Outcome-Driven Thresholds for Increased Home Blood Pressure Variability

Abstract

Increased blood pressure (BP) variability predicts cardiovascular disease, but lack of operational thresholds limits its use in clinical practice. Our aim was to define outcome-driven thresholds for increased day-to-day home BP variability. We studied a population-based sample of 6238 individuals (mean age 60.0±12.9, 56.4% women) from Japan, Greece, and Finland. All participants self-measured their home BP on ≥3 days. We defined home BP variability as the coefficient of variation of the first morning BPs on 3 to 7 days. We assessed the association between systolic/diastolic BP variability (as a continuous variable and in deciles of coefficient of variation) and cardiovascular outcomes using Cox regression models adjusted for cohort and classical cardiovascular risk factors, including BP. During a follow-up of 9.3±3.6 years, 304 cardiovascular deaths and 715 cardiovascular events occurred. A 1 SD increase in systolic/diastolic home BP variability was associated with increased risk of cardiovascular mortality (hazard ratio, 1.17/1.22; 95% confidence interval, 1.06–1.30/1.11–1.34; P =0.003/<0.0001) and cardiovascular events (hazard ratio, 1.13/1.14; 95% confidence interval, 1.05–1.21/1.07–1.23; P =0.0007/0.0002). Compared with the average risk in the whole population, risk of cardiovascular deaths (hazard ratio, 1.66/1.84; 95% confidence interval, 1.27–2.17/1.42–2.37; P =0.0002/<0.0001) and events (hazard ratio, 1.46/1.42; 95% confidence interval, 1.21–1.76/1.17–1.71; P <0.0001/0.0004) was increased in the highest decile of systolic/diastolic BP variability (coefficient of variation>11.0/12.8). Increased home BP variability predicts cardiovascular outcomes in the general population. Individuals with a systolic/diastolic coefficient of variation of day-to-day home BP >11.0/12.8 may have an increased risk of cardiovascular disease. These findings could help physicians identify individuals who are at an increased cardiovascular disease risk.