Interferences With Thyroid Function Immunoassays: Clinical Implications and Detection Algorithm

The slow clearance, prolonged half-life, and high serum concentration of thyroxine (T4) are largely due to strong binding by the principal plasma thyroid hormone-binding proteins, thyroxine-binding globulin (TBG), transthyretin (TTR), and albumin. These proteins, which shield the hydrophobic thyroid hormones from their aqueous environment, buffer a stable free T4 concentration for cell uptake. Free rather than bound T4 is subject to homeostatic control by the hypothalamic-pituitary thyroid axis. Although it is not a protease inhibitor, sequence analysis identifies TBG as a member of the serine protease inhibitor (serpin) family of proteins. Proteolytic cleavage of TBG appears to be a mechanism for site-specific release of T4 independently of homeostatic control. TBG probably facilitates the transport of maternal T4 and iodide to the fetus, although this remains to be proven. High-affinity cellular binding sites for TTR have been described; however, their function and that of choroid plexus synthesis of TTR and transport of T4 into the cerebrospinal fluid remain unclear. Albumin, with the lowest T4 affinity and fastest T4 release of the major T4-binding proteins may promote quick exchange of T4 with tissue sites. The affinity of albumin for T4 is increased by histidine substitution for arginine 218 in the most common form of dysalbuminemic hyperthyroxinemia. However, proline and alanine substitutions at the same site have a similar effect, suggesting that arginine 218 interferes with T4 binding.

Heparin can cause an artifactual elevation in the concentration of unbound (free) thyroxine (T4) in the plasma, particularly when measured by equilibrium dialysis. The lipase released into the plasma by heparin acts on substrate (triglycerides; TG) in the plasma in vitro to release nonesterified (free) fatty acids (FFA), which, in high concentrations, inhibit the binding of T4 to its plasma binding proteins. This artifact occurs only in the presence of sufficient substrate (serum TG greater than approximately 180 mg/dL), and is most pronounced in methods requiring long incubation times. We observed this artifact in a patient receiving intralipid and subcutaneous (sc) heparin. Plasma-free T4, when measured by equilibrium dialysis, was elevated, but was normalized when the in vitro generation of FFA during equilibrium dialysis was prevented by prior treatment of the sample with protamine to inhibit lipoprotein lipase and with an antibody to hepatic triglyceride lipase. This observation caused us to investigate formally whether heparin, at standard sc doses or at iv doses even lower than those that are commonly used to flush iv lines (100-300 U), could also cause this artifact. We gave increasing doses of heparin at weekly intervals to each of three normal volunteers and measured FFA generation in their plasma (supplemented with 250 mg/dL triglycerides) under conditions simulating equilibrium dialysis. We found that, indeed, iv doses of heparin as low as 0.08 U/kg (5.6 U in a 70-kg subject) as well as a standard dose of sc heparin (5000 U) could release significant lipase activity into the plasma and, in the setting of sufficient substrate, cause enough in vitro generation of FFA to artifactually increase the serum-free T4 concentration when measured by equilibrium dialysis. These results indicate that equilibrium dialysis may not always be the best method for assessing serum-free T4 concentrations in hospitalized patients, and should be taken into account when interpreting previous studies demonstrating inhibitors of T4-serum protein binding in sera from hospitalized patients.

The case of a 39-year-old woman who was referred for weight gain and amenorrhoea is reported. Laboratory evaluation showed high levels of thyroid-stimulating hormone (TSH). The patient was started on increasing doses of levothyroxine for subclinical hypothyroidism. TSH remained persistently raised and the patient became thyrotoxic. Evaluation at another laboratory showed normal levels of TSH, raising the possibility of interfering substances. TSH levels were normalised with the addition of mouse serum to the patient's sample, confirming the presence of human anti-mouse antibodies as the interfering substance in the TSH assay.