COMPARING PREDICTIVE POWER OF DEATH BY AMBULATORY PULSE PRESSURE AND ITS COMPONENTS FOR 24 HOURS, DAYTIME AND NIGHTTIME IN PATIENTS AGED OVER 60 YEARS

Objective: Pulse pressure (PP), the systolic-minus-diastolic BP difference, has been previously shown to be a sum of two components: ‘elastic PP’ (elPP) and ‘stiffening PP’ (stPP) associated, respectively, with arterial stiffness at the diastolic pressure, and its variation during the early systole. Both components can be obtained from repeated BP readings. The study aims to investigate relationships between the predictive power for death of PP with that of its components in a large cohort of patients older than 60 years using 24-hour ambulatory BP monitoring (24hABPM).Design and method: We analyzed 24hABPM records of adequate quality provided by the Spanish Ambulatory Blood Pressure Registry of patients referred by their physicians. Outcomes were death from all causes or from cardiovascular disease. The estimates of associations between outcomes and the predictors PP, elPP and stPP (all divided by 5mmHg), were determined by utilizing Cox proportional regression analysis to two models: one included PP only, and the other both elPP and stPP that appeared independent. Analysis applied separately to 12 sub-models, including 3 time domains (24 hours, daytime and nighttime), 2 outcomes, with-or-without adjustment to age, sex(0,1), BMI, smoking status(0,1), diabetes(0,1), dyslipidemia(0,1), mean diastolic BP and pulse rate, SD(systolic BP), antihypertensive treatment(0,1), and previous cardiovascular disease(0,1). The hazard ratio (HR) per 5mmHg increase for each association was defined as exp(estimate). Results: Of the 21,048 patients included [mean(SD) age 70.7(6.9) years, 54% males] 4,714 deaths occurred during the follow-up period of 8.7(2.7) years. The combining of stPP and elPP in the same Cox model was justified for all time domains. Using the 12 sub-models PP-estimate founded close to the average between elPP-estimate and stPP-estimate, where the relative difference was 1.8%(3.8%), and the correlation between PP-estimate and the average between the elPP-estimate and stPP-estimate was rather high (see figure). It means that HR(PP) equals nearly SQR[HR(elPP)·HR(stPP)]. Conclusions: The estimate of the association of PP with the death-related outcomes was found practically equal to the average between the corresponding estimates of the PP components that reflect arterial stiffness and its variation during the early systole.