Aaron Tansy, MD
Atrial fibrillation (AF) is the most common cardiac arrhythmia afflicting adults, and confers an estimated 5-fold increased risk for both ischemic/thromboembolic events and death.  The use of oral anticoagulation (OAC) has been well established to diminish this risk significantly, and is therefore a staple of stroke prevention in the management of AF.  However, despite OAC compliance, some patients still sustain new strokes.  These events occur likely for a variety of reasons that may run the gamut from transient sub-therapeutic OAC levels to poorly treated coexisting risk factors. 
To better understand this occurrence, Ida Albertsen and colleagues present a meta-analysis of recent randomized controlled trials of AF stroke prevention to identify clinical factors that may worsen risk of stroke in AF despite compliant use of warfarin.  After performing a thorough database search and screening, they selected six studies for analysis which identified the following factors that independently increased risk of stroke in treated AF: age >/ 75, female sex, prior stroke/TIA, naïve vitamin K-antagonist status, moderate and severe renal impairment, prior aspirin use, Asian race, and a CHADS2 score >/ 3.  Risk of stroke appeared highest for those either with a prior ischemic event, severe renal impairment, or high CHADS2 score.  Independently, diabetes, hypertension, and heart failure did not worsen stroke risk.  Data regarding etiology/mechanism of new strokes (thromboembolic, ICH, small-vessel disease) were not available to the authors for analysis.
Although lacking information about etiology of new strokes in treated AF, this group’s study provides some surprising and not-so surprising results.  Regarding the former, Asian ethnicity was found to increase stroke risk in treated AF.  But, a few previous studies with largely white populations suggested a decreased risk for Asians.  Why the discrepancy?  Ethnicity may play a true role, or may reflect other underlying factors such as imbalance in ethnic diversity of study populations or even those socioeconomic.  Thus, more ethnically diverse and global epidemiological studies are necessary to better address this finding.  Regarding the latter, prior antiplatelet use increases risk of stroke in treated AF.  Likely this finding reflects concurrent non-AF stroke risk factors.  However, surprisingly, only in aggregate as reflected in the CHADS2 score, did the majority of these well-established independent risk factors for non-AF stroke increase stroke risk in treated AF.  Further studies are required to tease apart the independent risks of each of these in the setting of treated AF, and how, in aggregate, they may confer heightened risk even despite OAC compliance.  Lastly, above all, with the advent of non-vitamin K-antagonist OACs, additional studies are warranted to understand if these newer drugs may be superior to warfarin for stroke prevention in the high-risk sub-groups that this analysis identified.