• atrial fibrillation;
  • brain natriuretic peptide;
  • in-hospital mortality;
  • ischaemic stroke;
  • transient ischaemic attack

Background and purpose:  Atrial fibrillation (AF) is the most powerful predictor of early death in patients with acute ischaemic stroke. We investigated whether the plasma brain natriuretic peptide (BNP) level on admission can serve as a biological marker of in-hospital death in acute ischaemic stroke and transient ischaemic attack (TIA) patients with AF.

Methods:  We prospectively enrolled ischaemic stroke and TIA patients with AF within 24 h of onset and measured plasma BNP on admission. Patients were divided into two groups: the deceased group, who died during hospitalization, and the survival group. The factors associated with in-hospital death were investigated by multivariate logistic regression analysis.

Results:  A total of 221 patients with AF were enrolled. Death occurred in 24 (10.9%) patients. The mean ± SD of the plasma BNP level of the deceased group was significantly higher than that of the survival group (714.1 ± 716.3 vs. 320.0 ± 380.7 pg/ml, < 0.0001). The optimal cutoff level, sensitivity, and specificity of BNP levels to distinguish the deceased group from the survival group were 320 pg/ml, 79.2, and 69.0%, respectively. Multivariate logistic regression analysis demonstrated that age per 10 years increase (OR, 3.56; 95% CI, 1.728–7.346, = 0.0006), internal carotid artery occlusion (OR, 10.20; 95% CI, 2.525–41.177, = 0.0011), NIHSS score of >17 (OR, 4.68; 95% CI, 1.137–19.286, = 0.0325), and plasma BNP level of >320 pg/ml (OR, 4.74; 95% CI, 1.260–17.800, = 0.0213) were independent factors associated with in-hospital death.

Conclusion:  The plasma BNP level on admission can predict in-hospital death in acute ischaemic stroke and TIA patients with AF.