SOURCE: Vespa PM. JAMA 2015;312:2562-2563.

Clinicians have commonly overestimated the risk of intracerebral hemorrhage (ICH) during anticoagulant therapy. Indeed, such misapprehensions have sometimes led to failure to employ warfarin (and probably other agents) when indicated for atrial fibrillation. There is little dispute that novel oral anticoagulants (apixiban, dabigatran, rivaroxaban) are simpler to use, since they do not require monitoring and are essentially free of food interactions. Clinical trials with novel oral anticoagulants (NOACs) have consistently documented that NOACs are associated with lesser risk of ICH, which is certainly a good thing … but how much of a good thing?

First, it may come as a surprise that in the modern era, large clinical trials of warfarin treatment in atrial fibrillation demonstrate ICH events consistently below 1% per year. Since the risk of thrombotic stroke in atrial fibrillation — even at a CHADS score of 1 — is approximately 3% per year, the risk:benefit ratio is strongly in favor of anticoagulation.

Atrial fibrillation mega-trials (each > 10,000 patients) have been completed with the three FDA-approved NOACs, each agent demonstrating a reduction in ICH compared to warfarin, with an overall odds ratio of 0.49 — essentially half the risk of ICH with NOACs compared to warfarin. ICH is a devastating consequence of anticoagulation, and although more than 99% of patients treated with warfarin for atrial fibrillation per year will not suffer ICH, the ability to reduce risk for such a dreaded event is an important consideration.