| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Laboratoire Universitaire de Biophysique, Unité d’Analyses Endocriniennes, CNRS UPRES-A 2 Service de Médecine Interne et Nutrition, Hôpital de Hautepierre, 67098 Strasbourg Cedex, France;
2
Service d’Onco-Hématologie, Hôpital de Hautepierre, 67098 Strasbourg Cedex, France;
a address correspondence to this author at: Institut de Physique Biologique, Faculté de Médecine, F-67085 Strasbourg Cedex, France
Thyrotropin (TSH) has now been generally accepted as the first-line marker when screening for thyroid diseases. However, free thyroxine (FT4) measurements remain useful to confirm the diagnosis of thyroid diseases and to evaluate the thyroidal status when TSH measurements appear to disagree with the clinical picture, as may be the case in patients with nonthyroidal illnesses (1).
Because it represents only a minute fraction of the total serum T4, measuring FT4, i.e., T4 not bound to binding proteins, remains difficult. This is especially true for sera from patients with nonthyroidal illnesses that have an altered binding capacity (2)(3). A variety of in vitro methods have been developed to estimate FT4. The low-molecular weight radioactive-labeled analog method was much criticized (2) and was soon completely withdrawn from the market. The most widely used methods are now the two-step assay and the one-step labeled-antibody assay. The aim of our study was to compare FT4 concentrations in sera from patients with nonthyroidal illnesses, measured with three recent nonisotopic automated techniques and with three manual RIAs: two immunoextraction assays (one two-step and one one-step assay), and a direct equilibrium dialysis (ED) assay considered as the reference method (2).
The patient panel consisted of 20 hospitalized subjects on the seventh day following a bone marrow transplant, an example of severe acute illness. The bone marrow transplant was indicated for several hemopathies (n = 12) or solid cancers (n = 8). All patients were euthyroid before the treatment, and none had been known to have a patent thyroid dysfunction in the past. The preparation regimen before the transplant consisted of chemotherapy alone (n = 11) or chemotherapy associated with total-body radiotherapy (n = 9). After the transplant (allogenic in 9 cases and autogenic in 11 cases), patients received immunosuppressive drugs such as glucocorticosteroids (prednisone, 10–20 mg/day), cyclosporin, methotrexate, and calcium folinate. They also received granulocyte growth factors, antibiotics, antiviral and antifungal drugs, low-molecular weight heparinate (1000–10 000 units/day), and several comfort medications. All of the procedures we followed were in accordance with the Helsinki Declaration of 1975 and the subsequent 1996 amendments. The patient sera were collected for routine analysis, were kept frozen at -20 °C, and were analyzed shortly after thawing.
Two automated FT4 methods were labeled-antibody immunoassays: the electrochemiluminescent Elecsys FT4 assay from Roche Diagnostics performed on the 2010 analyzer; and the enhanced chemiluminescent Vitros ECI FT4 assay from Ortho-Clinical Diagnostics. The third automated assay was a non-isotope-labeled analog technique: ADVIA Centaur FrT4 from Bayer Diagnostics. The two-step RIA was the GammaCoat FT4 from DiaSorin, and the one-step RIA was the labeled antibody kit Amerlex-MAB FT4 from Ortho-Clinical Diagnostics. The manual direct ED technique was the Nichols FT4 method (Nichols Institute Diagnostics). In addition to the FT4 assays, TSH was measured using four third-generation automated immunometric assays: the Immulite TSH-3G from Diagnostic Products Corporation performed on the Immulite analyzer; the Elecsys TSH from Roche Diagnostics performed on the Elecsys 2010 analyzer; the Vitros ECI TSH from Ortho-Clinical Diagnostics; and the ADVIA Centaur TSH-3 from Bayer Diagnostics. Total T4 was measured with the Elecsys T4 assay from Roche Diagnostics performed on the 2010 analyzer.
