Pooled measures on the relationship of eNOS gene 4b/a, T-786C and G894T polymorphism with ischemic stroke.

DHWE: deviated from Hardy–Weinberg equilibrium in controls.
aTwo articles (Kim et al. Meng et al.) in 4b/a and T-786C and one article (Di et al.) in G894T for recessive model were not sufficient to calculated pooled OR.
bOne article (Meng et al.) in 4b/a and T-786C and one article (Di et al.) in G894T for recessive model were not sufficient to calculated pooled OR.
cp-values was for observed pooled odds ratios. when I2≤50%, it was for FEM, otherwise it was for REM.
*Pooled ORs were not calculated for only two articles left sufficient to calculate pooled ORs after excluding articles with OR>3.0.
FEM, fixed effect model; REM, random effect model
Dominant model: aa + ba vs. bb for 4b/a, CC + TC vs. TT for T-786C, and TT + GT vs. GG for G894T.
Recessive model: aa vs. ba + bb for 4b/a, CC vs. TC + TT for T-786C, and TT vs. GT + GG for G894T.
Codominant model: a vs. b for 4b/a, C vs. T for T-786C, and for T vs. G for G894T.