High pulse pressure does not predict cardiovascular mortality in a general population of hypertensives

Abstract

Pulse pressure (PP) has been claimed to be superior to systolic blood pressure (SBP) and diastolic blood pressure (DBP) as a predictor of cardiovascular (CV) risk. In this study we assessed the hypothesis that high PP would predict CV mortality in a general population, and tested the predictive value of a novel expression of PP designed to overcome the confounding influence of the strong co-linearity between PP and SBP. A cohort of 15,406 men and women aged 45 – 64 years from a general population were screened for BP and other cardiovascular risk factors in the 1970s and followed-up for dates and causes of death by linkage with the Registrar General (Scotland). A measure of PP not confounded by the strong relationship with SBP (r2 = 0.84) was derived for each individual (conditional PP). Conditional PP = actual PP - expected PP (average PP in population with the same DBP, adjusted for age and gender). Conditional PP was tested for statistical independence from SBP and then included in a Cox proportional hazard model along with other BP measures and risk factors as a predictor of CV risk. By 2001, 8,192 (53%) individuals were dead, 51% from CV causes: 32% coronary artery disease (CAD), 12% stroke. After adjustment for age, smoking and serum cholesterol, SBP, DBP and PP were each significant (p < 0.001) single predictors for CAD, stroke and all cause mortality, in males and in females. Except for stroke mortality in females where DBP was the best predictor, SBP was most strongly correlated with risk. In all cases, PP was by far the least informative variable. For CAD mortality, this rank order of risk prediction was consistent across the age range. After adjustment for confounding factors including SBP, conditional PP was inversely associated with risk; low conditional PP was associated with highest risk, while high conditional PP at entry predicted the lowest risk. The effect of the low conditional PP as a risk factor appeared to be consistent for different cause specific-outcomes and across the range of SBP, and to be independent of age. In conclusion, a high PP in middle-aged individuals does not predict cardiovascular mortality and SBP is the most consistent and reliable predictor.