Critical evaluation of kinetic schemes for coagulation

Computational models of the coagulation cascade are used for a wide range of applications in bio-medical engineering such as drug and bio-medical device developments. However, a lack of robustness of numerical models has been highlighted when studying clinically relevant scenarios. In order to develop more robust models, numerical simulations need to be confronted with realistic situations relevant to clinical practice. In this work, two well-established numerical representations of the coagulation cascade initiated by the intrinsic and extrinsic systems, respectively, were compared with thrombin generation assays considering realistic pathological conditions. Proper modifications were needed to align the in vitro and in silico data, namely; adapting initial conditions to the thrombin assay system, omitting reactions irrelevant to our case study, and improving the fitting of some reaction rates. The modified models were able to capture the experimental trends of thrombin generation for a range of concentrations of factors XII, XI, and VIII for cases in which the coagulation cascade is triggered through the extrinsic and intrinsic systems. Our work emphasizes that when existing coagulation cascade models are extrapolated to experimental settings for which they were not calibrated, careful adjustments must be made. We show that the two coagulation models used in this work can predict physiological conditions, but when studying pathological conditions, proper modifications are needed to improve the numerical results.