Flow of selection for studies through review. BP, blood pressure; CIs, confidence intervals; RRs indicates relative risks.* Only the latest of the published duplicate studies from the same cohort was included if they offered the same outcome messages. However, one of these studies offered additional messages for subgroup analysis according to BP and gender , which could not be derived from the primary included study , so it was re-included when performing the subgroup analyses. † Data were derived from 18 prospective cohort studies (two articles were from the Strong Heart Study and reported the risk factors for CHD  and stroke , respectively).