False-Positive Total T3 Using the Ortho Vitros Immunoassay in Pediatric Populations

TO THE EDITOR: Total triiodothyronine (T3) (TT3) measurements can provide clinical utility elucidating T3-dominant Graves' disease and other thyrotoxicosis conditions. Thyroid hormone assays, including TT3, are susceptible to protein-based interferences, usually caused by endogenous antibodies. Interferences should be considered when the thyroid hormone concentrations are inconsistent with clinical impression (1). We detected an interference with the Ortho® Vitros TT3 assay causing falsely increased concentrations in 20 pediatric patients, some of whom had a history of high TT3 for years, suggesting that the interference was not limited to a single reagent lot. The erroneous results were questioned by endocrinology physicians due to incongruence of laboratory values with clinical features. Parallel testing on an alternative assay (Roche e601 T3) suggested the Ortho Vitros results were falsely increased (Table 1). A set of samples unaffected by interference (n = 25) were used for correlation between the Roche and Ortho assays and demonstrated the following regression statistics: Roche T3 = 1.11 (Vitros T3) + 0.19,R2 = 0.95. Communicationwith a second children's hospital that also relies on the Ortho Vitros TT3 immunoassay identified 7 patients with suspected falsely increased total T3 levels thatwerealsoevaluatedusingan alternative assay (Roche e602 or E170) (Table 1). The mean positive bias of the Vitros assay for TT3 in the combined cohort of 20 specimens was 3.15 ng/mL (SD 1.64 ng/mL). Minimal bias was observed in 2 pediatric hyperthyroid control samples (Table 1). In the cohort of patients with falsely increased total T3, other thyroid markers [thyroid-stimulating hormone (TSH), free thyroxine (T4), and free T3] were unaffected. Further, there was no discernable association between detectable antithyroglobulin or antithyroid peroxidase and the observed interference. Historic studies have reported similar false elevations using RIAs and defined an endogenous antibody-based mechanism (2). To test this hypothesis, we measured TT3 in 9 patient samples and 2 hyperthyroid controls before and after protein precipitation with ethanol, which we suspected would remove interference based on a publication by Després and Grant (2). Ethanol (800 μL) was added to 200 μL patient serum and incubated at room temperature for 60 min. The precipitate was removed by centrifugation at 18407g for 15 min. The supernatant was evaporated under nitrogen at 40 °C, and the resulting precipitate was reconstituted using 200 μL Vitros TT3 calibrator 1 (2). Ethanol precipitation significantly decreased the bias between assays (−0.06 ng/mL bias; SD 0.14 ng/mL) while the control specimens were unaffected (Table 1). Further support for a high– molecular weight (likely protein) interference was supported by incubation of a specimen suspected to be falsely increased with a heterophile blocking reagent (Scantibodies). This reagent is composed of a mixture of immunoglobulins specifically chosen to target heterophilic antibodies. Compared to the control sample, which showed little change on incubation, the specimen in question showed a 33% decrease in immunoreactivity. All experiments suggest the presence of a high–molecular weight interference, which we hypothesize to be endogenous antibodies that may be specific to T3 (3). Regardless of the