There are a number of limitations of this meta-analysis that need to be considered when interpreting the results. The trials included in this analysis, as well as prior meta-analyses, may not have taken into account differences in the time-in-therapeutic- range of the INR. This feature of warfarin use has been shown to vary significantly among the sites of the ACTIVE W trial.44 The definition of stroke may have varied between studies. For example, Hart et al42 included intracranial hemorrhage in the primary outcome of ‘‘all strokes,’’ whereas we defined stroke as the acute onset of a focal neurological deficit of presumed vascular origin lasting for 24 hours or more. The number of studies included in the current analysis was relatively small, and the earliest study was performed in 1989 and the latest in 2010, and various treatment regimens were used over this 2-decade period. This highlights the need for additional high-quality studies to address this issue. Patient characteristics of the included studies were not homogeneous, and all but 2 studies enrolled patients >70 years of age, and thus, the results may not be applicable to younger age groups. Most of the included studies reported data that was analyzed on an intent-to-treat basis, and thus, the pooled effect estimates may have been underestimated.
The only antiplatelet that was studied was aspirin, and newer antiplatelet drugs are available and their effectiveness compared with anticoagulation has yet to be determined. Lastly, several of the studies included in the analysis were found to have a high risk of performance or detection bias.