Adherence to Dabigatran and Longitudinal Outcomes
Adherence to Dabigatran and Longitudinal Outcomes
The objective of this study was to describe adherence and its association with clinical outcomes in a national cohort of Veterans treated with dabigatran. We found that in patients with non-valvular atrial fibrillation, adherence to dabigatran in the first year was good for a majority of patients. However, more than one-quarter of patients had sub-optimal adherence to dabigatran and poor adherence was associated with an increased risk for stroke and all-cause mortality. There was no association between dabigatran adherence and non-fatal bleeding or MI. This report is one of the first nationwide cohort study assessing patterns of adherence to dabigatran in an integrated healthcare system and its association with outcomes.
Prior studies have suggested that the differences between clinical trial efficacy and real-world effectiveness are primarily due to patient selection and an increased potential for medication non-adherence. In the RE-LY trial, dabigatran was associated with a lower risk of stroke and systemic embolism compared to warfarin with comparable rates of major hemorrhage. Subsequent analyses of real-world cohorts have shown that over a third of atrial fibrillation patients fail to receive appropriate, guideline-directed oral anti-coagulant or anti-platelet therapies. Additionally, under-treatment of eligible patients with oral anti-coagulants was found to be associated with a two-folds higher odds of a thromboembolic event underscoring the importance of appropriate therapy utilization. Nonetheless, real-world cohort studies evaluating dabigatran against warfarin have shown comparable safety and effectiveness. For example, Larssen et al examined a national cohort of Danish patients on dabigatran with atrial fibrillation and found lower rates of stroke and mortality without an increased risk of bleeding or MI with dabigatran compared to warfarin. Similarly, the FDA conducted a mini-sentinel study to examine rates of adverse events on dabigatran and found comparable bleeding rates with dabigatran and warfarin. However, real-world patient adherence to dabigatran therapy was not in any of these studies examined. While The RE-LY and RE-COVER trials reported over 95% adherence to dabigatran, only 72% of patients were noted to be adherent in our analyses highlighting critical differences between a randomized setting and the real-world scenario. Furthermore, internal national VA metrics show that 68% of patients receiving warfarin are adherent. Because of these reasons, the real-world effectiveness of dabigatran may be lower than that observed in the RE-LY trial.
The results of this study support that initiating dabigatran in atrial fibrillation patients in of itself is not adequate to reduce the stroke risk. Since dabigatran and the other new oral anticoagulation agents do not require routine laboratory monitoring, they may require closer clinical follow-up to ensure adequate adherence. Potential reasons for suboptimal adherence to dabigatran in clinical practice may include lack of close follow-up such as that seen in a trial-setting, twice daily dosing, gastro-intestinal side effects, poly-pharmacy given the high co-morbidity burden and financial reasons. Studies evaluating adherence to warfarin in a setting similar to ours have shown that adherence improves with attendance at high-performing anticoagulation clinics. Consistent with this observation, individuals transitioned from warfarin had a higher odds of being adherent to dabigatran compared to individuals started on dabigatran de novo. Thus, multi-modal interventions aiming at improving adherence to dabigatran may be beneficial and future studies should evaluate the effect of regular follow-up on dabigatran adherence.
Our study should be interpreted in light of several considerations. First, our study cohort comprises exclusively of US Veterans. This cohort is predominantly male and has a higher prevalence of factors associated with poorer adherence. Therefore, our results do not necessarily generalize to cohorts under-represented in this study (for example women, or non US populations). Nonetheless, the VA provides a unique opportunity to assess medication adherence since it has a closed pharmacy system and high reliability and validity of its data sources. Furthermore, as patients can obtain dabigatran with lower co-pay at VA pharmacies compared to non-VA pharmacies, it is unlikely that patients obtained dabigatran from other sources. Second, due to the retrospective nature of our study, there may be an under-estimation of event rates in our cohort. However, we included all hospitalizations at non-VA facilities paid for by the VA and all-cause mortality as an outcome to ensure complete capture given high sensitivity of the VA vital index file. Cause-specific mortality could not be ascertained to examine non-cardiovascular and non-bleeding causes of death in the study cohort. Third, while we utilized pharmacy databases that accurately capture medication dispensing, we do not know whether dispensed medications were actually taken. Currently, laboratory tests for dabigatran levels are not available and other surveillance measures such as electronic monitors recording bottle opening are not readily feasible. Moreover, refill compliance has been shown to be an accurate marker of patient adherence in closed pharmacy systems such as the VA when measured at multiple points in time. Further, we accounted for all in-hospital stays in our calculation of medication gaps and PDC as well as physician cancellation of the medication. Therefore our estimates of refill adherence are more specific than prior measures. Also, since we analyzed PDC as a time varying covariate, we addressed the potential for survival bias. Fourth, dabigatran was approved for use relatively recently. Accordingly, our follow-up duration is relatively short and may explain the higher PDC compared to prior reports on other cardiovascular medications. Future studies should evaluate adherence to dabigatran over a longer duration of follow-up. Lastly, due to the observational nature of our study, residual confounding in assessment of the association of adherence and outcomes cannot be completely eliminated. However, we accounted for covariates known to modify this association and used robust statistical techniques to assess the association between poor adherence and outcomes.
In conclusion, in this national cohort study, we found that the majority of patients who initiated therapy with dabigatran for atrial fibrillation were adherent. However, more than one-quarter of patients were non-adherent and lower adherence was associated with increased risk of stroke/death. These findings suggest the advantages of dabigatran relative to warfarin in terms of laboratory monitoring and reduced interactions must be weighed against the implications of non-adherence on patient outcomes. Further, these findings highlight the need for concerted efforts to bolster adherence to dabigatran to ensure optimal patient outcomes. Future studies should evaluate interventions aiming at improving dabigatran adherence.
