Multicenter in vitro thromboelastography and thromboelastometry standardization

TY - JOUR

T1 - Multicenter in vitro thromboelastography and thromboelastometry standardization

AU - Goggs, Robert

AU - Borrelli, Antonio

AU - Brainard, Benjamin M.

AU - Chan, Daniel L.

AU - de Laforcade, Armelle

AU - Goy-Thollot, Isabelle

AU - Jandrey, Karl E.

AU - Kristensen, Annemarie T.

AU - Kutter, Annette

AU - Marschner, Clara B.

AU - Rannou, Benoît

AU - Sigrist, Nadja

AU - Wagg, Catherine

PY - 2018

Y1 - 2018

N2 - Objective: To establish and compare the repeatability and reproducibility of activated thromboelastography (TEG) and thromboelastometry (ROTEM) assays. Design: Multicenter in vitro test standardization. Setting: Veterinary academic centers. Animals: Test samples were obtained from normal, healthy dogs. Sixty identical 5 mL aliquots of canine platelet-rich plasma collected by apheresis, frozen in 6% dimethyl sulfoxide, were tested initially. Sixty identical 6 mL aliquots of canine fresh frozen plasma with admixed cryoprecipitate were subsequently evaluated. Interventions: None. Measurements and Main Results: Frozen study samples, quality controls, reagents, and consumables were distributed to participating centers (7 TEG and 3 ROTEM). TEG centers analyzed study samples with kaolin and tissue factor activated assays; ROTEM centers ran proprietary ellagic acid activated and tissue factor activated assays. All machines underwent quality control prior to sample analysis. Within- and between-center coefficients of variation (CVs) were calculated and compared using Mann–Whitney tests and calculation of intraclass correlation coefficients. Within and between centers, individual parameters for both TEG and ROTEM assays were comparable. Both within-center and between-center CVs varied markedly (0.7–120.5% and 1.4–116.5%, respectively) with assay type, instrument, and parameter. CVs for equivalent parameters were not significantly different between the 2 platforms. Intraclass correlation coefficients suggested moderate agreement between centers. In general, individual parameter CVs for platelet-rich plasma samples were lower in TEG centers, while CVs for canine fresh frozen plasma with admixed cryoprecipitate samples were lower in ROTEM centers. Conclusions: More variation within and between centers was identified than anticipated, but some parameters such as alpha angle were repeatable and reproducible. Sample types for future multicenter standardization efforts will require further optimization and may need to be adapted separately to each platform. Individual centers using viscoelastic tests for evaluation and management of clinical patients should take steps to minimize preanalytical and analytical sources of variation.

AB - Objective: To establish and compare the repeatability and reproducibility of activated thromboelastography (TEG) and thromboelastometry (ROTEM) assays. Design: Multicenter in vitro test standardization. Setting: Veterinary academic centers. Animals: Test samples were obtained from normal, healthy dogs. Sixty identical 5 mL aliquots of canine platelet-rich plasma collected by apheresis, frozen in 6% dimethyl sulfoxide, were tested initially. Sixty identical 6 mL aliquots of canine fresh frozen plasma with admixed cryoprecipitate were subsequently evaluated. Interventions: None. Measurements and Main Results: Frozen study samples, quality controls, reagents, and consumables were distributed to participating centers (7 TEG and 3 ROTEM). TEG centers analyzed study samples with kaolin and tissue factor activated assays; ROTEM centers ran proprietary ellagic acid activated and tissue factor activated assays. All machines underwent quality control prior to sample analysis. Within- and between-center coefficients of variation (CVs) were calculated and compared using Mann–Whitney tests and calculation of intraclass correlation coefficients. Within and between centers, individual parameters for both TEG and ROTEM assays were comparable. Both within-center and between-center CVs varied markedly (0.7–120.5% and 1.4–116.5%, respectively) with assay type, instrument, and parameter. CVs for equivalent parameters were not significantly different between the 2 platforms. Intraclass correlation coefficients suggested moderate agreement between centers. In general, individual parameter CVs for platelet-rich plasma samples were lower in TEG centers, while CVs for canine fresh frozen plasma with admixed cryoprecipitate samples were lower in ROTEM centers. Conclusions: More variation within and between centers was identified than anticipated, but some parameters such as alpha angle were repeatable and reproducible. Sample types for future multicenter standardization efforts will require further optimization and may need to be adapted separately to each platform. Individual centers using viscoelastic tests for evaluation and management of clinical patients should take steps to minimize preanalytical and analytical sources of variation.

KW - canine

KW - dog

KW - repeatability

KW - reproducibility

KW - thromboelastography

KW - thromboelastometry

U2 - 10.1111/vec.12710

DO - 10.1111/vec.12710

M3 - Journal article

C2 - 29604163

AN - SCOPUS:85046399587

SN - 1476-4431

VL - 28

SP - 201

EP - 212

JO - Journal of Veterinary Emergency and Critical Care

JF - Journal of Veterinary Emergency and Critical Care

IS - 3

ER -