Day-by-Day Variability of Home Blood Pressure and Incident Cardiovascular Disease in Clinical Practice

We assessed the relationship between day-by-day home blood pressure (BP) variability and incident cardiovascular disease (CVD) in clinical practice. J-HOP study (Japan Morning Surge-Home Blood Pressure) participants underwent home BP monitoring in the morning and evening for a 14-day period, and their BP levels and BP variability independent of the mean (VIM) were assessed. Incident CVD events included coronary heart disease and stroke. Cox models were fitted to assess the home BP variability-CVD risk association. Among 4231 participants (mean±SD age, 64.9±10.9 years; 53.3% women; 79.1% taking antihypertensive medication), mean (SD) home systolic BP (SBP) levels over time and VIMSBP were 134.2 (14.3) and 6.8 (2.5) mm Hg, respectively. During a 4-year follow-up period (16 750.3 person-years), 148 CVD events occurred. VIMSBP was associated with CVD risk (hazard ratio per 1-SD increase, 1.32; 95% confidence interval [CI], 1.15-1.52), independently of mean home SBP levels over time and circulating B-type natriuretic peptide levels or urine albumin-to-creatinine ratio. Adding VIMSBP to the CVD prediction model improved the discrimination (C statistic, 0.785 versus 0.770; C statistic difference, 0.015; 95% CI, 0.003-0.028). Changes in continuous net reclassification improvement (0.259; 95% CI, 0.052-0.537), absolute integrated discrimination improvement (0.010; 95% CI, 0.003-0.016), and relative integrated discrimination improvement (0.104; 95% CI, 0.037-0.166) were also observed with the addition of VIMSBP to the CVD prediction models. In addition to the assessments of mean home SBP levels and cardiovascular end-organ damage, home BP variability measurements may provide a clinically useful distinction between high- and low-risk groups among Japanese outpatients.