Background Although the proportion of the total burden of incident stroke that can be attributed to underuse of anticoagulation in patients with known prior AF is uncertain, it could be considerable given the severity of AF-related stroke.

Methods We studied all incident ischaemic strokes (IIS) from 2002–2008 in a population-based study and determined 1-month and 6-month outcome (modified Rankin scale—mRS) in AF-related stroke versus all other cases.

Results AF-related stroke accounted for 30.2% (222/734) of all IIS. There was no difference in premorbid mRS or living arrangements between the AF and non-AF groups (age-adjusted p=0.45 and 0.06 respectively), but of all IIS that were disabling or fatal at 6 months, 43.2% (156/361) were AF-related and 30.7% (111/361) occurred in patients with known prior AF. Of the 111 patients with IIS that were disabling or fatal at 6 months and known prior AF, 88 had premorbid CHADS2 score ≥2, but only 11 (12.5%) were on warfarin prior to the stroke. Similarly, of 222 patients who were institutionalised or dead at 6 months after the stroke, 85 (38.3%) had known prior AF, of whom 71 had CHADS2 score ≥2 but only 6 (8.5%) were on warfarin prior to stroke.

Conclusion In the context of very low use of warfarin for primary prevention, about 1 in 3 disabling or fatal IIS occur in patients with known prior AF. The majority of this considerable burden of major stroke could be prevented by greater use of anticoagulation in eligible patients.