Repeatability and variability of total T4 measurements at three German veterinary laboratories

Abstract

To evaluate the reproducibility of serum testing for total thyroxine (T4) in three German laboratories. Serum was taken from 53 dogs with suspected hypothyroidism and divided into five aliquots. Three aliquots of each sample were marked with different names and sent to one laboratory (two aliquots simultaneously and one aliquot 1-3days later). The remaining aliquots were sent to two other laboratories. Laboratory1 used an enzyme immunoassay for T4 measurements, whereas laboratories2 and 3 used a chemiluminescence immunoassay. The agreement between the three laboratories (values within or below the reference interval) was determined using the Cohen's Kappa test. The intra- and interassay variability was calculated for each laboratory and the agreement between samples submitted to the same laboratory was also determined using the Cohen's Kappa test. For n=23/41 patients tested simultaneously in the three laboratories, all three values were either uniformly below, within or above the respective reference interval. The Cohen's Kappa value for intra- and interassay agreement was 1.0 in laboratory2 (n=15, complete agreement), 0.33 in laboratory1 (n=16) and 0.37 (intra-) and 0.19 (interassay agreement) in laboratory3 (n=16, low agreement). There was a low agreement between laboratories1 and 2 and between laboratories1 and 3 (κ=0.30 and 0.25, respectively), while a high agreement was determined between laboratories2 and 3 (κ=0.68). The intrassay variability of laboratories1, 2 and 3 was 13.6%, 5.0% and 10.4%, the interassay variability 17.2%, 5.1% and 17.4%, respectively. The differences in the measurement of thyroxine concentrations of the same serum sample in different laboratories and at different time points in the same laboratory underline the high relevance of interpreting laboratory results in context with the clinical signs of hypothyroidism as well as other laboratory values such as TSH concentration.