Publication bias checked by Egger’s regression test, showed that it may have a substantial impact on total prevalence estimate (Egger; bias: 7.0, P < 0.0001) (Figure 3).