No longitudinal study addressed whether systolic blood pressure level (SBPL) or within-visit variability (SBPV) predict arterial properties or vice versa. In families randomly recruited from a Flemish population, we determined SBPL and SBPV from five consecutive blood pressure readings. The indexes of SBPV were variability independent of the mean, the difference between maximum and minimum SBPL, and average real variability. We measured carotid intima-media thickness and distensibility by ultrasound and carotid-femoral pulse wave velocity by tonometry (SphygmoCor, version 8.2). Effect sizes were computed for 1-s.d. increments in the predictors, while accounting for covariables and family clusters. Among 1087 participants (50.4% women; mean age, 41.8 years), followed up for 2.55 years (median), higher SBPL predicted (P < or = 0.019) higher carotid intima-media thickness (+15 μm), lower carotid distensibility (-1.53 10(-3) kPa(-1)) and faster carotid-femoral pulse wave velocity (+0.285 m s(-1)) at follow-up, whereas none of the SBPV indexes predicted the arterial traits at follow-up (P> or = 0.11). In a subset of 713 participants, followed up for another 3.14 years, lower carotid distensibility predicted (P<0.01) higher SBPL (+2.57 mm Hg), variability independent of the mean (+0.531 units), difference between maximum and minimum SBPL (+1.75 mm Hg) and average real variability (+0.654 mm Hg). Higher carotid-femoral pulse wave velocity predicted a 1.11 mm Hg increase SBPL (P=0.031). In conclusion, temporality and effect size suggest that SBPL but not within-visit SBPV cause arterial stiffness and carotid intima-media thickness. Carotid stiffness, independent of SBPL, predicts within-visit SBPV, possibly because baroreflexes originating from a stiff carotid artery wall are impaired. Finally, stiffness of the aorta contributes to the age-related SBPL possibly, because faster returning reflected waves augments SBPL.