BACKGROUND For patients undergoing catheter ablation of atrial fibrillation (AF), left atrial size is a predictor of recurrence of AF during follow-up. For this reason, major clinical trials have used a left atrial diameter (LAD) of more than 5.0 or 5.5 cm, assessed by echocardiography, as an exclusion criterion for patients deemed candidates for ablation of AF. However, whether LAD accurately reflects true left atrial size has not been systematically investigated. Therefore, the purpose of this study was to test the hypothesis that LAD, measured by echocardiography, accurately correlates to left atrial volume measured by computed tomography (CT). RESULTS We included 50 patients (mean age 56 +/- 12 years, five female) with symptomatic AF (40% paroxysmal, 60% persistent), referred for catheter ablation. In each patient, transthoracic echocardiography was performed. Additionally, all patients underwent CT using a 64-slice CT scanner. Left atrial volume was calculated by manually tracing left atrial area on each CT cross-sectional image. Patients had a mean LAD measured by echocardiography of 4.5 +/- 0.7 cm, ranging from 2.9 to 5.7 cm. Left atrial volume measured by CT ranged from 67 mL to 270 mL with a mean value of 146 +/- 49 mL. A poor correlation was noted between LAD and left atrial volume, r = 0.49 (P < 0.001). CONCLUSIONS LAD measured by echocardiography correlates poorly with left atrial volume measured by CT in patients with AF. As a result, selecting patients with AF for treatment with catheter ablation should not be based on an echocardiographic-derived LAD alone.