Separating cardioembolic from large artery stroke has important treatment implications. We investigated whether echocardiography could improve Cardioembolic Stroke (CES) prediction compared with traditional measures and cholesterol biomarkers.

Data from 40 consecutive patients presenting with acute ischemic stroke which included brain and carotid imaging, ECG, echo, serum cholesterol and apolipoproteins were independently reviewed. Patients were classified into two groups: a) CES, defined by sustained or paroxysmal atrial fibrillation and <50% stenosis of a perfusing cerebral artery; b) Large artery stroke (LAS) defined as > 50% stenosis of an ipsilateral perfusing cerebral artery, with no evidence of AF on monitoring or evidence of small artery disease on neuroimaging and confirmed by an independent neurologist.

Other than the CES group being older, the baseline characteristics of the two groups were similar. Left Atrial Volume (indexed for body surface area, LAVi) was significantly larger in CES (57.9 +/- 19.4 vs 31.1 +/- 8.3ml/m2, p<0.01), with a simple equation that utilised age, LAVi and E wave accurately predicting 90% of CES. The difference in LAVi for CES was beyond that anticipated from the presence of AF alone. No differences in any of the lipid biomarkers were observed.

These finding indicate that LAVi is the most important predictor of CES due to atrial fibrillation and is highly predictive of patients with CES due to atrial fibrillation. Cholesterol biomarkers offered no additional discriminatory value.


atrial fibrillation, apolipoproteins, diagnostic techniques and procedures, echocardiography, ischemia, stroke

Find in your library