ABSTRACT
BACKGROUND: Total thyroxine assays continue to be an integral part of thyroid status testing. We experienced a significant number of elevated total thyroxine values in patients with normal thyroid stimulating hormone and thyroxine binding globulin concentrations using the Beckman Access and hypothesized that these were due to a sample matrix effect. METHODS: We compared the total thyroxine assays on the Beckman Synchron, Beckman Access, and Dade Dimension using individual patient specimens and two uniform matrices: a serum pool and an albumin-based matrix that were both supplemented with L-thyroxine to span the linear assay range. RESULTS: Access total thyroxine values in individual patient specimens exhibited sporadic positive bias as high as 77 nmol/l (6 microg/dl) when compared to the Synchron and Dimension. The use of uniform matrices had little effect on the Synchron in comparison to the Dimension but significantly improved the agreement between the Access and the Dimension or Synchron as indicated by a statistically significant improvement in correlation coefficients. CONCLUSIONS: The Access total thyroxine assay is prone to a variable and clinically significant positive bias that is mediated by a component of the sample matrix.
