Atrial fibrillation detection in cryptogenic stroke and a prediction score using imaging examination.

Background: Paroxysmal atrial fibrillation (AF) is considered one of the causes of cryptogenic stroke (CS). Early detection of AF is crucial for secondary stroke prevention. We aimed to identify the incidence of AF and characteristics of patients with CS using an insertable cardiac monitor (ICM), magnetic resonance imaging (MRI), and transesophageal echocardiography.

Methods: This multicenter, prospective, observational study enrolled patients with CS from 24 hospitals in Japan. The study recruitment spanned from February 2020 to July 2021. Following the implantation of ICM, we investigated the first AF detection within the initial 6, 12, and 24 months of follow-up and compared patient clinical characteristics, including MRI findings and transesophageal echocardiography data between patients with AF (AF group) and without AF (non-AF group) at baseline and after ICM implantation. Additionally, we examined predictors of AF to develop a predictive score for AF detection.

Results: Of the 202 patients, AF was detected in 26.2 % of patients at 24 months, and 7.4 % of patients experienced a stroke recurrence. The incidence of hyperlipidemia was lower in the AF group compared to the non-AF group. The AF group had more large-vessel occlusion and an elevated cardiothoracic ratio compared to the non-AF group. A risk-scoring system, the LOOK-AF score, was constructed using these four clinical variables.

Conclusions: Our study found AF in over a quarter of patients with CS in two years. The LOOK-AF score measured at admission can be used to predict AF in patients with an ICM.

Trial Registration: UMIN Clinical Trial Registry, UMIN000039809.