BACKGROUND Current guidelines make no outcome-based recommendations on the optimal measurement schedule for home blood pressure (BP).

METHODS We enrolled 4,802 randomly recruited participants from three populations. The participants were classified by their (i) cross-classification according to office and home BP (normotension, masked hypertension, white-coat hypertension, and sustained hypertension) and (ii) home BP level (normal BP, high normal BP, grade 1 and 2 hypertension), while the number of home measurement days was increased from 1 to 7. The prognostic accuracy of home BP with an increasing number of home BP measurement days was also assessed by multivariable-adjusted Cox models.

RESULTS Agreement in classification between consecutive measurement days indicated near perfect agreement (κ ≥ 0.9) after the sixth measurement day for both office and home BP cross-classification (97.8% maintained classification, κ = 0.97) and home BP level (93.6% maintained classification, κ = 0.91). Over a follow-up of 8.3 years, 568 participants experienced a cardiovascular event, and the first home BP measurement alone predicted events significantly (P ≤ 0.003). The confidence intervals (CIs) were too wide and overlapping to show superiority of multiple measurement days over the first measurement day (hazard ratios per 10mm Hg increase in systolic BP at initial day, 1.11 [CI 1.07–1.16]; that at 1–7 days, 1.18 [CI 1.12–1.24]). Masked hypertension, but not white-coat hypertension, was associated with increased cardiovascular risk, irrespective of the number of home measurement days.

CONCLUSION Even a single home BP measurement is a potent predictor of cardiovascular events, whereas seven home measurement days may be needed to reliably diagnose hypertension.