Results:
A total of twenty eligible studies with 3,610 atrial fibrillation patients (1,564 patients for ICE and 2,046 patients for TEE) were enrolled. Compared with TEE group, there was no significant difference in procedural success rate (RR=1.01; 95% CI: 1.00,1.02; P =0.171; I²=0.00%), total procedural time [weighted mean difference (WMD) = -5.58; 95%CI: -15.97, 4.81; P =0.292; I²=96.40%], contrast volume (WMD=-2.61; 95%CI: -12.25, 7.02; P =0.595; I²=84.80%), and fluoroscopic time (WMD=-0.34; 95%CI: −2.09, 1.41; P =0.705; I²=82.80%) in the ICE group. Subgroup analysis revealed ICE showed less contrast use than TEE in the lower proportion paroxysmal atrial fibrillation group and lower proportion hypertension.