We thank Zheng et al. for their interest in our publication.1 We agree that the effect of the risk factor of smoking for nonlobar ICH vs lobar ICH might be underestimated. Heterogeneity is indeed considerable in the direct comparison of nonlobar ICH vs lobar ICH. If the study by Inagawa et al. (e-Ref #15)—not Booth et al. as mentioned in the comment—is excluded from the analysis, the pooled RR for nonlobar vs lobar ICH is 1.32 (1.09–1.59; I2 = 21%; p value for heterogeneity = 0.27). However, that study fulfilled the inclusion criteria and hence was included. In the comparison of nonlobar ICH vs controls, we found no support for smoking as a risk factor for non-lobar ICH (pooled RR 1.09 [0.92–1.31]), also not when we excluded Inagawa’s study (pooled RR 1.11 [0.91–1.37]). As we discussed, the included studies used different definitions of smoking—never, current or past—instead of burden of smoking, expressed in number of pack years; this may have contributed to the heterogeneous results.