Hemodynamics affects factor XI/XII anticoagulation efficacy in patient-derived left atrial models

Atrial fibrillation (AF) is a common arrhythmia that disrupts blood circulation in the left atrium (LA), causing stasis in the left atrial appendage (LAA) and increasing thromboembolic risk. In patients at sufficiently high risk, anticoagulation is indicated. This benefit may be counterbalanced by an increased risk of bleeding. Novel anticoagulants under development, such as factor XI/XII inhibitors, may be associated with a lower bleeding risk. However, their efficacy in preventing thrombosis is not fully understood. We hypothesized that patient-specific flow patterns in the LA and LAA not only influence the risk of thrombosis but also the effectiveness of anticoagulation agents.

To test our hypothesis, we simulated blood flow and the intrinsic coagulation pathway in patient-specific LA anatomies with and without factor XI/XII inhibition. We included a heterogeneous cohort of thirteen patients, some in sinus rhythm and others in AF, four of whom had an LAA thrombus or a history of transient ischemic attacks. We used computational fluid dynamics based on 4D CT imaging and a detailed 32-coagulation factor system to run 247 simulations. We analyzed baseline LA flow patterns and evaluated various factor XI/XII inhibition levels. Implementing a novel multi-fidelity coagulation modeling approach accelerated computations by two orders of magnitude, enabling many simulations to be performed.

The simulations provided spatiotemporally resolved maps of thrombin concentration throughout the LA, showing that it peaks inside the LAA. Coagulation metrics based on peak LAA thrombin dynamics suggested patients could be classified as having no, moderate or high thromboembolic risk. High-risk patients had slower flows and higher residence times in the LAA than those with moderate thromboembolic risk, and they required stronger factor XI/XII inhibition to prevent thrombin growth. These data suggest that the anticoagulation effect was also related to the LAA hemodynamics.

The methodology outlined in this study has the potential to enable personalized assessments of coagulation risk and to tailor anticoagulation therapy by analyzing flow dynamics in patient-derived LA models, representing a significant step towards advancing the application of digital twins in cardiovascular medicine.