Relationship of Age With Stroke and Death in Anticoagulated Patients With Nonvalvular Atrial Fibrillation: AMADEUS Trial.

BACKGROUND AND PURPOSE The prevalence of atrial fibrillation increases with age, but age-specific data on the incidence of stroke and death in anticoagulated patients with atrial fibrillation are more limited, particularly with regard to comparisons of relative risks of clinical outcomes between the different age strata in relation to quality of anticoagulation control among warfarin users. METHODS We investigated the incidence of adverse outcomes between tertiles of age groups (age, <67 [n=722]; age, 67-74 [n=747]; and age, ≥75 [n=824]) in 2293 patients with atrial fibrillation participating in warfarin arm in the AMADEUS (Evaluating the Use of SR34006 Compared to Warfarin or Acenocoumarol in Patients With Atrial Fibrillation) trial. The average time in therapeutic range was calculated as a measure of anticoagulation control and related to clinical outcomes. RESULTS Absolute rates for stroke/systemic embolism (SSE), cardiovascular death, or any clinically relevant bleeding increased with increasing age strata. The combined end point of cardiovascular death and SSE was the highest in the top tertile (adjusted hazard ratio, 2.63; 95% confidence interval, 1.23-5.63) compared with the middle and lowest tertiles (P for trend=0.01). For bleeding, there was no significant difference in relative risks between the age strata (P for trend=0.55 in the warfarin group and in the warfarin group with time in therapeutic range≥60%, P for trend=0.60). The quality of anticoagulation control (time in therapeutic range) significantly correlated with any clinically relevant bleeding (r=-0.91; P<0.001) and cardiovascular death/SSE rates (r=-0.76; P=0.01). CONCLUSIONS Elderly patients with atrial fibrillation have higher absolute risks of cardiovascular death, SSE, and bleeding, but relative risks of clinically relevant bleeding are not significantly different with increasing age strata. A significant inverse relationship between time in therapeutic range and bleeding and cardiovascular death/SSE emphasizes the importance of good quality anticoagulation control.