-T4 treatment of SCHO when is TSH is above 10 mIU/L [68,69] but below 10 mIU/L recommend L-T4 treatment based on patient-specific risk factors (Figure 4) . There is active debate concerning the efficacy of treating SCHO to prevent progression [532-535], or improve renal [536,537], cardiovascular [474,524,531,538-543], or lipid [544-546] abnormalities that can be associated with SCHO [69,547].
Thyroid Dysfunction and Pregnancy
It is well documented that overt hypo- or hyperthyroidism is associated with both maternal and fetal complications [548-550]. However, the impact of maternal subclinical thyroid dysfunction remains controversial . No maternal or fetal complications appear associated with subclinical hyperthyroidism during pregnancy [258,551]. First trimester “gestational hyperthyroidism” is typically transient and hCG-related, as described above. In contrast, short-term and long-term outcome studies of maternal subclinical hypothyroidism  are complicated by heterogeneity among studies arising from a myriad of factors influencing TSH cutoffs, such as gestational stage, TSH method used, maternal TPOAb status, and current and pre-pregnancy iodine intake [277,454]. Using gestational age-specific reference intervals the frequency of SCHO in first trimester pregnancy approximates 2-5 percent [552-556]. A number of studies have reported that subclinical hypothyroidism is associated with increased frequency of maternal and fetal complications, especially when TPOAb is positive [557-559]. Maternal complications have included miscarriage [474,548,560-562], preeclamsia [548,563], placental abruption , preterm delivery [552,562,564] and post-partum thyroiditis . Fetal complications have included intrauterine growth retardation and low birth weight [258,548,566-568] and possible impaired neuropsychological development [550,569,570]. It remains controversial whether L-T4 treatment of SCHO in early gestation decreases risk of complications [559,562,564,571].
Trimester-Specific TSH Reference Ranges.
As with non-pregnant patients, TSH is the first-line test used for assessing thyroid status during pregnancy when gestation-related TSH changes occur [66,67,253,254,555,556,572]. In the first trimester, there is a transient rise in FT4 caused by high hCG concentrations stimulating the TSH receptor – because hCG shares some homology with TSH [254,308,309,573,574]. The degree of TSH suppression is inversely related to the hCG concentration and can be quite profound in patients with hyperemesis who have especially high hCG [271,575-577]. As gestation progresses, TSH tends to return towards pre-pregnancy levels . Recent studies from different geographic areas with diverse iodine intakes have using different TSH methods have reported higher trimester-specific TSH upper limits than recommended by previous guidelines [253,269,271,454,556,578-580]. In response, the American Thyroid Association has recently revised their pregnancy guidelines [66,74] to replace trimester-specific reference limits by a universal upper TSH limit of 4.0 mIU/L, when TPOAb is negative and local reference range data is not available. However, at this time between-method biases (Figure 3) clearly preclude proposing universal TSH or FT4 reference ranges that would apply to all methods and all populations [52,53,267,271,447]. It is critical that the IVD manufacturers respond to the urging of the IFCC C-STFT and harmonize their TSH methods to increase the feasibility of establishing TSH universal reference limits for pregnancy [52,53]. Requiring each institution to establish their own trimester-specific reference ranges for thyroid tests is impractical, given the costs, logistics and ethical considerations involved in recruiting the more than 400 disease-free pregnant women needed to establish reliable ranges for each trimester . Only after methods are re-standardized (FT4) or harmonized (TSH), will it be feasible to propose trimester-specific reference ranges that would apply across methods. However, such ranges would still be influenced by differences in ethnicity  and iodine intake, especially pre-pregnancy iodine intake that influences thyroidal iodine stores . There is also a current need to reevaluate optimal TPOAb cutoffs needed to exclude those individuals with thyroid autoimmunity whose inclusion skews TSH upper limits [271,280,454,574,581,582].
Clinical Utility of TSH Measurement
In the outpatient setting the reliability of TSH testing is not influenced by the time of day of the blood draw, because the diurnal TSH peak occurs between midnight and 0400 [583-586]. Third-generation TSH assays (FS ~0.01 mIU/L) have now become the standard of care because they can reliably detect the full spectrum of thyroid dysfunction from overt hyperthyroidism to overt hypothyroidism, provided that hypothalamic-pituitary function is intact and thyroid status is stable [24,57,216,242,359,414,587,588]. TSH is also used for optimizing L-T4 therapy – a drug with a very narrow therapeutic index [359,589,590]. Because TSH secretion is slow to respond to changes in thyroxine status there is no need to withhold the L-T4 dose on the day of the blood test . In addition, targeting the degree of TSH suppression relative to recurrence risk plays a critical role in the management of thyroid cancer [72,591-593].
