A multicenter study of transient global amnesia for the better detection of magnetic resonance imaging abnormalities

Background and purpose The detection rate of diffusion-weighted (DWI) hyperintense lesions varies widely in patients with transient global amnesia (TGA). The aim was to examine the association of hyperintense lesions on DWI magnetic resonance imaging (MRI) with patient characteristics, precipitating factors, clinical presentation and MRI settings in patients with TGA. Methods In this multicenter retrospective observational study, using the standardized diagnosis entry system of electronic health records of four tertiary medical centers in the Kansai district of Japan, TGA patients (n = 261) who underwent brain MRI within 28 days of onset were examined. When the onset time was unavailable, the discovery time was used. Results Diffusion-weighted hyperintense lesions were observed in 79 patients (30%). There were no significant differences in age, sex, vascular risk factors, precipitating factors or clinical presentation between patients with and without DWI lesions. The detection rate increased linearly 24 h after onset and then reached a plateau of 60%-80% by 84 h. After 84 h, the detection rate decreased rapidly. In a multivariate logistic regression model, MRI examination 24-84 h after onset (odds ratio 7.00, 95% confidence interval 3.50-13.99) and a thin-slice (<= 3 mm) DWI sequence (odds ratio 7.59, 95% confidence interval 3.05-18.88) were independent predictors of DWI lesions. Conclusions This study suggests that DWI hyperintense lesions in TGA are not associated with patient characteristics and clinical presentation. Brain MRI examination 24-84 h after onset and thin-slice DWI sequences enhance the detection of DWI lesions in TGA patients.