Impact of medication adherence on risk of ischemic stroke, major bleeding and deep vein thrombosis in atrial fibrillation patients using novel oral anticoagulants.

Objectives: Our study examined the impact of adherence to novel oral anticoagulants [NOACs - dabigatran and rivaroxaban] on ischemic-stroke (IS), major-bleeding (MB), deep-vein-thrombosis and pulmonary-embolism (DVTPE) risk in a large, nationwide, propensity-matched sample.

Methods: A retrospective cohort study utilized data from a US commercial managed-care database (2010-2012). Adult patients with ≥1 diagnosis of atrial fibrillation/flutter (ICD-9 427.31/32), >1 prescription of NOACs and CHADS-VASc score ≥1 were included. Patients were categorized as adherent versus nonadherent (using proportion of days covered [PDC ≥80%]) based on their NOAC use up to 6 months and those continued its use up to 12 months. The patients were matched using propensity score (based on inverse probability treatment weighting) and the risk of IS, MB, DVTPE outcomes was evaluated for the matched cohorts' post-adherence (exposure) assessment using multivariable Cox regression.

Results: A total of 3,629 and 1,946 patients with at least 6 and 12 months of NOAC use were included. Based on a PDC threshold of ≥80%, adherence rates at 6 and 12 month usage were 77% and 76%, respectively. Patients with lowest adherence were from the South, had low stroke risk and EPO/HMO insurance. Using Cox models with matched cohorts, nonadherence within the first 6 months' use was significantly associated with higher risk of IS and DVTPE (IS: hazard ratio [HR] = 1.82, p = .002; DVTPE: HR = 2.12, p = .010) and the risk increased with nonadherence for the prolonged period of 12 months' use (IS: HR = 2.08, p = .022; DVTPE: HR = 5.39, p = .003). The risk of MB was not different (p > .05) between adherent and nonadherent groups for both 6 month and 12 month cohorts.

Conclusion: Adherence to NOACs for both 6 months and prolonged use (up to 12 months) was associated with a reduction in IS and DVTPE risk, but did not substantially increase risk of MB. Further studies on newer, individual NOACs and older populations are warranted.