Xuxi Zhang

Publication Bias According to Supplementary Figure 2, we found a little publication bias because of P = 0.045. Therefore, we performed Trim and fill method (Supplementary Figure 3) to test the bias again, and the results showed that the results were steady and there was no publication bias. Combined the two results of publication bias, we got a conclusion that although there was a little publication bias among our 20 studies, the bias was not serious and could be accepted. Supplementary Figure 2 Egger’s and Begg’s Test Supplementary Figure 3 Trim and fill method 158 Chapter 6