“…Studies reporting smoking (n = 8 to be meta‐analyzed) 13,18,23,25,30,35‐37 showed high heterogeneity (weighted SSQ for error = 37.48; df = 6; P <0.001). As such, the regression analysis from the random‐effects model demonstrated that, even in the lack of statistical significance, there is a near‐trend significance ( P = 0.11).…”