Variation between studies in reported relative risks associated with hypertension: Time trends and other explanatory variables

Objectives. This study examined the variation in reported relative risks of coronary heart disease and stroke associated with hypertension and determined reasons for the interstudy variation. Methods. Studies published since 1970 were examined that reported the absolute number of events and person-years for men by age and hypertensive status. The data were pooled in Poisson regression models with the coronary heart disease or stroke rate as the dependent variable. Independent variables were hypertensive status, age at entry, age of study, duration of follow-up, diastolic blood pressure cutoff point, and interactions of all these variables with hypertensive status. Results. The reported relative risks associated with hypertension ranged from 1.45 to 2.77 for coronary heart disease and from 1.86 to 5.78 for stroke. Smaller relative risks were found in more recent studies, in studies with long followup, and in studies using a lower cut-off point to define hypertension. Conclusions. Part of the interstudy variation in relative risks associated with hypertension can be attributed to differences in duration of follow-up, blood pressure cutoff point, and time at which blood pressure was measured, suggesting declining relative risks over time.