Spurious conclusions on analog free thyroxine assay performance.
I read with misgivings the recent communication by Fritz et al. (1) alleging that because under some circumstances an analog free thyroxine ([FT.sub.4]) immunoassay correlates total [T.sub.4] and [FT.sub.4] values, it does not measure [FT.sub.4] but something akin to [T.sub.4]. I have shown repeatedly through mass action analyses (2-4) that many experiments, including the one reported by Fritz et al. (1) and others described in earlier papers (5, 6), cannot meaningfully address the working of any of these assays. Allegations of shortcomings cannot be substantiated if they depend on experiments using artificial [T.sub.4] solutions in which serum [T.sub.4] binding proteins are either lacking or are present at concentrations insufficient to prevent overlarge [T.sub.4] abstraction by the assay antibody (conditions strictly invalid for [FT.sub.4] measurements).
The amount of antibody that can be used in analog [FT.sub.4] assays having a radioactive probe is limited (3). Constraints on the stability of molecules radiolabeled with carrier-free "I atoms dictate the minimum usable amount of labeled probe for adequate tracer detection in the assay (3). To enable measurement, the minimum amount of stripping antibody must remove at least 1% of the total [T.sub.4] from a normal serum sample (2-4). Although this amount far exceeds the actual number of [FT.sub.4] molecules present in serum (approximately 0.02% of the total), the procedure remains valid because the serum-bound free equilibrium for [FT.sub.4], bound [T.sub.4] is negligibly disturbed, and the measured [FT.sub.4] concentrations (obtained through calibration of standards) are not significantly affected (3-5). Sequestrations of more than approximately 3% invalidate any results. Because the [FT.sub.4] concentration in ultrafiltrates or dialysates remains equivalent to only 0.02% of the total [T.sub.4] (i.e., the ambient, undiluted [FT.sub.4] concentration), any assay designed to sample 1% of [T.sub.4] cannot detect this small amount, and total [T.sub.4] assays similarly fail.
Results from varying serum (and therefore serum [T.sub.4]-binding protein) concentrations can also be explained quite differently. I understand that this early analog immunoassay retained a "TBG effect". That is, residual binding of the labeled analog to [T.sub.4]-binding globulin (TBG; and probably other proteins) was large enough to correlate [FT.sub.4] values with TBG (partly but not completely mitigated by a [T.sub.4] blocking agent). Thus, when serum amounts are altered by dilution or concentration, 2 connected events ensue. First, a more concentrated serum will have a higher concentration of TBG relative to antibody probe and act as a high-TBG serum, resulting in increased [FT.sub.4]. Conversely a diluted sample will mimic a low-TBG serum giving a low result. Second, intrinsic effects of dilution will occur. If in undiluted serum the assay samples 1%-2% of the total [T.sub.4], in a 4-fold diluted serum, it will sample 4%-8%, an interval that lies well outside the window of validity of the assay, insofar as too much [T.sub.4] is sequestered by the antibody probe to maintain equilibrium between [FT.sub.4] and the [T.sub.4] bound to the serum proteins. The assay essentially "collapses" into one more closely related to [T.sub.4], and will do so in all cases in which too much [T.sub.4] is sampled by the antibody.
Much of the work of Nelson et al. (5, 6) aimed at probing the working of analog [FT.sub.4] assays is not applicable to the topic under discussion, but these investigators seem to suggest that in regard to these [FT.sub.4] assays any experiment should be meaningful in any situation (whether physiologic or not). The cardinal property of direct 1-step assays (analog or labeled antibody) is that, although they are designed for all physiologic circumstances (i.e., they work with all sera regardless of serum [T.sub.4]-binding protein concentrations), they cannot be used with artificial solutions outside the assay's limits of validity. This point has been argued elsewhere (2-4).
Another questionable implication (5, 6) is that analyses of the performance of this particular assay may apply generally to all such assays. On the contrary, [FT.sub.4] assays are individual with regard to serum protein effects, assay ingredients, robustness to dilution, and percentage sampling by the antibody probe. General conclusions on the class of tests cannot be drawn by examining a single assay (3). Assays with residual analog binding to serum proteins cannot be easily compared with newer examples that lack this characteristic.
Finally, although ultrafiltration, equilibrium dialysis, and direct analog assays have the same aims, in regard to sampling and dilution these different assays represent opposite ends of a spectrum of valid methods. Experiments performed by one group of methods do not always apply to the other, because ultrafiltration and dialysis measure [FT.sub.4] in the absence of serum proteins and bound [T.sub.4], whereas analog assays accommodate their presence.
Editor's Note: Dr. Fritz declined to reply for publication.
(1.) Fritz KS, Wilcox RB, Nelson JC. A direct free [T.sub.4] immunoassay with the characteristics of a total [T.sub.4] immunoassay. Clin Chem 2007;53:91115.
(2.) Wilkins TA, Midgley JEM, Barron N. Comprehensive study of a thyroxin-analog based assay for free thyroxin ("Amerlex [FT.sub.4]"). Clin Chem 1985; 31:1644-53.
(3.) Midgley JEM. Direct and indirect free thyroxine assay methods: theory and practice [Review]. Clin Chem 2001;47:1353-63.
(4.) Midgley JEM. The design of one-step free thyroxine assays: a critique [Review]. J Clin Ligand Assay 2005;28:110-15.
(5.) Nelson JC, Wilcox RB. Protein-bound [T.sub.4] dependence; the uncontrollable variable in free [T.sub.4] assays. Exp Clin Endocrinol 1994;102:102-9.
(6.) Nelson, JC, Wang R, Asher DT, Wilcox RB. The nature of analogue-based free thyroxine estimates. Thyroid 2004;14:1030-6.
John E.M. Midgley
North Lakes Clinical
6 High Wheatley
Ilkley LS29 SRX,
|Printer friendly Cite/link Email Feedback|
|Author:||Midgley, John E.M.|
|Article Type:||Letter to the editor|
|Date:||Sep 1, 2007|
|Previous Article:||HPLC method for plasma vitamin [K.sub.1]: effect of plasma triglyceride and acute-phase response on circulating concentrations.|
|Next Article:||Presence of the hemochromatosis S65C mutation leads to failure of amplification in a multiplex C282Y/H63D PCR.|