Figure 12 shows the results from this VA Health System evidence-based synthesis program.  Although the results were not statistically significant, there was a clear trend, with a 17% reduction in all-cause mortality, low-strength evidence for benefit against VTE-related mortality, and a suggestion of benefit for the novel anticoagulants.[6] The target-specific NOACs showed a reduction in recurrent DVT/PE of about 15%, a numerical decrease in major bleeding, but a higher rate of treatment discontinuation.  

Of course it is always important to look at the confidence intervals when interpreting data, and although all of these endpoints demonstrated a numerical decrease (except for discontinuation), none of the endpoints achieved statistical significance because the confidence intervals crossed the line of unity (the number 1), for no effect.  

Nevertheless, the authors of this well-done synthesis concluded that newer anticoagulants are no worse than dose-adjusted warfarin for the major clinical outcomes, which is the most objective interpretation of these data.  Hylek E. Am J Med 2013; published on-line at http://education.amjmed.com/00000.