Hospitalized Patients with Nonthyroidal Ilnesses (NTI)
Routine thyroid testing in the hospital setting is not recommended because thyroid test abnormalities are frequently seen in euthyroid sick patients [238,594]. Non-thyroidal illness, sometimes called the “sick euthyroid syndrome” is associated with alterations in hypothalamic/pituitary function and thyroid hormone peripheral metabolism often exacerbated by drug influences [100,218,239,245,595]. T3 levels typically fall early in the illness followed by a fall in T4 as the severity of illness increases. [24,244,595-597]. As thyroid hormone levels fall TSH typically remains unchanged, or may be low early in the illness, especially in response to drug therapies such as dopamine or glucocorticoid [100,101,218]. During the recovery phase, TSH frequently rebounds above the reference range . However, high TSH may also be seen associated with psychiatric illness . It is important to distinguish the generally mild, transient TSH alterations typical of NTI from the more profound and persistent TSH changes associated with hyper- or hypothyroidism [24,238,244].
Misleading TSH Measurements
TSH can be diagnostically misleading either because of (a) biological or (b) technical factors. from heterophile antibodies (HAbs) or endogenous TSH autoantibodies are the most common causes of a falsely high TSH [299,329,599].
Biologic factors causing TSH diagnostic dilemmas
Unstable thyroid function
TSH can be misleading when there is unstable thyroid status – such as in the early phase of treating hyper- or hyperthyroidism or non-compliance with L-T4 therapy -when there is a lag in the resetting of pituitary TSH to reflect a new thyroid status . During such periods of instability TSH will be misleading and FT4 will be the more diagnostically reliable test.
Pituitary dysfunction is rare in ambulatory patients . TSH measurement is unreliable in cases of both central hypothyroidism and central hyperthyroidism caused by TSH-secreting adenomas [215,217,219,508].
Central Hypothyroidism (CH)
Central hypothyroidism (CH) is rare (1/1000 as prevalent as primary hypothyroidism, 1/160,000 detected by neonatal screening) [509, 601]. CH can arise from disease at either the pituitary or hypothalamic level, or both . A major limitation of using a TSH-centered screening strategy is that this strategy will miss a diagnosis of CH, because the TSH isoforms secreted in CH are abnormally glycosylated and bio-inactive, yet will be detected as paradoxically normal TSH by current IMA methods despite the presence of clinical hypothyroidism [215, 217, 602]. The clinical diagnosis of CH can be confirmed biochemically as a low FT4/normal-low TSH discordance. Serum FT4 should be used to optimize L-T4 replacement therapy. In the absence of clinical suspicion, investigations for pituitary dysfunction should only be initiated after ruling-out technical interference.
TSH-secreting pituitary adenomas
TSHomas are characterized by near-normal TSH despite clinical hyperthyroidism . Since this is a rare (0.7%) type of pituitary adenoma, technical interference causing paradoxically high TSH, such as a TSH autoantibody should be excluded before initiating inconvenient and unnecessary pituitary imaging or dynamic (T3 suppression or TRH stimulation) diagnostic testing. TSHomas are characterized by discordance between the clinical presentation and a paradoxically non-suppressed TSH despite high thyroid hormone levels and clinical hyperthyroidism . This clinical/biochemical discordance reflects adenoma secretion of TSH isoforms with enhanced biologic activity that cannot be distinguished from bioactive TSH by IMA methods. Failure to diagnose the pituitary as the cause of the hyperthyroidism can lead to inappropriate thyroid ablation. The treatment of choice is surgery but in cases of surgical failure somatostatin analog treatment has been found effective . Note that the biochemical profile (high thyroid hormones and non-suppressed TSH) is similar to that seen with thyroid hormone resistance syndromes . When pituitary imaging is equivocal, genetic testing may be necessary to distinguish between these two conditions.
