Mathematical models of coagulation-are we there yet?

Area of Science:

• Biochemistry
• Computational Biology
• Hematology

Background:

• Mathematical models of coagulation aim to simulate thrombin generation and understand hemostasis regulation by coagulation factors.
• Current models exhibit variability in reactions and rates, with limited validation due to scarce, coherent datasets, raising questions about their utility.

Purpose of the Study:

• To systematically evaluate existing coagulation models against a large dataset of plasma coagulation factor levels from 348 individuals with normal hemostasis.
• To identify the reasons for discrepancies between model predictions and experimental data.

Main Methods:

• Compared predictions from various coagulation models with experimentally measured thrombin generation.
• Quantified and compared model performance in predicting thrombin generation and analyzed the influence of individual reactions and reaction rates.

Main Results:

• No current model accurately predicted hemostatic response across the entire cohort; all generated dissimilar thrombin generation curves compared to experimental data.
• Identified key reactions causing differential model predictions and highlighted the impact of experimental uncertainty on variability.
• Determined reactions with significant influence on measured thrombin generation, including the contribution of factor XI.

Conclusions:

• Systematic assessment using large datasets reveals avenues for improving coagulation models.
• An accurate model reflecting individual hemostatic variations could aid in assessing antithrombotics and advance precision medicine.