Amyloid beta (Abeta) may disturb cerebral autoregulation by damaging the wall of small cerebral blood vessels and by direct negative vasoactive properties. We assessed whether previous and concurrent plasma Abeta(1-40) and Abeta(1-42) levels were associated with an impaired CO2-induced cerebral vasomotor response. In the longitudinal population-based Rotterdam Study we measured plasma Abeta levels and cerebral vasomotor reactivity to hypercapnia with transcranial Doppler ultrasonography (TCD) in 441 people, aged 60-90 years. We performed age and sex adjusted logistic regression analysis. Plasma Abeta levels assessed on average 6.5-year before TCD were linearly associated with an impaired CO2-induced cerebral vasomotor response (odds ratio 1.48 (95%CI 1.19;1.84) per standard deviation increase in Abeta(1-40), and 1.36 (95%CI 1.09;1.70) per standard deviation increase in Abeta(1-42)). Such an association was not present for Abeta assessed concurrently with the TCD measurement. Persons whose plasma Abeta(1-40) levels had decreased in the 6.5-year period preceding TCD measurements were more likely to have an impaired CO2-induced vasomotor reactivity. Overall our observations are most compatible with plasma Abeta levels representing vascular Abeta deposits years later resulting in impaired CO2-induced vasomotor reactivity.