Resistance to Thyroid Hormone (RTH)
Resistance to thyroid hormone is biochemically characterized by high thyroid hormone (FT4 +/- T3) levels and a non-suppressed, sometimes slightly elevated TSH without signs and symptoms of thyroid hormone excess . Early cases of resistance to thyroid hormone were shown to result from mutations in the thyroid hormone receptor B . More recently the definition of RTH has been broadened to include other causes of thyroid hormone resistance – mutations in the thyroid hormone cell membrane transporter MCT8, and a range of genetic thyroid hormone metabolism defects (SBP2) . These resistance syndromes display a spectrum of clinical and biochemical profiles may need to be identified by specialized genetic testing.
Activating or Inactivating TSH Receptor Mutations
Non-autoimmune hyperthyroidism resulting from an activating mutation of the TSH receptor (TSHR) is rare [426-433]. A spectrum of loss-of-function TSHR mutations (TSH resistance) causing clinical and subclinical hypothyroidism despite high thyroid hormone levels, have also been described [434-436]. Because TSHR mutations are a rare cause of TSH/FT4 discordances, technical interferences should first be excluded before considering a TSHR mutation as the cause of these discordant biochemical profiles.
Technical Factors causing TSH Diagnostic Dilemmas
Causes of technical interferences with TSH measurement are similar to those discussed for thyroid hormone tests.
Non Analyte -Specific Interferences
Heterophile Antibodies (HAbs) can cause falsely high TSH IMA tests [289,294,300,328-330, 609]. The HAb in some patient’s sera interfere strongly with some manufacturers tests but appear inert in others . This is why re-measurement in a different manufacturers assay should be the first test for interference. A fall in TSH in response to blocker-tube treatment is typically used to confirm HAb interference
Anti-Reagent Antibody Interferences.
As discussed for free hormone tests,,,,,, some patients have antibodies that target test reagents (such Rhuthenium) that cause interference with TSH and/or free hormone tests. It should be noted that the anti-Rhuthenium antibodies of different patients may affect different analytes to different degrees [339-342].
Tests employing Streptavidin-Biotin
reagents are prone to interferences from antibodies targeting either Streptavidin  or biotin reagents . Alternatively, high dose biotin ingestion has been known to produce interference with thyroid and other tests in an analyte-specific, platform-specific manner [346-350].
Analyte-specific interferences typically result from autoantibodies targeting the analyte. Depending on the analyte and test formulation, autoantibody interferences typically cause falsely-high test results, but can cause falsely-low test results, as in the case of Tg autoantibodies. It should be noted that transplacental passage both heterophile antibodies or anti-analyte autoantibodies (i.e. TSHAb or T4Ab) have the potential to interfere with neonatal screening tests [351-354]. Specifically, maternal TSH autoantibodies can cross the placenta and may cause a falsely high TSH screening test in the newborn mimicking congenital hypothyroidism, whereas maternal T4 autoantibodies could cause falsely high neonatal T4 masking the presence of congenital hypothyroidism [230, 353].
TSH Autoantibodies (TSHAb)/”Macro TSH”.
Analytically suspicious TSH measurements are not uncommon  and have been reported in up to five percent of specimens subjected to rigorous screening . In recent years there have been a number of reports of TSHAb, often referred to as “macro” TSH, causing spuriously high TSH results in a range of different methods [610,611]. The prevalence of TSHAb approximates 0.8 percent, but can as high as ~1.6 percent in patients with subclinical hypothyroidism. Showing a lowering of TSH in response to a polyethylene glycol (PEG) precipitation of immunoglobulins is the most convenient test for TSHAb [599,611]. Alternatively, column chromatography can show TSH immunoactivity in a high molecular weight peak representing a bioinactive TSH-immunglobulin complex [599,611].
TSH variants are a rare cause of interference . Nine different TSH beta variants have been identified to date . These mutant TSH molecules may have altered immunoactivity and be detected by some TSH IMA methods but not others . The bioactivity of these TSH mutants is variable and can range from normal to bio-inert , resulting in discordances between the TSH concentration and clinical status  and/or discordant TSH/FT4 relationships . These TSH genetic variants are one of the causes of congenital hypothyroidism [614-616].
