P797Association of office and ambulatory blood pressure with mortality and cardiovascular endpoints

Background Blood pressure (BP) is the strongest modifiable risk factor predicting death and cardiovascular complications. However, which BP index has greater clinical relevance is not established. Purpose To compare associations of adverse health outcomes with commonly used BP indexes. Methods Longitudinal population-based cohort study of 11,135 adults recruited in Europe, Asia and South America with baseline observations collected from 1985 until 2010 and last follow-up year ranging from 2006 to 2016. Multivariable-adjusted analyses addressed the associations of primary endpoints (total mortality and composite CV endpoints) and secondary endpoints (CV mortality, cardiac and coronary endpoints and stroke) with conventional BP, automated office BP and ambulatory BP indexes. Model performance was assessed by the integrated discrimination improvement (IDI) and change in the area under the curve (ΔAUC). Results The study included 49.3% women. Median age was 54.7 years. Over 13.8 years (median), 2836 participants died (18.5 per 1000 person-years) and 2049 (13.4 per 1000 person-years) experienced a cardiovascular endpoint. Both endpoints were related to all single systolic BP (SBP) indexes (P<0.001). In models that included both 24-h and nighttime SBP, the HRs for total mortality and the cardiovascular endpoint in relation to 24-h SBP were 0.98 (95% confidence interval, 0.92–1.06) and 1.17 (1.08–1.28) and HRs associated with nighttime SBP were 1.18 (1.11–1.27) and 1.12 (1.04–1.21), respectively. Otherwise, depending on the SBP index combined with 24-h or nighttime SBP, HRs for both endpoints associated with 24-h or nighttime SBP retained significance (P≤0.004), ranging from 1.12 (1.07–1.17) to 1.56 (1.41–1.74). In models that included either 24-h or nighttime SBP, HRs associated with other SBP indexes weakened, lost significance or reversed. IDI and ΔAUC for adding 24-h or nighttime SBP to other SBP indexes were consistently greater than unity except for combining 24-h and nighttime SBP. These findings were consistent for secondary endpoints, for systolic and diastolic BP, and in sensitivity analyses performed to evaluate the influence of antihypertensive drug treatment at baseline, various definitions of day and night, and the contribution of different cohorts to the overall results. Conclusions Associations of outcomes were strongest for 24-h and nighttime ambulatory BP, which outperformed all other BP indexes in predicting 10-year absolute risk and which should therefore be prioritised in clinical practice. Acknowledgement/Funding The European Union, the European Research Council, the European Research Area Net for Cardiovascular Diseases, et al.