摘要：Thromboelastic testing provides an assessment of a patient’s coagulation and fibrinolytic systems. In recent years, thromboelastic testing has become an important point of care technique. However, its direct connection with the underlying biochemistry of coagulation and clot formation is not obvious. Toward this issue, we describe a validated reduced order mathematical model of coagulation and fibrinolysis, consisting of 22 ordinary differential equations, which described clot formation from initiation of the coagulation cascade through the degradation of polymerized fibrin by plasmin. We trained the model via leave one out cross validation on ROTEM measurements, a common thromboelastic test, on four patients, and then predicted ROTEM trajectories on four unseen patients, in whole blood and whole blood with the addition of 2 nM tissue plasminogen activator. Following model validation, sensitivity analysis suggested which biochemical interactions and species controlled the system response. Lastly, we investigated if we could estimate protein concentrations from commonly reported thromboelastic metrics. These estimation studies suggested we could (on average) relearn the initial fibrinogen concentration to within 20% of its true value. Taken together, this work presents a model which connects the underlying biochemistry of coagulation and clot formation in patients to a common point of care thrombelastographic test.
