An accurate etiological classification is key to optimize secondary prevention after ischemic stroke, but the cause remains undetermined in one third of patients. Several studies pointed out the usefulness of circulating gene expression markers to discriminate cardioembolic (CE) strokes, mainly due to atrial fibrillation (AF), while only exploring them in small cohorts. A systematic review of studies analyzing high-throughput gene expression in blood samples to discriminate CE strokes was performed. Significantly dysregulated genes were considered as candidates, and a selection of them was validated by RT-qPCR in 100 patients with defined CE or atherothrombotic (LAA) stroke etiology. Longitudinal performance was evaluated in 12 patients at three time points. Their usefulness as biomarkers for AF was tested in 120 cryptogenic strokes and 100 individuals at high-risk for stroke. Three published studies plus three unpublished datasets were considered for candidate selection. Sixty-seven genes were found dysregulated in CE strokes. CREM, PELI1, and ZAK were verified to be up-regulated in CE vs LAA (p = 0.010, p = 0.003, p < 0.001, respectively), without changes in their expression within the first 24 h after stroke onset. The combined up-regulation of these three biomarkers increased the probability of suffering from CE stroke by 23-fold. In cryptogenic strokes with subsequent AF detection, PELI1 and CREM showed overexpression (p = 0.017, p = 0.059, respectively), whereas in high-risk asymptomatic populations, all three genes showed potential to detect AF (p = 0.007, p = 0.007, p = 0.015). The proved discriminatory capacity of these gene expression markers to detect cardioembolism even in cryptogenic strokes and asymptomatic high-risk populations might bring up their use as biomarkers.
Keywords: Atrial fibrillation; Biomarker; Cardioembolic; Etiology; Microarrays; Stroke.