Score for the targeting of atrial fibrillation: a new approach to diagnosing paroxysmal atrial fibrillation.

Background And Purpose: Detecting paroxysmal atrial fibrillation (PAF) after ischaemic stroke is challenging. There are several methods to increase the detection rate of PAF, but it is first necessary to identify subgroups of patients at risk. In a previous study, we established a clinicoradiologic score that predicts atrial fibrillation (AF) in stroke patients. The purpose of the present study is to validate this score specifically for PAF patients.

Methods: 500 consecutive ischaemic stroke patients were examined in our stroke unit. A blind evaluation of the STAF (score for the targeting of atrial fibrillation) was performed for each patient with or without AF. Firstly, we established the reproducibility of the STAF performance by comparing areas under the receiver operating characteristic curves in the preliminary and present studies. Secondly, to validate the predictive value of the STAF in occult AF, areas under the receiver operating characteristic curves were compared for each type of AF. Thirdly, the best threshold value was calculated.

Results: AF was detected in 145 cases including 45% of paroxysmal forms. There is no significant score performance difference (p = 0.192) between the preliminary and prospective cohort areas under the receiver operating characteristic curves. This confirms the reproducibility of STAF performance. The area under the receiver operating characteristic curve for the PAF group was 0.907 versus 0.911 for the permanent AF group (p = 0.906). The diagnostic value of the STAF is as good in permanent as paroxysmal AF. In PAF, a STAF ≥ 5 has a sensitivity of 91% and a specificity of 77%.

Conclusions: Due to its reproducibility and predictive value, the STAF can be used by neurologists as part of a novel diagnostic strategy for occult AF.