Biological variation of canine serum thyrotropin (TSH) concentration

The aim of the present study was to estimate the between-dog, within-dog and analytical components of variance for serum thyrotropin (TSH) in healthy dogs, and to use these components of variance to 1) estimate the critical difference for significance between serial results; 2) assess the utility of the conventional population-based reference interval; 3) set a desirable performance standard for analytical imprecision; and 4) estimate the number of samples required for determination of the true mean value for an individual dog. Using the Immulite test system, TSH was measured in serum samples collected weekly for five weeks from eight clinically healthy dogs. Results were subjected to nested analysis of variance. Between-dog variation was 43.6%, within-dog variation was 13.6%, analytical variation was 8.8%, the one-sided critical difference was 37.8%, the index of individuality was 0.4, the maximum allowable analytical imprecision was 6.8%, and the number of samples required to determine the true mean value in a single dog was 40. In practical terms, the present study indicated that the analytical imprecision of canine serum TSH measurement should be < 7%, and that comparing a single serum TSH measurement from an individual dog to the conventional population-based reference range may be too insensitive to detect small but important changes in the serum TSH level of that particular dog. In addition, when treating a hypothyroid dog, serum TSH, measured on a weekly basis, should decrease by at least one-third before any effect of exogenous thyroxine supplementation can be said to have influenced the serum TSH level.