Differential influence of distinct components of increased blood pressure on cardiovascular outcomes: from the atherosclerosis risk in communities study

Abstract

Elevation in blood pressure (BP) increases risk for all cardiovascular events. Nevertheless, the extent to which different indices of BP elevation may be associated to varying degrees with different cardiovascular outcomes remains unclear. We studied 13340 participants (aged 54 ± 6 years, 56% women and 27% black) of the Atherosclerosis Risk in Communities Study who were free of baseline cardiovascular disease. We used Cox proportional hazards models to compare the relative contributions of systolic BP, diastolic BP, pulse pressure, and mean arterial pressure to risk for coronary heart disease, heart failure, stroke, and all-cause mortality. For each multivariable-adjusted model, the largest area under the receiver-operating curve (AUC) and smallest -2 log-likelihood values were used to identify BP measures with the greatest contribution to risk prediction for each outcome. A total of 2095 coronary heart disease events, 1669 heart failure events, 771 stroke events, and 3016 deaths occurred during 18 ± 5 years of follow-up. In multivariable analyses adjusting for traditional cardiovascular risk factors, the BP measures with the greatest risk contributions were the following: systolic BP for coronary heart disease (AUC=0.74); pulse pressure for heart failure (AUC=0.79); systolic BP for stroke (AUC=0.74); and pulse pressure for all-cause mortality (AUC=0.74). With few exceptions, results were similar in analyses stratified by age, sex, and race. Our data indicate that distinct BP components contribute variably to risk for different cardiovascular outcomes.

Keywords: blood pressure; cardiovascular diseases; epidemiology; hypertension; outcomes assessment.

Figure

Twenty-year risk estimates for each of the main outcomes are plotted for participants (denoted by black dots) across varying levels of systolic blood pressure (SBP) and pulse pressure (PP). Actual risk estimates are denoted by labeled contour lines, and the gradient of relatively lower versus higher risk for a given outcome is denoted by white (lowest), yellow to orange (moderate), and red (highest) background colors. As shown in Panel A, risk for coronary heart disease increased steadily with elevations in both SBP and PP. Panel B shows heart failure risk increasing primarily in association with rising PP, as denoted by more vertical contour lines representing risk estimates. Panel C shows risk for stroke increasing primarily in association with increments of SBP, as denoted by more horizontal contour lines in the upper range of SBP values. Panel D shows risk for all-cause mortality increasing in association with PP, as denoted by more vertical contour lines particularly in the upper range of PP values.