Aims To study the relation between visit-to-visit variability of blood pressure (BP) and cardiovascular risk in patients with stable coronary heart disease. Methods and results In 15 828 patients from the STABILITY trial (darapladib vs. placebo in patients with established coronary heart disease), BP variability was assessed by the standard deviation (SD) of systolic BP, the SD of diastolic BP, maximum BP, and minimum BP, from 5 measurements (baseline and months 1, 3, 6, and 12) during the first year after randomisation. Mean (SD) average BP during the first year of study was 131.0 (13.7) mmHg over 78.3 (8.3) mmHg. Mean (SD) of the visit-to-visit SD was 9.8 (4.8) mmHg for systolic and 6.3 (3.0) mmHg for diastolic BP. During the subsequent median follow-up of 2.6 years, 1010 patients met the primary endpoint, a composite of time to cardiovascular death, myocardial infarction, or stroke. In Cox regression models adjusted for average BP during first year of study, baseline vascular disease, treatment, renal function and cardiovascular risk factors, the primary endpoint was associated with SD of systolic BP (hazard ratio for highest vs. lowest tertile, 1.30, 95% CI 1.10-1.53, P = 0.007), and with SD of diastolic BP (hazard ratio for highest vs. lowest tertile, 1.38, 95% CI 1.18-1.62, P < 0.001). Peaks and troughs in BP were also independently associated with adverse events. Conclusion In patients with stable coronary heart disease, higher visit-to-visit variabilities of both systolic and diastolic BP are strong predictors of increased risk of cardiovascular events, independently of mean BP.