STAF score predict paroxysmal atrial fibrillation in patients with recurrent acute ischemic stroke / 国际脑血管病杂志

ABSTRACT

Objective:To investigate the predictive value of STAF (score for the targeting of atrial fibrillation) score for paroxysmal atrial fibrillation (PAF) in patients with recurrent acute ischemic stroke.Methods:Consecutive patients with recurrent acute ischemic stroke hospitalized in the Department of Neurology, Huizhou Municipal Central Hospital from January 2014 to January 2020 were enrolled retrospectively. All patients were divided into a sinus rhythm (SR) group and a PAF group. PAF was defined as no PAF during the last stroke hospitalization or after discharge, and PAF was found by routine ECG, ambulatory ECG or long-term ECG monitoring after this recurrent acute ischemic stroke. Multivariate logistic regression analysis was used to evaluate the independent related factors of PAF in patients with recurrent acute ischemic stroke. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of STAF for PAF in patients with recurrent acute ischemic stroke. Results:A total of 234 patients with recurrent acute ischemic stroke were enrolled, including 174 males (74.4%), aged 67.15±10.63 years, and 199 (85.0%) in the SR group and 35 (15.0%) in the PAF group. Univariate analysis showed that the proportion of patients in age >62 years old (94.3% vs. 63.8%; χ2=12.777, P<0.001) and left atrial enlargement (54.3% vs. 11.1%; χ2=40.379, P<0.001) of the PAF group was significantly higher than those of the SR group. Multivariate logistic regression analysis showed that age (odd ratio [ OR] 1.071, 95% confidence interval [ CI] 1.028-1.114; P=0.001) or age >62 years ( OR 17.512, 95% CI 2.881-106.453; P=0.002), left atrial enlargement ( OR 17.511, 95% CI 6.298-48.687; P<0.001), absence of vascular etiology ( OR 8.562, 95% CI 2.534-28.923; P=0.001), STAF score ( OR 2.715, 95% CI 1.969-3.744; P<0.001) and STAF score ≥5 ( OR 12.714, 95% CI 5.636-28.681; P<0.001) were independently associated with PAF. ROC curve analysis showed that the area under the curve of STAF for predicting PAF in patients with recurrent acute ischemic stroke was 0.848 (95% CI 0.779-0.917), and the sensitivity and specificity of STAF ≥5 for predicting PAF were 58.1% and 89.4%, respectively; the area under the curve for predicting PAF in patients with recurrent ischemic stroke of undetermined cause was 0.809 (95% CI 0.663-0.956), and the sensitivity and specificity of STAF ≥5 for predicting PAF were 53.8% and 91.3%, respectively. Conclusions:PAF is likely to exist in patients with recurrent acute ischemic stroke and aged >62 years, left atrial enlargement and absence of vascular etiology. STAF has medium predictive value for PAF in patients with recurrent acute ischemic stroke, but the sensitivity is not high.