Statistical Analysis

Statistical analysis was performed using the GraphPad Instat computer software (version 3.06 for Windows, GraphPad Software Inc., La Jolla, CA, USA). All results were presented as mean ± SD values. Normal distribution of quantitative variables was preliminary tested using the Kolmogorov–Smirnov test. Patients were dichotomized into two groups according to the severity of neurological deficit evaluated by NIHSS score: mild ischemic stroke was defined as presenting NIHSS score of 0–5 and severe ischemic stroke as presenting NIHSS score of 6–41. We also performed a sample size and power analysis to define the number of subjects to include in the ESUS cohort: a sample size of 226 was estimated considering a chi-squared test, with alpha of 0.05, power of 0.95 and a medium effect size of 0.25.

Mann-Whitney test was performed to compare age, electrocardiographic and echocardiographic markers of atrial cardiopathy in patients with mild stroke and severe stroke. The two-sided Fisher's exact test was performed to evaluate statistical correlation between categorical variables in both groups of patients. To determine the factors associated with severe ESUS a multivariate logistic regression was performed (with the stepwise procedure based on Akaike's criterion) and corresponding Odds Ratio (OR) and their confidence intervals (CI) were calculated. A P value below 0.05 was considered statistically significant.