The Figure shows the conclusions of a recent study from Denmark featuring a direct comparison of the CHADS₂ and CHA₂DS₂-VASc scoring systems, using a nationwide registry of patients admitted to the hospital with AFib who were not anticoagulated. [48] The registry totaled almost 75,000 patients and represents the largest real world cohort ever investigated. 
The main outcome measured was stroke and systemic embolism.  In patients with a score of 0 by either score, the event rate per 100 patient-years was 1.67 with CHADS₂ versus 0.78 with CHA₂DS₂-VASc.  In patients with a score of 1, the event rate per 100 patient-years was 4.75 with CHADS₂ and 2.01 with CHA₂DS₂-VASc.  
The important point to remember is that for a female age 66 with an MI, the score will be 3 according to CHA₂DS₂-VASc and 0 in CHADS₂.  This illustrates why the lower scores in CHA₂DS₂-VASc are associated with a better accuracy of low risk – because more patients who are classified as low-to- intermediate risk by CHADS₂ are actually at risk, whereas their risk is more accurately classified by CHA₂DS₂-VASc.  

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