Scarce data comparing real-world outcomes between apixaban and vitamin K antagonist (VKA) users with nonvalvular atrial fibrillation (NVAF) are available. We sought to assess the effectiveness and safety of newly-initiated apixaban vs. VKA in German NVAF patients.

We performed a retrospective analysis in German outpatients using IMS Disease Analyzer data. Adults newly-initiated on apixaban or a VKA from January 2013 to March 2015 with a diagnosis of NVAF on the day of the first qualifying oral anticoagulant (OAC) prescription (index date) or any time during 1 year prior, and at least 1 year of follow-up were included. Patients experiencing a prior event in the composite endpoint, receiving an OAC before the index date, >1 OAC on the index date or switched to another OAC during follow-up were excluded. Apixaban and VKA users were 1:1 propensity-score matched. We evaluated the composite of ischaemic stroke, transient ischaemic attack (TIA), myocardial infarction (MI) or intracranial haemorrhage (ICH) in the year after OAC initiation. Cox regression was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs).

In total, 835 apixaban and 835 VKA users were matched. Forty-one composite events were identified. Hazard of the composite endpoint did not differ between apixaban and VKA users (HR=0.87, 95%CI=0.47–1.60). Ischaemic stroke and MI occurred at dissimilar (albeit not statistically significant) rates between apixaban and VKA therapy (HR=1.51, 95%CI=0.54–4.24) and (HR=0.33, 95%CI=0.11–1.03). Only two patients (both in the apixaban cohort) experienced an ICH.

Apixaban and VKA therapy were associated with a similar impact on the composite endpoint in real-world German practice. Additional investigation is needed to evaluate the numeric trends of ischaemic stroke and decreased number of MIs observed with apixaban, as well as the high rate of reduced dose apixaban use found in this analysis.