Risk of Ischemic Stroke, Hemorrhagic Stroke, Bleeding, and Death in Patients Switching from Vitamin K Antagonist to Dabigatran after an Ablation

Abstract

Background: Safety regarding switching from vitamin K antagonist (VKA) to dabigatran therapy in post-ablation patients has never been investigated and safety data for this is urgently needed. The objective of this study was to examine if switch from VKA to dabigatran increased the risk of stroke, bleeding, and death in patients after ablation for atrial fibrillation.

Methods: Through the Danish nationwide registries, patients with non-valvular atrial fibrillation undergoing ablation were identified, in the period between August 22nd 2011 and December 31st 2015. The risk of ischemic stroke, hemorrhagic stroke, bleeding, and death, related to switching from VKA to dabigatran was examined using a multivariable Poisson regression model, where Incidence rate ratios (IRR) were estimated using VKA as reference.

Results: In total, 4,236 patients were included in the study cohort. The minority (n = 470, 11%) switched to dabigatran in the follow up period leaving the majority (n = 3,766, 89%) in VKA treatment. The patients in the dabigatran group were older, were more often males, and had higher CHA2DS2-VASc, and HAS-BLED scores. The incident rates of bleeding and death were almost twice as high in the dabigatran group compared with the VKA group. When adjusting for the individual components included in the CHA2DS2-VASc and HAS-BLED scores, the multivariable Poisson analyses yielded a non-significant IRR (95%CI) of 1.64 (0.72-3.75) for bleeding and of 1.41 (0.66-3.00) for death associated with the dabigatran group, compared to the VKA group. A significant increased risk of bleeding was found in the 110mg bid group with an IRR (95%CI) of 4.49(1.40-14.5).

Conclusion: Shifting from VKA to dabigatran after ablation was associated with twice as high incidence of bleeding compared to the incidence in patients staying in VKA treatment. The only significant increased risk found in the adjusted analyses was for bleeding with 110mg bid dabigatran and not for 150mg bid. Since there was no dose-response for bleeding, the switch from VKA to dabigatran in itself was not a risk factor for bleeding.

Fig 1. Flowchart of the study cohort.

Legend: NVAF = non-valvular atrial fibrillation; VKA = Vitamin-K-antagonist

Fig 2. Poisson Regression.

Incidence rate ratios of ischemic or hemorrhagic stroke, myocardial infarction, bleeding, death, or a composite of these. Legend: Number of dabigatran users n = 470. Composite is a combined endpoint of ischemic and hemorrhagic stroke, bleeding, and death. All patients with prior events of the investigated endpoint were excluded in the analysis. Models were adjusted for: CHA2DS2-VASc and HAS-BLED components. IR = Incidence Rate, PY = Person years, CI = Confidence Intervals

Fig 3. Poisson Regression.

Incidence rate ratios of ischemic or hemorrhagic stroke, myocardial infraction, bleeding, death, and a composite of these. Legend: Prior dabigatran users defined as dabigatran use prior to ablation. De novo dabigatran users defined as no dabigatran use prior to ablation. One year follow up defined as follow up time for a maximum time of 1 year after ablation. 110 mg dabigatran bid defined as prescribed with 110mg dabigatran at switch date. 150mg dabigatran bid defined as prescribed with 150mg dabigatran at switch date. Composite is a combined endpoint of ischemic and hemorrhagic stroke, bleeding, and death. All patients with prior events of the investigated endpoint were excluded in the analysis. Models were adjusted for: components in CHA2DS2-VASc and HAS-BLED