MISDIAGNOSIS IN DISCORDANT FREE T4 AND TSH CONCENTRATIONS

DETECTING ASSAY INTERFERENCE BY METHOD COMPARISON

Abstract

INTRODUCTION
Thyroid function tests (TFTs) are important tools in diagnosing thyroid disorders. In rare cases, TFTs can be misleading due to assay interference which may result in false diagnosis. We describe a patient who presented with neck swelling and falsely elevated free thyroxine (FT4) due to assay interference and discuss a useful strategy to demonstrate assay interference.

CASE
A 40-year-old lady with thyrotoxicosis was previously treated with carbimazole until 2017. Her care was subsequently transferred to another centre. Here, she was found to have elevated FT4 (52.5 pmol/L) and nonsuppressed TSH (0.95 mIU/L) using Siemens® platform. She was started on carbimazole 5 mg OD. After a month of treatment, the tests revealed elevated FT4 [127.9 pmol/L, normal value (NV) 11.5 to 22.7] and elevated TSH [34.76 mIU/L, NV 0.55 to 4.78). Carbimazole was increased to 20 mg OD. Her serial FT4 and TSH levels were persistently elevated despite her being euthyroid. In view of discordant biochemical and clinical presentation, thyroid hormone assay interference was suspected. Her TFTs were repeated in 3 different platforms, which revealed low FT4 and elevated TSH concentrations suggestive of severe hypothyroidism. Assay interferences are usually due to interfering antibodies. By methods of comparison using 3 different two-step immunoassays (Abbott Architect, Beckman DxI 800 and Cobas Roche), all showed low FT4 concentrations and above reference range TSH levels, suggestive of hypothyroidism. Only the Siemens® platform showed a falsely elevated fT4, demonstrating an assay interference with this immunoassay.

CONCLUSION
Assay interference should be considered when there is a discrepancy between clinical picture and biochemical results to prevent inappropriate management. Method of comparison between immunoassays is a useful strategy to demonstrate the presence of autoantibodies as the source of assay interference.