Background: Therapeutic anticoagulation with Warfarin among patients with atrial fibrillation reduces the risk of stroke. To achieve this risk reduction, it is important that the TTR is > 65%. Novel oral anticoagulants are noninferior or superior to Warfarin in reducing the risk of stroke and do not require maintaining a narrow therapeutic window. The goals of this study were to determine the TTR in a safety-net hospital system as well as predictors of poor TTR. Second, we set out to determine the association between medication nonadherence and TTR. If nonadherence is a significant independent predictor of TTR, this implies that a universal policy of switching to NOACs without considering patient adherence may not improve outcomes.
Methods: We queried the Parkland Health and Hospital System electronic medical record for all patients from January 2010 to 2016 with a diagnosis of atrial fibrillation or flutter on EKG who were on warfarin and had INR results available during the study period. TTR was calculated using the Rosendaal method. Patients were grouped into TTR quartiles. Adherence to Warfarin was defined by the proportion of days covered to the medication (0 – 100%). Multivariable linear regression modeling was used to assess the associations between demographic, comorbid conditions, and adherence and TTR.
Results: A total of 2,626 patients were included in the analysis. There was significant variability in the TTR with the median TTR of 50% and IQR of 31 - 66. Younger age, black race, alcohol and drug use history, use of antiplatelet medications, and systolic blood pressure >160 was associated with having TTR in the lowest quartile (<31%). Medicare status was associated with lower likelihood of being in the lowest quartile of TTR. Finally, lower adherence was associated with increased likelihood of having poor TTR (Table).
Conclusions: TTR to Warfarin in this underserved population is suboptimal. Given that adherence to Warfarin is independently associated with poor TTR, in the absence of strong adherence interventions, a policy of universal NOAC adoption is unlikely to significantly improve outcomes. If patients are transitioned to NOACs, this data implies that a targeted adherence intervention will be necessary to ensure that the medication is effective in reducing the risk of stroke.