Underestimates and overestimates of total thyroxine concentrations caused by unwanted thyroxine-binding protein effects.

The first evidence of unwanted serum protein effects on analogue-based total thyroxine (T4) determinations came from a study that varied serum protein concentrations while total T4 concentrations were constant. The present study approached this issue by varying total T4 concentrations while protein concentrations were constant. Four analogue-based total T4 immunoassays were applied to solutions that contained either free T4 without binding protein, a T4-binding protein without T4, or protein-bound T4. When total T4 concentrations were 3-12 microg/dL, the assays reported total T4 determinations that ranged from none detected to 23 microg/dL. These T4 determinations reflected the protein to which T4 was bound, in addition to the level of T4. Total T4 was underrepresented when T4 was unbound, or thyroxine-binding globulin (TBG) bound. Total T4 was overrepresented when T4 was albumin-bound, or transthyretin-bound. There were substantial disparities among assays applied to the same total T4 solutions. These assays reported no detectable T4 when applied to T4-binding protein solutions without protein-bound T4. Nonetheless, T4-binding proteins contributed to the underestimates and overestimates of protein bound T4. Different forms of protein bound T4 were quantified differently, evidence that protein-T4 complexes persist during quantification. We attribute the unexpected total T4 values to a combination of incomplete protein-bound T4 release from T4-binding proteins during quantification, and variably inaccurate quantifications of the protein-bound T4 that remained.

The nature of analogue-based free thyroxine estimates.

Clinical laboratories often use analogue-based immunoassays to estimate serum free thyroxine (FT(4)) concentrations. These assays yield FT(4) estimates that correlate closely with thyroxine (T(4)) binding protein concentrations. This correlation implies that either T(4) binding proteins or protein bound T(4) contribute to analogue-based FT(4) values. To study the contributions made by T(4) binding proteins to these FT(4) estimates further, four analogue-based FT(4) assays were applied to: (1) FT(4) solutions without T(4) binding proteins, (2) to T(4) binding protein solutions without T(4), and (3) to total T(4) solutions containing T(4) binding protein, FT(4), and protein-bound T(4). The FT(4) estimates obtained with these solutions ranged from 0.2-8.6 ng/dL, when FT(4) concentrations ranged from less than 0.2-12,000 ng/dL. In the FT(4) solutions, gravimetrically determined FT(4) concentrations were 500-12,000 ng/dL (0.5-12.0 microg/dL) without protein-bound T(4), and the FT(4) estimates obtained were 0.3-6.9 ng/dL. In the total T(4) solutions, dialyzable FT(4) concentrations were less than 0.2-59 ng/dL, retained T(4) concentrations were 499.8-11,441 ng/dL, and the analogue-based FT(4) estimates obtained were 0.2-8.6 ng/dL. Similar FT(4) estimates (0.2-8.6 ng/dL and 0.3-6.9 ng/dL) were obtained with similar concentrations of either protein-bound T(4) or FT(4). Similar test results were associated with similar total T(4) concentrations, not similar FT(4) concentrations. Protein-bound T(4) and T(4) binding protein contributed variably to test results. T(4) quantifications included large analytical losses that are unaccounted for. These assays passed tests of correlation with FT(4) concentrations, but they failed tests of specificity for FT(4) and accuracy in T(4) quantification.