This Figure reports the efficacy endpoints by ITT analysis in tabular form.  It appears that the magnitude of superiority favors dabigatran, although it is necessary to stress again that this is not a statistically valid direct comparison, because the populations are not identical.  Importantly, whereas the mean CHADS₂ score in ROCKET AF was 3.5 (meaning these patients were a high-risk group),[97] only about a third of the patients in RE-LY[96] and ARISTOTLE[99] had CHADS₂ scores ≥3, and so those two subgroups are reported separately in the lower part of the table.

 Reiffel JA. Am J Med 2013; 126: 00-00.