Data extraction
Data from eligible studies were extracted in a predesigned Microsoft Office Excel 2007 form by two independent reviewers (I.A. and M.K.) and crosschecked for any disagreements, which were resolved after consensus with a senior (G.G.). In case of eligible studies that did not clearly report the primary outcome of interest, we contacted authors electronically (via email) to obtain the required data. For one study that provided effect size as median with interquartile range (IQR), median was used as mean and the standard deviation was obtained from IQR (standard deviation=IQR/1.35). In 10 out of 17 studies the endpoint was defined as AF recurrence, whereas in the other 6 studies the endpoint included atrial tachyarrhythmias (AF, atrial flutter, atrial tachycardia). Furthermore, in the majority of eligible studies (11/16) the LA-PLSsys was measured in sinus rhythm, whereas, accordingly to Hang et al.[9], peak and average systolic strain and strain rate showed no statistical difference between sinus rhythm and AF at baseline echocardiographic parameters. It should be noted that all strain measurements were performed before the catheter ablation procedure.