Fig. 2

Funnel plot comparing proportion versus the standard error of proportion for the outcome of (a) primary union and (b) secondary union. The assessment of the Egger test was coupled with an informal visual inspection of the funnel plot where circles represent studies included in the meta-analysis. The solid vertical line indicates no union in terms of proportion. The outer dashed line indicate the triangular region within which 95 % of studies are expected to lie in absence of both bias and heterogeneity (random effect pooled proportion ± 1.96 × standard error of pooled proportion). Asymmetry about the pooled proportion line is consistent with the presence of publication bias