Figure 12 illustrates some data that support the ESC treatment consensus statement from the previous Figure.[9]  The data come from the large Swedish Atrial Fibrillation Cohort.[10] The graph in the upper left-hand corner is the most instructive: the 2 curves represent the net clinical benefit for patients with a CHA₂DS₂-VASc stroke-risk score of 0–2 (actually closer to 1–2), ie, the group with AFib but the lowest risk of stroke and a HAS-BLED bleeding-risk score of ≥3 (the vertical ordinate), ie, patients with AFib and a high risk of bleeding. 

In other words, in these patients with a low risk of stroke and a high risk of bleeding, physicians will surely ask, "Do I really want to be given oral anticoagulant therapy here?"

Nevertheless, the “proportion surviving” curves (the y-axis) show that out to 4 years (the x-axis) there is a highly statistically significant (P<.00001) increase in probability of survival for the patients on oral anticoagulation therapy, even in this group of patients.