Editors' Note: Location-Specific Risk Factors for Intracerebral Hemorrhage: Systematic Review and Meta-Analysis.

In their meta-analysis of 42 studies of patients with intracerebral hemorrhage (ICH, n = 26,174 patients), Jolink et al. report on the updated risk factors associated with lobar and nonlobar ICH. Hypertension remains an independent predictor of ICH, irrespective of location, whereas Black race, Hispanic ethnicity, male sex, diabetes, alcohol use, and underweight body habitus were also associated with nonlobar ICH. Zheng et al. comment that the effect of tobacco use may be underestimated in this meta-analysis and furthermore that the published data on tobacco use meet the Cochrane criteria for heterogeneity and thus may make their assessment on this issue unreliable. The readers also comment that the exclusion of one study from this meta-analysis would reduce the heterogeneity of included studies and yield a pooled effect estimate that indicates tobacco use is a statistically significant predictor of nonlobar ICH. The study investigators agree that the relationship between tobacco use and ICH may be underestimated, but they cannot exclude a study which meets criteria for inclusion. The heterogeneity of findings regarding tobacco use may be explained by varying definitions of tobacco use (never, remote, or current) and by differing amounts of use, which may contribute to a higher risk of ICH. However, based on the available data, the investigators conclude that there is insufficient evidence to support that tobacco use is associated with nonlobar ICH. In their meta-analysis of 42 studies of patients with intracerebral hemorrhage (ICH, n = 26,174 patients), Jolink et al. report on the updated risk factors associated with lobar and nonlobar ICH. Hypertension remains an independent predictor of ICH, irrespective of location, whereas Black race, Hispanic ethnicity, male sex, diabetes, alcohol use, and underweight body habitus were also associated with nonlobar ICH. Zheng et al. comment that the effect of tobacco use may be underestimated in this meta-analysis and furthermore that the published data on tobacco use meet the Cochrane criteria for heterogeneity and thus may make their assessment on this issue unreliable. The readers also comment that the exclusion of one study from this meta-analysis would reduce the heterogeneity of included studies and yield a pooled effect estimate that indicates tobacco use is a statistically significant predictor of nonlobar ICH. The study investigators agree that the relationship between tobacco use and ICH may be underestimated, but they cannot exclude a study which meets criteria for inclusion. The heterogeneity of findings regarding tobacco use may be explained by varying definitions of tobacco use (never, remote, or current) and by differing amounts of use, which may contribute to a higher risk of ICH. However, based on the available data, the investigators conclude that there is insufficient evidence to support that tobacco use is associated with nonlobar ICH.