The CHA₂DS₂-VASc score has been validated and is widely used to stratify the risk of thromboembolism in patients with non-valvular atrial fibrillation (AF). We sought to investigate whether this score could also be useful to predict the risk of stroke and death in patients with valvular AF.

Between 1998 and 2011, 1,592 consecutive patients, hospitalised for AF, 300 with valvular AF (mitral and/or aortic valve disease) and 1,292 with non-valvular AF were enrolled in the cohort. All patients were followed-up at least 6 months and cardiovascular events recorded. The end-point was defined as the first occurrence of stroke or death. The Cox analysis was adjusted on warfarin, antiplatelet and antiarrhythmic treatments at discharge.

Mean age was 73±14 years in valvular AF and 68±15 in non-valvular AF (p=0.0001). At baseline, in the valvular AF group CHA₂DS₂-VASc score were = 0 for 14 (5%) patients, = 1 for 28 (9%), ? 2 for 258 (86%). Non-valvular AF CHA₂DS₂-VASc scores were = 0 for 158 (12%), = 1 for 189 (15), ?2 for 945 (73%). The difference was statistically significant (p<0.0001). During a mean follow-up of 4.6±3.5 years, the patients with valvular AF experienced 154 (51%) and the patients with non-valvular AF experienced 409 (32%) strokes or deaths. The Kaplan-Meier curves (figure) show that patients with a CHA₂DS₂-VASc score ?2 were at higher risk of stroke or death. The adjusted Cox model, showed that valvular AF (HR, 1.57, 95%CI 1.30-1.89, p<0.0001) and a CHA₂DS₂-VASc score ?2 (HR, 5.30, 95%CI 3.77-7.45, p<0.0001) were predictors of risk of stroke or death.

These results suggest that a CHA₂DS₂-VASc score ?2 is associated with a higher risk of stroke and deaths, at mid-term follow-up, in patients with valvular AF (figure next page).