Validity of analog free thyroxin immunoassays.

I have briefly illustrated the way in which the main features of analog methods may be readily predicted by consideration of elementary physicochemical laws. (Editorial limitations on the length of this presentation have prevented exploration of other issues such as the reasons for, and effects of, the hitherto unexplained inclusion by manufacturers of large amounts of albumin (12, 18) in kit reagents.) The main implication of our analysis is that, to conform genuinely to the principles of "unbound analog" free hormone immunoassay, an analog must bind to serum proteins to a maximal extent of approximately 10% (but preferably less) in the absence of antibody. The notion that an analog is suitable for use in this context merely if it binds to serum proteins with lower affinities than does the native hormone (6) is demonstrably fallacious. Current analog methods thus neither adhere to the principles of unbound analog free hormone assay nor do they survive classic dilution tests of free hormone assay validity. Demonstrably they do not, in a general sense, measure the free hormone in serum. They may nevertheless yield roughly "correct" values in pregnancy (and thus, arguably, possess clinical value) if the analog fortuitously distributes in an optimal manner between serum proteins, implying approximate balance between the biasing effects caused by the protein changes occurring in pregnancy. However, such methods yield incorrect estimates in all other situations in which the binding characteristics of test samples are disturbed (e.g., by the presence of binding competitors, drugs, abnormal proteins, etc. or when "unbalanced" changes in protein concentration occur), this being the fundamental cause of their diagnostic unreliability. Wilkins, Midgley, and their colleagues have consistently opposed these conclusions, although they have never (using their computer model or by other rigorous means) demonstrated them to be incorrect, nor indeed have they ever formally substantiated the physicochemical propositions on which they themselves claim the methodology rests. Furthermore, they have repeatedly misinterpreted experimental results yielded by the Amerlex kits (such as the effects thereon of serum dilution, NEFA, drugs, etc.), discreetly abandoning (and even totally reversing) their original claims (though not withdrawing them) when these appeared no longer tenable. Perhaps the most serious consequence of these events is the resulting confusion in the literature. This is of major importance in physiological research (28), but it also clearly has major implications in clinical chemistry.(ABSTRACT TRUNCATED AT 400 WORDS)