4. Discussion
To the best of our knowledge, this is the first study to comprehensively develop and evaluate different remedial strategies during non-adherence events in edoxaban treated patients with NVAF. We developed an optimised remedial strategy based on non-adherence scenarios using PK and PD population modelling. Unlike previous studies that focused on PK, our study considered both PK and PD 17-19. iFXa percentage was used as a PD marker in our study, which provides a closer relationship to the clinical outcome than exposure and can help to achieve a more appropriate remedial strategy.
In our study, we used the on-therapy range to overcome the challenge of lacking a defined therapeutic range. The on-therapy range provides a dynamic goal based on the patient’s individual demographic and indicated dosage regimen, which mimics the real-life situation. Additionally, we performed sensitivity analysis to extend our recommendation to patients receiving P-gp inhibitors and/or with low body weight.
According to our PK/PD modelling and simulation, the EHRA recommendation is not optimal and can only be applied when the delay ≤ 11 hours. These results are similar to those of our previous study, showing that EHRA also does not provide optimal recommendations for rivaroxaban missing dose20. Our study shows that from the PK/PD perspective, there is potential room for improvement in the remedial strategy in current practice. The current practice has missed the opportunity to help patients address their missed or delayed drug in a more appropriate manner.
In practice, patients can have different risks of bleeding and thrombosis, which can be assessed through scoring systems such as HAS-BLED or CHA2DS2-VASc for bleeding and stroke, respectively 29, 30. Other risk factors, including concomitant drug use, can also alter a patient’s risk of bleeding and thrombosis. Optimal remedial strategies can be modified based on the patient’s risk of bleeding and thrombosis. It is necessary that healthcare providers balance patients’ risk of bleeding and thrombosis by considering the deviation time above and below the on-therapy range.
Our study has several limitations. First, the results of our study represent the values from the modelling population and may not be extrapolated to patients who do not match the modelling population, especially in patients who take edoxaban for other indications, including venous thromboembolism. Second, the results of our study were not evaluated in real-world patients. However, due to ethical reasons, PK/PD modelling and simulation is the preferred approach for developing and testing remedial strategies in the target population.