The Figure shows the ITT analysis of the results.[97] For rivaroxaban vs warfarin these results were 2.1 vs 2.4 events per 100 patient-years, respectively.  The difference was just short of statistical nonsignificance, meaning that rivaroxaban was noninferior, but just short of superior, to warfarin; put another way, the efficacy of rivaroxaban was at least as good as the efficacy of warfarin. 

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