Predicting the outcomes of acute ischaemic stroke in atrial fibrillation: the role of baseline CHADS2, CHA2DS2-VASC and HAS-BLED score values

Objective Atrial fibrillation (AF)-related risk of stroke is commonly assessed using the CHADS(2) or CHA(2)DS(2)-VASc score, whilst the oral anticoagulation-related bleeding risk can be estimated by the HAS-BLED score. We investigated the association of these scores with outcomes of AF-associated strokes, defined as symptomatic intracranial haemorrhage (sICH), favourable outcome (modified Rankin Scale [mRS] 0-2) or death. Methods Analyses of prospective data on stroke patients with non-valvular AF treated in the Stroke Unit from January 2009 to June 2012 were performed. Results Of 787 patients with an acute ischaemic stroke, 131 had AF (16.6%, median age 70, range 62-76 years). Of those, 6 patients (4.6%) had sICH, 49 (37.4%) died and 55 (42.0%) had a favourable 90-day outcome. HAS-BLED score of >= 3 was associated with sICH both in the univariate analysis (OR 15.13; 95% Cl 2.11-108.25, P = 0.007) and in the multivariable model (OR 19.96; 95% Cl 2.23-178.81, P = 0.007), which also included the baseline neurological deficit score (NIHSS), intravenous thrombolysis or the use of antiplatelet/anticoagulant therapy. The CHADS(2) and CHA(2)DS(2)-VASc scores were associated with 90-day mortality in the univariate analyses (OR 1.47; 95% Cl 1.11-1.95 and OR 1.36; 95% Cl 1.08-1.69, respectively, both P = 0.008). The CHA(2)DS(2)-VASc score was inversely related to the favourable outcome in the univariate analysis (OR 0.80; 95% Cl 0.65-0.99, P = 0.042). Conclusion HAS-BLED was found to have an independent predictive value on the occurrence of sICH regardless of the treatment (thrombolysis or conservative therapy). A trend toward statistical relation to the influence of the CHA2DS2-VASc values on the favourable outcome was registered.