Added predictive value of night-time blood pressure variability for cardiovascular events and mortality: the Ambulatory Blood Pressure-International Study.

The association of ambulatory blood pressure (BP) variability with mortality and cardiovascular events is controversial. To investigate whether BP variability predicts cardiovascular events and mortality in hypertension, we analyzed 7112 untreated hypertensive participants (3996 men) aged 52±15 years enrolled in 6 prospective studies. Median follow-up was 5.5 years. SD of night-time BP was positively associated with age, body mass index, smoking, diabetes mellitus, and average night-time BP (all P<0.001). In a multivariable Cox model, night-time BP variability was an independent predictor of all-cause mortality (systolic, P<0.001/diastolic, P<0.0001), cardiovascular mortality (P=0.008/<0.0001), and cardiovascular events (P<0.001/<0.0001). In contrast, daytime BP variability was not an independent predictor of outcomes in any model. In fully adjusted models, a night-time systolic BP SD of ≥12.2 mm Hg was associated with a 41% greater risk of cardiovascular events, a 55% greater risk of cardiovascular death, and a 59% increased risk of all-cause mortality compared with an SD of <12.2 mm Hg. The corresponding values for a diastolic BP SD of ≥7.9 mm Hg were 48%, 132%, and 77%. The addition of night-time BP variability to fully adjusted models had a significant impact on risk reclassification and integrated discrimination for all outcomes (relative integrated discrimination improvement for systolic BP variability: 9% cardiovascular events, 14.5% all-cause death, 8.5% cardiovascular death, and for diastolic BP variability: 10% cardiovascular events, 19.1% all-cause death, 23% cardiovascular death, all P<0.01). Thus, addition of BP variability to models of long-term outcomes improved the ability to stratify appropriately patients with hypertension among risk categories defined by standard clinical and laboratory variables.