Analytic measurements of free thyroxine.

This article has attempted to clarify two main areas of uncertainty relating to the clinical use of free TH determinations. The first centers on the validity of the free hormone hypothesis. This has been the subject of contentious debate for more than a decade, and final agreement among endocrinologists on this issue has yet to be reached. It is thus important that diagnosticians should not assume the validity of the hypothesis when assessing the analytic validity of free TH assay kits. In short, the reliability of the kits should be judged by analytic criteria (such as their theoretic basis and the results yielded on diluted samples) and not merely on the claim that they yield correct results in certain patient categories. Indeed, the finding that fT4 decreases in pregnancy (albeit to a limited extent) illustrates the danger of assuming the validity of the hypothesis when judging the reliability of fT4 assay kits. The second area of uncertainty relates to the true serum fT4 concentrations present in various clinical conditions, in consequence of doubts regarding the analytical validity of currently available methods. An attempt has been made in this presentation to clarify the physicochemical concepts underlying the more widely-used kits, and some of the reasons for their unreliability, although this also constitutes a contentious area. Errors generated in analogue-based free TH assay kits arise in part in consequence of the use of analogues that fail to conform to the physicochemical theory underlying "unbound analogue" assay, compounding errors resulting from the inclusion in assay reagents of various additives (e.g., albumin) that, inter alia, disturb the equilibrium between free and bound hormone moieties, and which also distort the effects of binding competitors. Regrettably different manufacturers have, in practice, relied on different analogues and/or reagent additives, buffers, blockers, and so forth, and different operating temperatures. Indeed different kits from the same manufacturer frequently differ in their formulation (see, for example, footnote on page 618), and yield widely differing results in certain sera. Also manufacturers may modify kit components from time to time without disclosure. Thus, though many analogue-based kits may arguably be diagnostically useful in well-defined clinical situations, total reliance cannot be placed in the results they yield in more complicated circumstances (such as those characterizing NTI), which may be diagnostically misleading.(ABSTRACT TRUNCATED AT 400 WORDS)