Another approach to the problem of thyroxine-binding protein alteration is to measure free T4 rather than total T4. The amount of protein-bound inactive T4 by itself has no direct influence on the serum level of metabolically active free hormone. The original Sterling technique involved separation of free from protein-bound T4 by a dialysis membrane after adding radioactive T4. The amount of free T4 was estimated indirectly by measuring total T4, obtaining the percentage of radioactivity in the dialysis fluid compared to total radioactivity added to the patient specimen measured before dialysis, and then calculating FT4 by multiplying the percentage of the dialysate radioactivity by total T4 quantity. This method generally gave normal results in patients with TBG abnormalities but frequently produced elevated results in patients with severe non thyroid illness. Several years later, Nelson and Tomei developed a modification of the dialysis method using a different dialysis solution buffer and measuring FT4 directly in the dialysis fluid using a more sensitive T4 immunoassay than was available to Sterling. Nelson’s results showed that most specimens were within reference range in both TBG abnormality and in severe non thyroid illness. Some investigators consider the Nelson equilibrium dialysis direct method to be the current FT4 gold standard. However, dialysis is time consuming, relatively expensive, and cannot be automated. Therefore, most laboratories use non dialysis immunoassay methods, which are commercially available based on several different principles but that are simple enough to be within the technical ability of most ordinary laboratories. The “two-step” FT4 is one such method; this involves tubes with anti-T4 antibody coating the tube walls. This antibody captures FT4 in patient serum but not T4 bound to serum proteins. The patient serum is then removed; the tube washed; and a solution containing T4 labeled with an isotope or an enzyme is added. The labeled T4 solution is removed after incubation. The amount of labeled T4 captured by the antibody on the tube surface is proportional to the amount of FT 4 in the patient sample (that is, how many antibody binding sites are occupied by patient FT4 and therefore not available to labeled T4). At present, most kit manufacturers use the “T4 analogue” method, because it is the easiest and least expensive. A synthetic molecule similar to T4 (T4 analogue) is created that will not bind to TBG but will compete with nonbound (free) T4 for anti-T4 antibody. This analogue is labeled with an isotope or enzyme system, so that the amount of analogue bound to the antibody is proportional to the amount of FT4 available. The analogue kits appear to function as well or slightly better than the FT4I in differentiating euthyroid persons from hyperthyroid and hypothyroid patients.

Drawbacks. Unfortunately, in patients with severe non thyroid illness most of the first-generation analogue kits were falsely decreased as often as the ordinary T4 methods and more often than the FT4I. Although the reasons for this have been disputed, the consensus indicates that the analogues bind to albumin to some degree and also are affected by nonesterified fatty acids. Albumin is often decreased in severe non thyroid illness. The manufacturers now attempt to “correct”their analogue kits in various ways, most often by adding a blocking agent that is supposed to prevent analogue binding to albumin. At present, most analogue kits are less affected by non thyroid illness than previously, but they still are affected, with a rate of false decrease about the same as the FT4I. However, not all FT4 kits perform equally well. In several multikit evaluations, one-step analog kits gave decreased values in severe non thyroid illness in about 40% of patients (range, 2%-75%) and increased values in about 1% (range 0%-9%). In several different dialysis and several two-step method kits, there were decreased values in about 20% of patients (range, 0%-81%) and increased values in about 12% (range, 0%-42%). There was considerable variation in results between different kits. Heparin increases free fatty acid concentration, which falsely decreases some of the FT4 kit results, particularly some analog methods; ordinary total T4 is not affected. Some two-step FT4 kits can be affected, producing mildly elevated results in some cases.