Total thyroxine testing: Comparison of an in-house test kit with radioimmuno- and chemiluminescent assays

Peterson ME, DeMarco CL, Sheldon KM.

in Scientific Proceedings (Abstract). American College of Veterinary Internal Medicine 2003;396.

Endocrine testing of dogs and cats is commonplace in clinical practice. Feline hyperthyroidism and canine hypothyroidism comprise the most frequently encountered endocrinopathies for each species, and measurement of serum total thyroxine (TT4) concentrations has become a valuable screening tool for the differential diagnosis of these disorders. The current study was designed to compare the serum TT4 concentrations measured using a validated in-clinic assay (SNAP T4t, IDEXX Laboratories Inc.) with the TT4 concentrations measured using 2 validated commercial laboratory technologies; radioimmunoassay (RIA) (Coat-a-Countt, Diagnostic Products Corp) and chemiluminescence (Immulitet, Diagnostic Products Corp.). Serum samples were collected from healthy dogs and cats as well as from animals with suspected thyroid disease. The 56 feline and 49 canine samples were divided into 3 aliquots and tested by each of 3 techniques. The results were analyzed for correlation and bias using standard statistical methods. The results obtained from the SNAP T4 test correlated well with both the RIA results (r 5 .91 and .90 for feline and canine samples respectively) and the Immulite results (r 5 .89 and .92 for feline and canine samples respectively). Bias analysis showed that both SNAP T4 and Immulite reported higher results than RIA for feline samples. The overall SNAP to RIA bias was 0.7 mg/dL The overall Immulite to RIA bias was 0.4 mg/dL. In all cases the bias between any 2 test methods for canine samples was small (0.2 mg/dL for SNAP: RIA and 20.1 mg/dL for Immulite : RIA). Using the laboratory tests as a benchmark, the overall accuracy of the Snap T4 test was excellent. Less than 5% of the total sample set gave Snap results that were discrepant from both the RIA and Immulite results by more than 1.0 mg/dL. Using the manufacturer’s reference ranges, only 5 samples from the total sample set would have given results that, based on the reference ranges, could have led to a different clinical decision depending on the test method used.Excellent concordance between Snap T4 results and either RIA or Immulite results was observed. All of these technologies can be used to accurately measure serum TT4 concentrations, but because of slight test-associated biases, the most consistent results will be obtained by using 1 test method exclusively.