B-005 Comparison of 3 assays for quantitation of D-dimer concentrations in canine plasma samples

Abstract Background D-dimers are the terminal degradation product of fibrin cross-linked by coagulation factor XIII. Increased plasma D-dimer concentrations indicate that both thrombin generation and plasmin-mediated fibrinolysis have occurred to a greater extent than under basal conditions. The most frequent and clinically important causes of increased plasma D-dimer concentrations are thromboembolic disorders including disseminated intravascular coagulation (DIC). Typically, D-dimers are quantitated using immunologic methods such as enzyme-linked immunosorbent assays and automated immunoturbidometric tests. For all immunologic assays, the sensitivity and specificity of the antibodies employed substantially impacts the utility of the test for a given species. Dogs are the most popular companion animal in the United States, with almost 90 million animals in 65 million households. Dogs also represent an important large-animal species for pharmaceutical and comparative biomedical research. Most D-dimer assays used to test canine plasma samples involve antibodies raised against human D-dimers, but there are substantial differences in the affinity of these antibodies for canine D-dimer epitopes Methods We compared the utility of 3 commercially available D-dimer assays for measurement of D-dimer concentrations in canine plasma samples. Two assays (Helena D-dimer, Helena Biosciences and HemosIL D-dimer, Instrumentation Laboratory) employed the same antibody clone (8D3), whereas the third assay (STA Liatest D-Di, Diagnostica Stago) employed a mixture of 2 distinct antibody clones (8D2 and 2.1.16). All 3 assays were performed on the same automated assay platform (STA Compact Max, Diagnostica Stago). Tested canine plasma samples were those submitted to the Comparative Coagulation Laboratory, Animal Health Diagnostic Center, Cornell University College of Veterinary Medicine, Ithaca, NY for routine analysis of D-dimer concentrations. Incurred samples stored frozen (-80°C) with previously measured D-dimer concentrations spanning a clinically relevant range were re-analyzed using 2 additional methods. Results were compared with scatterplots, Deming regression, Spearman’s correlation coefficients and Bland-Altman plots. Results Compared to the original reference method (HemosIL D-dimer) the D-dimer concentrations quantitated using the Helena D-dimer assay were linearly associated per Deming regression (Y=0.484X–140.1, P<0.0001). Deming regression showed no significant association between D-dimer concentrations measured with the HemosIL D-dimer assay and the STA Liatest assay. The Spearman correlation coefficients for these comparisons were 0.872 (P<0.0001) and 0.418 (P=0.0337) respectively. Bland-Altman plotting suggested a mean difference of 734 ± 672 ng/mL (DDU) between the HemosIL and the Helena assays with a greater degree of positive bias at high concentrations above a clinically relevant cutoff. Conclusion The 2 assays incorporating the 8D3 monoclonal antibody yielded concordant results for measurement of canine plasma D-dimer concentrations. The third assay using a distinct combination of antibodies had disparate values. Our results indicate that anti-human D-dimer antibodies vary in reactivity, and independent verification of assay performance is required for multi-species usage.