The individual results of TSH and FT4
measurements are reported in Table 1
. The FT4 values have
been plotted as the percentage of the mean reference value,
determined in the laboratory from >100 euthyroid control subjects
(Fig. 1
). The TSH results obtained with the four methods were in close
agreement. When a low threshold of 0.15 mIU/L was used, only
one sample (patient 11) was intermethod discordant, with borderline
results between 0.12 and 0.17 mIU/L. All other TSH results obtained
with the four assays were concordant, yielding 8 values within the
reference interval and 11 decreased values. This finding
confirms the good reliability of recent TSH assays, which, as in our
previous study (4), classified the sera in an equivalent
way. However, the Vitros ECI TSH median was lower than the three others
(Wilcoxon paired-rank test, P <0.07), and as already
reported with the Immulite and AxSYM TSH assays (5), a
negative bias between the Vitros ECI TSH results and the Elecsys,
Immulite, and ADVIA Centaur TSH results was observed at low values
<0.2 mIU/L.
|
|
The FT4 value obtained with the ED technique was
within the reference interval in 11 sera and increased in 9. To study
the agreement between the FT4 results obtained
with the different assays and with the dialysis technique, the
percentages of concordance were calculated and the degree of
concordance was measured using Cohen’s kappa coefficient (
)
(6). For this purpose, FT4 results
were classified as low, normal, or high, using the reference limits of
each assay. As generally accepted, concordance was judged to be very
good for >0.80 and good for 0.81 > > 0.60
(7).
FT4 concordance with the dialysis method was good
and very good for the Vitros ECI FT4 and the
GammaCoat FT4, respectively (Table 1
). It was
poor and not statistically significant for the Elecsys
FT4, and there was no concordance at all (
<0) for the ADVIA Centaur and Amerlex-MAB FT4
assays.
The dialysis method represents a fundamentally reliable approach for FT4 measurements, although it causes a dilution (1:13 in the Nichols kit); the effect of this dilution on FT4 determinations is difficult to assess if binding competitors are present (2). The transplanted patients were heparin treated. Heparin is known to induce, in vivo, a release of lipoprotein lipase from the vascular endothelium. The lipolytic activity of this enzyme in vivo, and/or in vitro during sample storage and assay, leads to an increase in the serum concentration of nonesterified free fatty acids (NEFAs), which are binding competitors of T4, and consequently to an increase in FT4 (8). The production of NEFAs is negligible during storage at -20 °C, but it could be particularly marked during the overnight dialysis step at 37 °C. Like ED, the GammaCoat assay is based on a sound physicochemical basis, and it has been shown to give values in close agreement with those of ED in many states, such as high and low binding protein concentrations, autoantibodies, low total T4, and nonthyroidal illness (9). Against ED, the GammaCoat assay has the advantage of involving a much shorter incubation time at 37 °C (20 min instead of 16 h), which should limit the generation of NEFAs in vitro. Increased FT4 results were also frequent with the GammaCoat assay: eight results were above the upper reference limit. Therefore, increased ED FT4 values should not be considered an in vitro artifact resulting from prolonged dialysis at 37 °C, but they may be a reflection of in vivo hyperthyroxinemia.
Among the other four FT4 assays, only the Vitros ECI showed an acceptable concordance with ED results. This new labeled-antibody assay, like ED, has been shown to be free of biases that depend on the serum binding capacity and does not include albumin in the assay reagents to buffer the effects of NEFAs (10). The other two labeled-antibody assays, the Elecsys and Amerlex-MAB, include albumin in their reagents. Albumin can induce negative biases when the serum binding capacity is reduced, as it is in sera from patients with nonthyroidal illnesses (10). The ADVIA Centaur FT4 reagents do not contain albumin. The negative bias observed with this method might be related to the sensitivity of the labeled analog to variations in serum albumin concentrations. The ADVIA Centaur analog is a high-molecular weight analog, IgG-T4 complex labeled with acridinium ester. In general, macromolecular analogs are not considered albumin-dependent, and they do not have the marked drawbacks of low-molecular weight radioactive analogs (11). However, it has been shown that, contrary to the Vitros ECI, the Chiron ACS:180 FT4, a previous formulation of ADVIA Centaur reagents, yielded decreased FT4 values when the serum binding capacity was reduced by dilution (5)(12) and yielded only moderately increased FT4 values when a competitor such as furosemide was added to serum (13). Our results are consistent with the sensitivity of the ADVIA Centaur assay to the serum binding capacity. The dependence on the serum binding capacity of the ADVIA Centaur kit, the Amerlex-MAB kit, and to a lesser extent the Elecsys kit, may also explain the decreased FT4 measurements observed with these three kits: the number of FT4 results below the lower limit of the reference interval for each kit were three, four, and one, respectively.