Discussion
The objective of this study was to describe adherence and its association with clinical outcomes in a national cohort of Veterans treated with dabigatran. We found that in patients with non-valvular atrial fibrillation, adherence to dabigatran in the first year was good for a majority of patients. However, more than one-quarter of patients had sub-optimal adherence to dabigatran and poor adherence was associated with an increased risk for stroke and all-cause mortality. There was no association between dabigatran adherence and non-fatal bleeding or MI. This report is one of the first nationwide cohort study assessing patterns of adherence to dabigatran in an integrated healthcare system and its association with outcomes.
Prior studies have suggested that the differences between clinical trial efficacy and real-world effectiveness are primarily due to patient selection and an increased potential for medication non-adherence. In the RE-LY trial, dabigatran was associated with a lower risk of stroke and systemic embolism compared to warfarin with comparable rates of major hemorrhage. Subsequent analyses of real-world cohorts have shown that over a third of atrial fibrillation patients fail to receive appropriate, guideline-directed oral anti-coagulant or anti-platelet therapies. Additionally, under-treatment of eligible patients with oral anti-coagulants was found to be associated with a two-folds higher odds of a thromboembolic event underscoring the importance of appropriate therapy utilization. Nonetheless, real-world cohort studies evaluating dabigatran against warfarin have shown comparable safety and effectiveness. For example, Larssen et al examined a national cohort of Danish patients on dabigatran with atrial fibrillation and found lower rates of stroke and mortality without an increased risk of bleeding or MI with dabigatran compared to warfarin. Similarly, the FDA conducted a mini-sentinel study to examine rates of adverse events on dabigatran and found comparable bleeding rates with dabigatran and warfarin. However, real-world patient adherence to dabigatran therapy was not in any of these studies examined. While The RE-LY and RE-COVER trials reported over 95% adherence to dabigatran, only 72% of patients were noted to be adherent in our analyses highlighting critical differences between a randomized setting and the real-world scenario. Furthermore, internal national VA metrics show that 68% of patients receiving warfarin are adherent. Because of these reasons, the real-world effectiveness of dabigatran may be lower than that observed in the RE-LY trial.
The results of this study support that initiating dabigatran in atrial fibrillation patients in of itself is not adequate to reduce the stroke risk. Since dabigatran and the other new oral anticoagulation agents do not require routine laboratory monitoring, they may require closer clinical follow-up to ensure adequate adherence. Potential reasons for suboptimal adherence to dabigatran in clinical practice may include lack of close follow-up such as that seen in a trial-setting, twice daily dosing, gastro-intestinal side effects, poly-pharmacy given the high co-morbidity burden and financial reasons. Studies evaluating adherence to warfarin in a setting similar to ours have shown that adherence improves with attendance at high-performing anticoagulation clinics. Consistent with this observation, individuals transitioned from warfarin had a higher odds of being adherent to dabigatran compared to individuals started on dabigatran de novo. Thus, multi-modal interventions aiming at improving adherence to dabigatran may be beneficial and future studies should evaluate the effect of regular follow-up on dabigatran adherence.
Our study should be interpreted in light of several considerations. First, our study cohort comprises exclusively of US Veterans. This cohort is predominantly male and has a higher prevalence of factors associated with poorer adherence. Therefore, our results do not necessarily generalize to cohorts under-represented in this study (for example women, or non US populations). Nonetheless, the VA provides a unique opportunity to assess medication adherence since it has a closed pharmacy system and high reliability and validity of its data sources. Furthermore, as patients can obtain dabigatran with lower co-pay at VA pharmacies compared to non-VA pharmacies, it is unlikely that patients obtained dabigatran from other sources. Second, due to the retrospective nature of our study, there may be an under-estimation of event rates in our cohort. However, we included all hospitalizations at non-VA facilities paid for by the VA and all-cause mortality as an outcome to ensure complete capture given high sensitivity of the VA vital index file. Cause-specific mortality could not be ascertained to examine non-cardiovascular and non-bleeding causes of death in the study cohort. Third, while we utilized pharmacy databases that accurately capture medication dispensing, we do not know whether dispensed medications were actually taken. Currently, laboratory tests for dabigatran levels are not available and other surveillance measures such as electronic monitors recording bottle opening are not readily feasible. Moreover, refill compliance has been shown to be an accurate marker of patient adherence in closed pharmacy systems such as the VA when measured at multiple points in time. Further, we accounted for all in-hospital stays in our calculation of medication gaps and PDC as well as physician cancellation of the medication. Therefore our estimates of refill adherence are more specific than prior measures. Also, since we analyzed PDC as a time varying covariate, we addressed the potential for survival bias. Fourth, dabigatran was approved for use relatively recently. Accordingly, our follow-up duration is relatively short and may explain the higher PDC compared to prior reports on other cardiovascular medications. Future studies should evaluate adherence to dabigatran over a longer duration of follow-up. Lastly, due to the observational nature of our study, residual confounding in assessment of the association of adherence and outcomes cannot be completely eliminated. However, we accounted for covariates known to modify this association and used robust statistical techniques to assess the association between poor adherence and outcomes.
In conclusion, in this national cohort study, we found that the majority of patients who initiated therapy with dabigatran for atrial fibrillation were adherent. However, more than one-quarter of patients were non-adherent and lower adherence was associated with increased risk of stroke/death. These findings suggest the advantages of dabigatran relative to warfarin in terms of laboratory monitoring and reduced interactions must be weighed against the implications of non-adherence on patient outcomes. Further, these findings highlight the need for concerted efforts to bolster adherence to dabigatran to ensure optimal patient outcomes. Future studies should evaluate interventions aiming at improving dabigatran adherence.
Source...