Back -- Table of Contents -- Forward

Section II
Pre-Analytic Factors

The important pre-analytic factors are:

• Physiologic Variables

• TSH/T4-T3 relationship
• Demographic/genetic factors
• Biologic variation
• Nutrition
• Age

• Pathologic Variables

• Thyroid gland dysfunction
• Hepatic or renal dysfunction
• Medications
• Systemic illnesses

• Specimen-related Variables

• Interfering factors

The negative feedback inhibition of thyroid hormones on TSH secretion by the pituitary thyrotroph mediates the physiologic control of thyroid function. At the level of the thyrotroph cell, it appears that this negative feedback is mediated by T3 generated in situ from free T4 (FT4) entering the cell. An understanding of the normal relationship between serum levels of FT4 and TSH is essential when interpreting thyroid tests, since T4 functions as the major circulating pro-hormone for its conversion product T3, which exerts biologic activity in the cell. Needless to say, an intact hypothalamic-pituitary axis is a prerequisite for diagnostically accurate TSH measurements!

In addition to physiologic variables, other pre-analytic factors that can affect the results of biochemical thyroid tests include: (a) individual patient variables, including severe nonthyroidal illness (NTI); (b) iatrogenic factors, such as medications, including thyroid medications and steroids; and (c) specimen variables, including interfering factors such as heterophilic antibodies (HAMA).

Fortunately, most pre-analytic variables have little effect on the most common thyroid test used in ambulatory patients -- serum TSH measurement. Pre-analytic variables and interfering substances present in specimens can affect protein-bound thyroid hormones, and thereby total and free thyroid hormone tests, more frequently than serum TSH. Both FT4 and TSH values may be diagnostically misleading in the hospital setting of severe nonthyroidal illness (NTI). Specifically, euthyroid patients frequently have abnormal serum TSH and/or total and free thyroid hormone concentrations as a result of NTI per se or secondary to medications. When there is a strong suspicion that one of these variables might affect test results, specialist advice from the expert physician or clinical biochemist is needed.

Pre-analytic issues involving the measurement of serum thyroid autoantibodies and serum thyroglobulin (Tg) are discussed in Sections III D and E, respectively.

A. Physiologic Variables

(a) The serum TSH/ FT4 relationship

When hypothalamic-pituitary function is normal, the negative feedback inhibition of thyroid hormones on pituitary TSH secretion produces an inverse relationship between serum FT4 and TSH concentrations. Thyroid function can be determined either directly, by measuring the primary thyroid gland product T4 (as FT4), or indirectly, by assessing the biologic effect of the thyroid hormones on pituitary TSH secretion. It follows that if the pituitary is functioning normally, serum TSH concentrations will be inversely related to thyroid function (high TSH in primary hypothyroidism/low TSH in hyperthyroidism). In fact, as the sensitivity and specificity of serum TSH assays have improved, it has become apparent that for most clinical situations, serum TSH measurement offers more sensitivity for detecting thyroid dysfunction than does a FT4 test. The reasons for this are two-fold:

• As shown in Figure 1, serum TSH and FT4 concentrations have an inverse log/linear relationship such that small abnormalities in FT4 will produce an amplified response in serum TSH (7).
• Twin studies show that individuals have a genetically determined FT4 setpoint (8). This dictates that any mild FT4 excess or deficiency will be sensed by the pituitary, relative to that individual's setpoint, and cause an amplified, inverse response in TSH secretion. Thus, in the early stages of developing thyroid dysfunction, a serum TSH abnormality will precede the development of an abnormal FT4 as a reflection of the exponential TSH response to FT4 changes and the fact that FT4 reference limits are constructed from a cohort of individuals with different FT4 setpoints.

It is currently believed that measurement of the serum TSH concentration is our most accessible indicator of thyroid status at the tissue level. For example, studies of mild (subclinical) thyroid hormone excess or deficiency (abnormal TSH/normal range FT4 and FT3) find changes in markers of thyroid hormone action on a variety of tissues (heart, brain, bone, liver and kidney) that reverse when treatment is initiated to normalize serum TSH (9).

It is important to recognize the situations when serum TSH is diagnostically misleading:

• Abnormal hypothalamic-pituitary function.
• Unstable thyroid status, i.e. the early phase of treating hypothyroidism or adjusting the L-T4 replacement dose (6 to 8 weeks), as well as following treatment for hyperthyroidism when it takes time (2 to 3 months) for the serum TSH to re-equilibrate to the new thyroid hormone status.
• TSH-producing pituitary tumors.
• Drugs that influence pituitary TSH secretion (ie. steroids and glucocorticoids).

Figure 1. The TSH/Free T4 Relationship

(b) Demographic/Individual Factors

For practical purposes, the patient variables listed below have no significant effect on the reference ranges for thyroid tests. The exceptions are age (neonatal period, early childhood and the elderly) and caloric deprivation that is associated with low serum total and free FT3 concentration, as measured by most methods.