THYROID SPECIFIC AUTOANTIBODIES (TRAB , TPOAB AND TGAB)
Tests for antibodies targeting thyroid-specific antigens such as thyroid peroxidase (TPO), thyroglobulin (Tg) and TSH receptors (TSHR) are used as markers for autoimmune thyroid conditions [37,617]. Over the last four decades, thyroid antibody test methodologies have evolved from semi-quantitative agglutination, complement fixation techniques and whole animal bioassays to specific ligand assays using recombinant antigens or cell culture systems transfected with the human TSH receptor [37,618621]. Unfortunately, the diagnostic and prognostic value of these tests has been hampered by methodologic differences as well as difficulties with assay standardization . Although most thyroid autoantibody testing is currently made on automated immunoassay platforms, methods vary in sensitivity, specificity and the numeric values they report because of standardization issues [44,582,620,623]. Thyroid autoantibody testing can be useful for diagnosing or monitoring treatment for a number of clinical conditions, however these tests should be selectively employed as adjunctive tests to other diagnostic testing procedures.
TSH Receptor Autoantibodies (TRAb)
The TSH receptor (TSHR) serves as a major autoantigen [624,625]. Thyroid gland stimulation occurs when TSH binds to TSHR on thyrocyte plasma membranes and activates the cAMP and phospholipase C signaling pathways . The TSH receptor belongs to the G protein-coupled class of transmembrane receptors. It undergoes complex posttranslational processing in which the ectodomain of the receptor is cleaved to release a subunit into the circulation . A TSH-like thyroid stimulator found uniquely in the serum of Graves’ disease patients was first described using a guinea pig bioassay system in 1956 . Later, using a mouse thyroid bioassay system this serum factor displayed a prolonged stimulatory effect as compared to TSH and hence was termed to be a “long-acting thyroid stimulator” or LATS [627,628]. Much later, the LATS factor was recognized not to be a TSH-like protein but an antibody that was capable of stimulating the TSH receptor causing Graves’ hyperthyroidism . TSH receptor antibodies have also become implicated in the pathogenesis of Graves’ opthalmopathy [629-632]. TRAbs are heterogeneous (polyclonal) and fall into two general classes both of which can be associated with autoimmune thyroid disorders – (a) thyroid stimulating autoantibodies (TSAb) that mimic that the actions of TSH and cause Graves’ hyperthyroidism and (b), blocking antibodies (TBAb) that block TSH binding to its receptor and can cause hypothyroidism [37,48,621,625,629,633,634]. Although TSH, TSAb and TBAb appear to bind to different sites on the TSH receptor ectoderm, TSAb and TBAb have similar affinities and often overlapping epitope specificities . In some cases of Graves’ hyperthyroidism, TBAb have been detected in association with TSAb [636,637] and the dominance of one over the other can change over time in response to treatment . Because both TSAb and TBAb can be present in the same patient, the relative concentrations and receptor binding characteristics of these two classes of TRAb may influence the severity of Graves’ hyperthyroidism and the response to antithyroid drug therapy or pregnancy [624,636,639-643]. For completeness, it should also be mentioned that a third class of “neutral” TRAb has also been described, of which the functional significance has yet to be determined [641,644].
Two different methodologic approaches have been used to quantify TSH receptor antibodies [40,620,633,645]: (i) TSH receptor tests (TRAb assays) also called TBII or TSH Binding Inhibition Immunoglobulin assays, and (ii) Bioassays that use whole cells transfected with human or chimeric TSH receptors that produce a biologic response (cAMP or bioreporter gene) when TSAb or TBAb are present in a serum specimen. In recent years automated immunometric assays using recombinant human TSHR constructs have been shown to have high sensitivity for reporting positive results in Graves’ disease sera [620,646]. However, assay sensitivity varies among current receptor versus bioassay methods 
Bioassay methods (TSAb/TBAb)
The first TSH receptor assays used surgical human thyroid specimens, mouse or guinea pig thyroid cells, or rat FRTL-5 cell lines to detect TSH receptor antibodies. These methods typically required pre-extraction of immunoglobulins from the serum specimen [626,633,647-652]. Later, TRAb bioassays used cells with endogenously expressed or stably transfected human TSH receptors and could use unextracted serum specimens [653-655]. Current TRAb bioassays are functional assays that use intact (typically CHO) cells transfected with human or chimeric TSH receptors, which when exposed to serum containing TSH receptor antibodies use cAMP or a reporter gene (luciferase) as a biological marker for any stimulating or blocking activity in a serum [40,42,620,648,651,653,656]. Bioassays are more technically demanding than the more commonly used receptor assays because they use viable cells. However, these functional assays can be modified to detect TBAb that may coexist with TSAb in the same sera and make interpretation difficult [40,657]. The most recent development is for 2nd generation assays to use a chimeric human/rat LH TSHR to effectively eliminate the influence of blocking antibodies. This new approach has shown excellent sensitivity and specificity for diagnosing Graves’ hyperthyroidism and clinical utility for monitoring the effects of anti-thyroid drug therapy .
