OBJECTIVE:

To evaluate which blood pressure measure is the best predictor of risk of total, ischemic, and hemorrhagic stroke.

METHODS:

The authors used a prospective cohort study among 11,466 men followed for incident stroke during a median of 19.4 years in the Physicians' Health Study. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were self-reported. They calculated relative risks (RRs) and 95% CIs for total, ischemic, and hemorrhagic stroke using Cox proportional hazards models. Model fit was compared using the chi(2) test statistic from likelihood ratio tests.

RESULTS:

During follow-up, 508 strokes occurred (411 ischemic, 89 hemorrhagic, and eight of unknown etiology). For each 10-mm Hg increase in SBP, the multivariable RRs were 1.31 (95% CI: 1.20 to 1.42) for total stroke, 1.28 (95% CI: 1.16 to 1.40) for ischemic stroke, and 1.38 (95% CI: 1.13 to 1.68) for hemorrhagic stroke. Although DBP, pulse pressure, and mean arterial pressure were all significant predictors of stroke risk, none was a significantly better predictor than SBP alone. Adding DBP did not significantly improve the model fit of SBP alone for any stroke type.

CONCLUSION:

In this large cohort of initially healthy men, systolic blood pressure was a consistent and significant predictor of total, ischemic, and hemorrhagic stroke. Systolic blood pressure alone was the only measure necessary to predict risk of total stroke or its major subtypes.