• gender
• race
• season
• phase of menstrual cycle
• cigarette smoking
• exercise
• fasting
• phlebotomy–induced stasis

Pregnancy raises the serum level of TBG that peaks at 12 to 14 weeks of gestation, with a parallel increase in serum TT4 and TT3 levels. There may be a slight rise in the maternal serum FT4 concentration in the first trimester with a corresponding mild decrease in the serum TSH level, but usually within normal reference limits (10). Hyperemesis amplifies these changes and may result in high FT4 and low TSH concentrations in the first trimester of pregnancy (11). As the pregnancy progresses, serum TSH levels remain normal and FT4 declines slightly, but still remains within normal reference limits (10, 12). Patients receiving L-T4 replacement therapy who become pregnant may require an increased dose to maintain normal serum TSH (13). Serum Tg concentrations typically rise during pregnancy in both normal subjects and patients with differentiated thyroid carcinomas (DTC) and return to baseline 6 to 8 weeks postpartum (10).

(c) Biologic Variation

The intra-individual serum levels of the thyroid hormones as well as their precursor protein, thyroglobulin (Tg) are quite stable over a 1 to 4 year period. Intra-individual variability is less than the variance between individuals (Table 1). The stability of intra-individual serum T4 concentrations reflects the long thyroxine half-life (7 days) and the genetic free T4 setpoint. The stability of intra-individual T3 concentrations reflects autoregulation of the rate of T4 to T3 conversion (14). All thyroid analytes display a greater inter-individual variability compared to intra-individual variability. This is especially the case for serum Tg concentrations because individuals have differences in thyroid mass, TSH status and may have conditions associated with thyroid injury (i.e. thyroiditis) – all conditions that influence serum Tg concentrations. Serum TSH concentrations display the most variability, both within and between individuals. This primarily reflects the short half-life of TSH (~60 minutes) together with its ultradian and diurnal variation, the latter peaking during the night and reaching a nadir sometime between 1000 to 1600 hrs (15). Since diurnal TSH fluctuations usually occur within the normal TSH reference range (~0.3 to 4.0 mU/L), diurnal variation rarely affects the diagnosis of thyroid disease and should not influence the scheduling of blood drawing in an outpatient setting.

B. Pathologic Variables

(a) Disease-related Variables

Medications

Medications can cause both in vivo and in vitro effects on thyroid tests. This may cause misinterpretation of laboratory data and lead to inappropriate diagnoses as well as unnecessary further testing, and escalating cost (19).

In Vivo effects: Overall, medications affect the TSH level less than thyroid hormone levels. For example, the well-known effect of estrogen on TBG concentrations which raises the serum TT4 level does not affect the serum TSH concentration. This is because TSH is controlled by the FT4 concentration that is independent of binding-protein effects.

Common medications of importance include:

• glucocorticoids
• propranalol
• dopamine
• furosemide
• phenytoin
• carbamazepine
• iodide/iodine
• amiodarone
• lithium

Glucocorticoids in large amounts can lower the serum T3 level and inhibit TSH secretion. Dopamine also inhibits TSH secretion and may even reduce the raised TSH level of primary hypothyroidism to normal in hospitalized patients. Phenytoin and Carbamazepine can result in a reduction of serum levels of TT4 and FT4; the effect on the FT4, however, is an artifact of commercial assay designs since it does not occur when a direct absolute FT4 method is used. These drugs do not affect the serum TSH concentration. Propranolol is sometimes used to treat manifestations of thyrotoxicosis and has an inhibitory effect on T4 to T3 conversion. Propranol given to non-thyroid disease individuals for their hypertension can cause an elevation in TSH as a result of the impaired T4 to T4 conversion. Iodide, such as iodine-containing solutions for sterilizing the skin, radioopaque dyes, Amiodarone, or oral iodide, can lead to either hyperthyroidism or hypothyroidism in those individuals who are predisposed (20, 21). Amiodarone, in addition has complex effects on thyroid gland function and may induce either hypothyroidism or hyperthyroidism in susceptible patients with thyroid autoantibodies (22). When Amiodarone causes hyperthyroidism, measurement of the serum T3 level may be needed for the diagnosis. When patients who take oral L-T4 are given amiodarone, the serum TSH concentration may be disproportionately elevated for the level of serum FT4. Lithium can cause hypothyroidism in as many as 5-10% treated patients, especially those with thyroid autoantibodies. In some cases it can cause hyperthyroidism.

In Vitro effects: Some therapeutic and diagnostic agents may be present in the serum specimen in sufficient concentrations to produce in vitro interference with some thyroid test methods. The therapeutic agents themselves (i.e. Furosemide) may competitively inhibit thyroid hormone binding to serum proteins in the specimen and cause an artefactual low result (23, 24). Alternatively, in the case of Heparin, in-vitro stimulation of lipoprotein lipase can liberate FFA, which inhibits T4 binding to serum proteins (19). In certain pathologic conditions such as uremia, abnormal serum constituents such as indole acetic acid may accumulate and interfere with thyroid hormone binding (25). Thyroid test methods employing fluorescent signals may be sensitive to the presence of fluorophor-related therapeutic or diagnostic agents in the specimen.

