• APTT;
  • fibrinogen;
  • hemostasis;
  • horse;
  • PT;
  • TEG


There is a paucity of information about the analytical performance of thromboelastography (TEG) in horses, specifically concerning components of variation among different analytical methods. Such data may be obtained by nested analysis of repetitive standardized sampling of healthy individuals.


The objectives were (1) to assess the relative susceptibility to sources of preanalytical variation in a highly standardized setting, (2) to directly compare and evaluate the observed analytical variation, and (3) to assess the applicability of population-based reference intervals.


Blood was collected once from 20 healthy adult horses and weekly for 5 consecutive weeks from 8 healthy adult horses. TEG analyses were performed on citrated whole blood using nonactivated TEG and 2 TEG assays with human recombinant tissue factor (TF) or kaolin as activators.


There were significant differences between nonactivated and activated assays for all measured parameters (P < .05). The intraindividual variation was significantly higher with the nonactivated assay compared with TF- and kaolin-activated protocols. Analytical variation was generally low and comparable for all 3 assays. Nonactivated TEG showed the highest degree of within-subject variation, TF-activated TEG (TF-TEG) intermediate and kaolin-activated TEG (K-TEG) the lowest degree of within-subject variation.


Nonactivated TEG is more sensitive to preanalytical variation than both TF-TEG and K-TEG. Analytical variation was low and comparable for all assays, but not all parameters reached objective analytical goals. The results further indicate that population-based reference intervals can be readily applied to TF-TEG, but not to nonactivated or K-TEG, where critical difference may provide a better interpretation criterion.