Accuracy of free thyroxine measurements across natural ranges of thyroxine binding to serum proteins

Nonthyroidal illness syndrome: evaluation of thyroid function in sick patients

Altered thyroid function tests as a consequence of illness have been recognized for many years, yet the cause and clinical implications remains uncertain. The routine testing of thyroid function in hospitalized patients should be discouraged, as the results are less predictive of primary thyroid disease than in ambulatory patients. Clinicians should be aware of the methods used for thyroid function testing, as the effect of illness on thyroid function varies among the different tests. The most commonly used free T4 assays likely are influenced significantly by nonthyroidal illness. Advances in understanding the basic mechanisms of thyroid hormone metabolism and thyroid hormone action have given insights into the changes in thyroid function tests as a consequence of nonthyroidal illness. In the future, thyroid hormone receptor isoform-specific agonists and antagonists may allow for more specific treatment of select patients with nonthyroidal illness syndrome.

Reference ranges for thyroid function tests in premature infants beyond the first week of life

OBJECTIVES

To establish reference ranges for the more sensitive assays of thyrotropin and the best available assays of free thyroxine in premature infants after the first week of life.

STUDY DESIGN

Free thyroxine measurements by direct equilibrium dialysis and thyrotropin measurements by third generation immunometric assay were measured in 120 healthy premature infants 25 to 36 weeks' gestation at birth and every 3 weeks until hospital discharge. Infants were stratified by postconceptional age. Differences in free thyroxine and thyrotropin levels among groups were determined by ANOVA. Correlations between hormone measurements and gestational and postnatal ages were sought by linear regression analysis. Reference ranges were determined as arithmetic (free thyroxine) and geometric (thyrotropin) mean+/-2 SD ranges.

RESULTS

From 120 infants, 164 samples were obtained and grouped by postconceptional age at sampling. Free thyroxine was not different among postconceptional age groups and did not correlate with gestational or postnatal age. The free thyroxine reference range based on these data was 10 to 33 pmol/l (0.8 to 2.6 ng/dl). Thyrotropin did not correlate with gestational age. There was a clinically trivial but statistically significant (r(2)=0.03, p<0.05) correlation of thyrotropin with postnatal age. The thyrotropin reference based on these data was 0.8 to 12 mU/l.

CONCLUSIONS

Free thyroxine was closely regulated in these premature infants and levels were similar to those in older children and adults, once the natal surge in thyrotropin has subsided. After the first week of life a single range for each hormone appeared appropriate for all premature infants until 40 weeks postconceptional age.

Direct and Indirect Free Thyroxine Assay Methods: Theory and Practice

Background: For the diagnosis of thyroid disease, measurement of “free hormone” is generally accepted as an appropriate measure. However, valid assays measuring the free fraction of thyroxine (FT4) ideally must perform without bias, despite large variations in the concentrations and affinities of serum T4-binding proteins in the population. Several approaches have been taken to overcome such bias, and these have created considerable controversy in the field over the past decade.

Approach: This review, from both a historical and an analytical standpoint, charts the progress made over more than 30 years in improvements to the performance of assays in common use for the measurement of FT4 in serum or plasma. It reexamines the theory behind early approaches to such assays [for example, the free thyroxine index (FTI) method], that preceded more accurate, two-step immunoassays or one-step analog techniques. It evaluates the continuous refinements to the latter assays that by now have largely supplanted the FTI approach and where the deficiencies that so exercised clinical chemists in the past have been virtually eliminated in the leading assays.

Content: The basic Mass Action theory underpinning all such methods is discussed by assessing how far each particular approach obeys the criteria the theory imposes. In this, it is not the intention of the review to dissect individual commercial or academic assays, but rather to give guidance where appropriate as to how any assay said to measure FT4 can be conveniently evaluated by those intending to use it. Examples are given where inappropriate tests may wrongly imply assay invalidity by misinterpreting how FT4 assays work.

Summary: Detailed knowledge of the underlying theory is essential when devising tests for direct FT4 assays, to ensure that such tests do not overstep the practical limits of assay validity.

Serum thyroxine binding capacity-dependent bias in five free thyroxine immunoassays: assessment with serum dilution experiments and impact on diagnostic performance

OBJECTIVES

To assess the presence and magnitude of a serum thyroxine binding capacity (sBC)-dependent bias in five free thyroxine (FT(4)) immunoassays, compared with equilibrium dialysis (ED). The exhibited bias is confronted with clinical results from previous studies to evaluate its impact on FT(4) determination in sera with various sBC.

DESIGN AND METHODS

The sera of three pregnant women and three non thyroidal ill (NTI) patients were serially diluted in an inert buffer to progressively decrease the sBC. FT(4) values were measured in diluted and undiluted samples with the six assays.

RESULTS

As a function of increasing dilution performed on pregnancy sera, except for ED and Vitros ECi FT(4,) the other four FT(4) assay results decreased to different degrees, in the following order: two-step GammaCoat RIA< Elecsys< ADVIA:Centaur< Amerlex-MAB RIA. In sera from NTI patients, the decrease was more marked and found at high dilution with the Vitros ECi assay. Data from previous studies showed that FT(4) measured with biased assays were decreased only in samples from very severely NTI patients with low sBC and that FT(4) results in pregnancy sera with high sBC were not significantly biased.

CONCLUSIONS

The dilution test is a sensitive alarm to assess sBC-dependent bias in FT(4) assays. For all FT4 assays and particularly when a bias is observed, documentation should be sought on the diagnostic performance of the assay and supported by a detailed clinical study including samples with low sBC. Physicians should still be educated about the limitations of FT(4) assays.

Free thyroid hormone measurement. A critical appraisal

The main purpose of free T4 and free T3 assays is to distinguish reliably between thyrotoxicosis, hypothyroidism, and the euthyroid state, an objective that cannot be attained with assays of total T4 and T3 because of hereditary and acquired variations in the concentrations of binding proteins. Effective correction for changes in the serum concentration of TBG can be achieved with numerous types of free hormone estimate, but other changes in binding are not well accommodated. Despite remarkable methodologic ingenuity, no current method reflects the free T4 concentration in undiluted serum under in vivo conditions. Equilibrium dialysis, widely considered the reference method for free T4 measurement, is also subject to error, either preanalytic, owing to generation of NEFA in the sample leading to an overestimate of free T4, or analytic with underestimation of the effect of competitors to increase free T4. Current approaches to free T4 measurement are vulnerable to several method-dependent artifacts: abnormal albumin binding of T4 or of the assay tracer, the inhibition of T4 binding to TBG by medications, and the effects of critical illness, especially in heparin-treated patients, pregnancy, and the abnormalities in sick premature infants. Because of systematic variation between methods (i.e., whether a technique is albumin dependent or prone to incubation or dilution artifacts), it is essential to consider methodologic details in evaluating free T4 estimates in these situations and whenever estimates of free T4 are clinically discordant. False-positive abnormalities are more frequent than false-negative results. When free T4 results are correlated with the serum TSH concentration with attention to the assumptions that define this relationship, the majority of false-positive results can be readily identified. If a free T4 anomaly remains unexplained on repeat sampling, it is appropriate to use an alternative free T4 method that depends on a different assay principle and to correlate the result with an authentic total T4 measurement.