Publication bias
No obvious publication bias was found on the CYP1A1 A2455G polymorphism
with CRC risk. However, a significant publication bias was observed on
the CYP1A1 T3801C (C vs. T: P = 0.013) polymorphism with CRC
risk. Figure 2 shows the non-parametric ‘trim and fill’ method funnel
plot. The result did not revised in overall analysis (C vs. T: OR =
0.94, 95% CI: 0.84–1.04) indicating add studies can not affect the
merging result.