(b) Severe Nonthyroidal Illness (NTI)

Patients who are seriously ill often have abnormalities in their thyroid tests but usually do not have thyroid dysfunction. The terms "nonthyroidal illness" or NTI, as well as "euthyroid sick" are often used to describe this subset of patients (26). The spectrum of changes in thyroid tests relates both to the severity and stage of illness and to some extent the method used (see Figure 2).

Most sick patients have low serum TT3 and FT3 concentrations, as measured by most methods (6). In severe conditions, such as septicemia, there is inhibition of binding of thyroid hormones to serum proteins and both serum TT4 and TT3 concentrations may become subnormal (26). In contrast, serum free hormone values (FT4 and FT3) are method dependent. Typically, the FT4 and FT3 estimate tests used by clinical laboratories report low values for such specimens, whereas normal or elevated values are seen more often when the measurements are made by a direct absolute method like equilibrium dialysis [see Section IIIBc] (27, 28). Serum TSH concentrations remain within normal limits in the majority of NTI patients, provided that no dopamine or glucocorticoid therapy is administered (29). However, in acute NTI there is often a transient fall in serum TSH to the 0.02-0.2 mU/L range followed by a rebound to mildly elevated values during recovery (29). It is important to use a TSH assay with a functional sensitivity below 0.02 mU/L in the hospital setting. This is because a reliable measurement of the degree of TSH suppression can help discriminate sick hyperthyroids, who typically have profoundly low serum TSH (< 0.02 mU/L), from patients with a minor transient TSH suppression due to NTI. Mildly elevated TSH is also less diagnostic for hypothyroidism in the hospital setting. Sick hypothyroid patients are best identified from the combination of low FT4 and elevated TSH (>20 mU/L) (30).

It is clear that the diagnosis and treatment of thyroid dysfunction in the presence of a severe NTI is not simple, and is best done with the help of an endocrine specialist. It is recommended that thyroid testing of NTI patients include both the serum FT4 and TSH measurement, using a TSH assay with functional sensitivity below 0.02 mU/L. In general, it is best to avoid thyroid testing acutely ill patients unless there is a family history of thyroid disease or clinical symptoms of significant thyroid dysfunction likely to affect patient outcome. Empiric treatment of low levels of serum TT4 has not improved outcome and is still experimental (31).

 

Figure 2. Changes in Thyroid Test Values in Nonthyroidal Illness (NTD)

Recommendations

For Physicians and Laboratories:

The assessment of thyroid function in hospitalized patients with acute or chronic NTI should be deferred until the illness has resolved, except when the patient's history or clinical features strongly suggest the presence of thyroid dysfunction that would adversely affect patient outcome. A combination of both FT4 and TSH measurement by an assay with functional sensitivity <0.02 mU/L, is recommended for evaluating the thyroid status of NTI patients.

 

C. Specimen Variables

(a) Stability

A recent study found no effect of exposure for 10 days at ambient temperature on total and free thyroid hormone, TSH or Tg concentrations. Other studies have reported that T4 in serum is stable for months when stored at 4ºC and years when frozen at –10ºC. TSH is stable for several years in frozen sera. TSH and TT4 in dried whole blood spots used in screening for neonatal hypothyroidism are also stable for several months when exposed to a desiccant. In general, TSH in serum is slightly more stable than T4.

(b) Serum Constituents

Hemolysis, lipemia, and hyperbilirubinemia do not produce significant interference in immunoassays, in general. However, free fatty acids can displace T4 from serum binding proteins, which may partly explain the low TT4 values often seen in NTI.

(c) Heterophilic Antibodies (HAMA)

These antibodies are encountered with moderate frequency in patient sera. They affect immunometric assay (IMA) methods more than competitive immunoassays by forming a bridge between the capture and signal antibodies, thereby creating a false signal, resulting in a high or low value artifact depending on the methodology see [Section IIIBe)] (32). The inappropriate value may not necessarily be abnormal, but merely be inappropriately normal. Since antibodies cross the placenta, they have the potential to influence the neonatal screening result. Manufacturers employ various approaches to neutralize HAMA effects on their methods with varying success.

(d) Sample Collection and Processing

Serum is preferred over EDTA- or heparinized-plasma by most manufacturers. For optimal results and maximum serum yield, it is recommended that whole blood samples be allowed to clot for at least 1 hour before centrifugation and separation. Storage of serum at –20ºC is recommended if the assay is to be delayed for more than 24 hours. Collection of serum in barrier gel tubes does not affect the results of most TSH and thyroid hormone tests.

Back -- Table of Contents -- Forward