TSH results were frequently decreased, and FT4
values obtained with the ED technique, GammaCoat, and Vitros ECI kits
were frequently increased. However, the concordance between the Elecsys
TSH results, chosen as example, and FT4
determined with these three kits was poor: the percentages of
concordance were between 60% and 65%; and the coefficient, lying
between 0.24 and 0.33, was not statistically significant (Table 1
). The
concordance was still lower with Elecsys FT4 and
absent ( <0) with the ADVIA Centaur and the Amerlex-MAB
FT4. The relationship between serum TSH and
FT4 has been well characterized in individuals
with normal hypothalamic and pituitary functions and stable
concentrations of thyroid hormones. The poor degree of agreement
between the FT4 and TSH values in our study might
be attributable in part to transient hyperthyroxinemia that did not
last long enough to induce a decrease in TSH to <0.15 mIU/L. Transient
hyperthyroxinemia may lead to high FT4
concentrations with TSH concentrations within the reference
interval. One reason for the discordant low TSH values in the
presence of FT4 values within the reference
interval may be related to the glucocorticoid treatment in these
patients, which decreases TSH secretion and lowers serum TSH
concentration without decreasing FT4 below the
lower limit of the reference interval (14). Finally, the
discordance between TSH values within the reference interval and
decreased FT4 values may be attributed to the
negative bias of the Amerlex-MAB, ADVIA Centaur, and Elecsys
FT4 assays. This bias might be related to
decreased serum binding capacity. The poor concordance we observed
between TSH and FT4 might also be related to the
impairment of the hypothalamic pituitary thyroid axis in a critical
situation such as bone marrow transplant.
In conclusion, despite a negative bias of the Vitros ECI TSH at concentrations <0.2 mIU/L, the concordance between all TSH methods was good in bone marrow-transplanted patients. The FT4 concentrations measured with recent automated methods were not equivalent. Among the three automated methods studied, which may gain widespread use, only the Vitros ECI showed a good concordance with the ED and GammaCoat kits, which frequently gave increased values and which are known to be technically reliable assays (2)(9). However, patients with nonthyroidal illnesses, such as those of this study, are not considered as hyperthyroid but as presenting with euthyroid sick syndrome (15). They do not benefit from treatment directed at normalizing FT4. FT4 results within the reference intervals were obtained in 85% of the patients with the Elecsys assay and in 80% of the patients with the ADVIA Centaur and Amerlex-MAB assays. If the "quality" of a FT4 method, in this situation, is judged on a "to treat or not to treat" basis, FT4 assay methods that more frequently produce values within the reference intervals, such as the Elecsys, ADVIA Centaur, and Amerlex-MAB, should be preferred. More than ever it is necessary to have a good knowledge of the FT4 method used to correctly interpret the results of thyroid function tests in patients with nonthyroidal illnesses.
Acknowledgments
This work was supported by the Hôpitaux Universitaires de Strasbourg. We thank Bayer Diagnostics and Ortho-Clinical Diagnostics for providing their TSH and FT4 reagents free of charge. We also thank N. Heider for carefully reviewing the English in the manuscript.
Footnotes
2>7004, Université Louis Pasteur, Faculté de Médecine, 67085 Strasbourg Cedex, France
fax 33-3-88-144861, e-mail sapin{at}ipb.u-strasbg.fr
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  G. van der Watt, D. Haarburger, and P. Berman Euthyroid Patient with Elevated Serum Free Thyroxine Clin. Chem., July 1, 2008; 54(7): 1239 - 1241. [Full Text] [PDF]
|
|
|
|
|
|
M. L. Rawlins and W. L. Roberts Performance Characteristics of Six Third-Generation Assays for Thyroid-Stimulating Hormone Clin. Chem., December 1, 2004; 50(12): 2338 - 2344. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. L. R. Williams, J. Simpson, C. Delahunty, S. A. Ogston, J. J. Bongers-Schokking, N. Murphy, H. van Toor, S.-Y. Wu, T. J. Visser, R. Hume, et al. Developmental Trends in Cord and Postpartum Serum Thyroid Hormones in Preterm Infants J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5314 - 5320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Sapin and M. d'Herbomez Free Thyroxine Measured by Equilibrium Dialysis and Nine Immunoassays in Sera with Various Serum Thyroxine-binding Capacities Clin. Chem., September 1, 2003; 49(9): 1531 - 1535. [Full Text] [PDF]
|
|
|
|
|
|
A. Price, C. Burgin, I. Catch, and M. Cruise Functional Sensitivity and Recovery of Thyroid-stimulating Hormone Clin. Chem., November 1, 2001; 47(11): 2067 - 2067. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