TSH Receptor (TRAb)/TSH Binding Inhibitory Immunoglobulin (TBII) Methods
TRAb methods detect serum immunoglobulins that bind TSHR but do not functionally discriminate stimulating from blocking antibodies. TRAb methods are based on standard competitive or noncompetitive principles. First generation methods were liquid-based whereby immunoglobulins in the serum inhibited the binding of 125I-labeled TSH or enzyme-labeled TSH to a TSH receptor preparation [40,658]. These methods used TSH receptors of human, guinea-pig or porcine origin . After 1990, a second-generation of both isotopic and non-isotopic methods were developed that used and immobilized porcine or recombinant human TSH receptors [40,659-661]. These second-generation methods were shown to have significantly more sensitivity for detecting Graves’ thyroid stimulating immunoglobulins than first-generation tests . In 2003 a third-generation of non-isotopic methods were developed that were based on serum immunoglobulins competing for immobilized TSHR preparation (recombinant human or porcine TSHR) with a monoclonal antibody (M22) [37,40,42,620,648,656,660,662-666]. 3rd generation assays have also shown a good correlation and comparable overall diagnostic sensitivity with bioassay methods [620,636,648,667,669]. Current third-generation tests have now been automated on several immunoassay platforms . However, between-method variability remains high and interassay precision often suboptimal (CVs > 10 %) despite the use of the same international reference preparation for calibration [622,670]. This fact makes it difficult to compare values using different methods and indicates that further efforts focused on additional assay improvements are needed [37,622,671].
Clinical Use of TRAb Tests
ELISA and RIA are two different types of immunoassays used in research and diagnosis.
Their principle is similar but their methods of detection and analysis are different.
ELISA is a methods where in color is produced out of an immune reaction and the color is estimated for qualitative analysis and quantitative analysis.
While RIA is also an immune reaction, it involves presence of radiation after the reaction.
So the radiation is measured as a function of quality and quantity of the sample under test.
For more details read our detailed articles on ELISA principle and also RIA principle.
First we shall see the similarities.
ELISA and RIA similarities
|Principle||Immunoassay type||Immuno assay type and involves formation of antigen and antibody.|
|Factors involved||Antigen and antibody reaction||Antigen and antibody reaction by immune methods.|
|Uses and applications||For research and diagnosis||Also useful in diagnosis and research|
|Cost of experimentation||Expensive and one has to rely on ready made industry kits and elisa plate reader||Expensive and one may need to procure radio isotopes and also radio film|
ELISA VS RIA differences
|The measurement relies largely||On development of color.||Production of radiation.|
|The key molecule is||Enzyme linked to antigen or antibody.||The radio active isotope.|
|The types of reacting substance to produce detectable change||Many enzymes like horse raddish peroxidase||Mostly a single iodine isotope.|
|Time of experiment||Is very short and one can measure very fast.||It is very time taking|
|Skill requirement||Minimal skill and knowledge required.||An efficient and highly skilled handler is needed to minimize the risk of exposure to radiation.|
|Permissions||No special permissions required.||Permissions or license required for use of radio active material|
|Disposal||Disposal of waste is simple and can be done along with medical waste if used in hospital.||Disposal requires special care to avoid radiation exposure by other people.|
|Applications||Disease diagnosis due to rapid experimentation||Largely for receptor binding studies and also to small extent for diagnosis.|
Also Elisa protocol has different types like direct elisa, indirect elisa, sandwhich elisa. Even there are multiple bindings like antigen-antibody-antigen or antibody-antigen-antibody.
The detection of color generated due to elisa reaction is done by simple photometer (visible spectrophotometer) where as in RIA, the measurement of the radioactive emission is done by using a gamma counter.
Both have different roles in research and also diagnosis. But currently use of RIA is declining in Health care while the ELISA technique has gained wide popularity. Especially for infectious disease, serum diagnosis by ELISA is too frequent as it is cheap and very rapid.
On the other hand RIA has got limited use due to safety issues and legal binding rule. Yet it is an indispensable tool in medical research as it helps in study of drug receptor binding and other micro-sturctural imaging.