Predictive value of repeated systolic blood pressure measurements for stroke risk. The Zutphen Study.

Abstract

The strength of the association between blood pressure and stroke incidence is dependent on the number of blood pressure measurements. Different summary variables of repeated blood pressure measurements taken during 10 years were evaluated in relation to the long-term risk of stroke in the Zutphen Study. During the period 1960-1970 repeated blood pressure measures were taken yearly in 603 men aged 50-69 years in 1970 in the town of Zutphen, The Netherlands. The individual average systolic blood pressure between 1960 and 1970, the predicted systolic blood pressure for 1970 (based on regression of blood pressure readings on time), and the single observed systolic blood pressure in 1970 were used as systolic blood pressure estimates. Their strength in predicting the 15-year stroke incidence was assessed using Cox proportional hazards models. Adjustment was made for the confounding effects of age, cigarette smoking, and serum total cholesterol. The average systolic blood pressure between 1960 and 1970 was the strongest predictor of 15-year stroke incidence. The strength of the association was underestimated by 55% when a casual systolic blood pressure measurement was used instead of 11 yearly measurements. It can be concluded that a casual blood pressure measurement leads to a substantial underestimation of the long-term stroke risk of an individual.