3.3.2 Contrast volume
All studies selected involving a total of six eligible studies reporting the contrast volume[6-8, 10, 13, 14]. The pooled results indicated that compared with the TEE procedure, the ICE procedure showed no significant difference (WMD -2.61; 95%CI: -12.25, 7.02; P =0.595; I²=84.80%) (Figure 6). The subgroup analysis showed that in the paroxysmal AF <50% subgroup, the ICE group’s contrast volume was significantly decreased compared with the TEE group [WMD -15.02; 95%CI: -27.08, -3.07; P = 0.015; I² = 78.60%]. Moreover, in the hypertension proportion < 90% subgroup, the contrast volume in the ICE group was much lower than that in the TEE group [WMD-12.95; 95%CI: -22.83, -3.07; P = 0.010; I²=62.90%] (Supplementary Table 3) .
Meanwhile, sensitivity analysis showed that no single study dominated the combined proportion and heterogeneity, ranging from -4.81 (95% CI: -15.29, 5.66) to 1.26 (95% CI: -8.55, 11.06). Moreover, Egger’s test was performed and result showed no publication bias (P=0.371). which indicated that the results were considered to be robust.
3.3.3Fluoroscopic time
A total of ten eligible studies reported the fluoroscopic time and the pooled result showed that the fluoroscopic time guided by ICE was significantly equivalent to that guided by TEE (WMD -0.34; 95%CI: −2.09, 1.41; P =0.705; I²=82.80%) (Figure 7 )[6-15]. Subgroup analysis was performed with a total of 8 subgroup factors for the fluoroscopic time, and the results are displayed inSupplementary Table 4 . No significant change was detected in the overall combined proportion by sensitivity analysis, ranging from-0.72 (95% CI: -2.48, 1.03) to 0.07 (95% CI: -1.74, 1.87). Moreover, no publication bias was shown in Egger’s test (P =0.941).