A comparison of assay methods for thrombin activatable fibrinolysis inhibitor (TAFI)

Abstract number: P0788

Guimarães^* A. H. C., Van Tilburg‡ N. H., Vos‡ H., Bertina‡ R.M., Rijken^* D.C., Guimarães† A. H. C.

‡Department of Hematology, Hemostasis and Thrombosis Reserch Center, Leiden University Medical Center, Netherlands †The Institure for Cardiovascular Research , VU Medical Center, Amsterdam, Netherlands ^*TNO Prevention and Health, Gaubius Laboratory, Leiden, Netherlands

Thrombin activatable fibrinolysis inhibitor (TAFI) antigen levels exhibit a large interindividual variability in which genetic control seems to play a major role. Several methods are currently available for determining TAFI plasma concentration. However, recent reports implied that genotype-dependent variation of TAFI concentration might include an artefact caused by different antibody reactivity towards TAFI isoforms. Especially the 1040C/T (Thr325Ile) polymorphism seems to be involved. In order to assess assay agreement we determined plasma TAFI levels in 92 healthy volunteers making use of different assays. An in-house ELISA using commercial sheep polyclonal antibodies (Affinity Biologicals Inc.) and a rocket immunoelectrophoresis using home-made rabbit polyclonal antibodies were used to determine TAFI antigen levels. A commercial chromogenic assay (Actichrome^® TAFI, American Diagnostica) was used as an alternative method for determining TAFI concentration in plasma. Each individual was also genotyped for the 1040C/T polymorphism. We found a significant association of TAFI levels with genotype as previously reported for antigen data. The C/C genotype corresponded to the highest TAFI levels and the T/T to the lowest for all the assays used (ELISA r = 0.524; P < 0.0001; immunoelectrophoresis r = 0.329; P = 0.0009 and Actichrome^® TAFI r = 0.358; P = 0.0004). The Actichrome^® TAFI assay was significantly correlated with the ELISA (r = 0.554, P < 0.0001) and with the immunoelectrophoresis (r = 0.677; P < 0.0001). From linear regression analysis in the genotype subgroups we concluded that the correlation between the Actichrome^® TAFI assay and the immunoelectrophoresis was not affected by the genotype. The correlation between the Actichrome^® TAFI and the ELISA, however, was genotype-dependent in line with a less efficient recognition of the Ile-325 variant in the ELISA. In conclusion, our results demonstrate that artefacts may indeed arise when measuring TAFI antigen levels by ELISA. Nevertheless, for the rocket immunoelectrophoresis and for the Actichrome^® TAFI assay no artefacts were detected, therefore confirming a true genotype-related variation of TAFI concentration.