Cerebral amyloid angiopathy (CAA) of the amyloid-beta (Abeta) type is the most common form of sporadic CAA and is now also accepted as an early and integral part of Alzheimer's disease (AD) pathogenesis. Cerebral amyloid angiopathy is a risk factor for haemorrhagic stroke and is believed to independently contribute to dementia. Rare forms of hereditary cerebral amyloidosis caused by mutations within the Abeta domain of amyloid precursor protein (APP) have been identified, where mutant Abeta preferably deposits in vessels because of a decreased fibrillogenic potential and/or increased vasotopicity. A review of factors involved in CAA caused by wild-type Abeta suggests that increased Abeta levels in brain without an increased Abeta42/Abeta40 ratio is one of the most important prerequisites for vascular amyloidosis. This is exemplified by CAA observed in APP duplication and Down's syndrome patients, neprilysin polymorphism patients and knockout mice and Swedish APP (KM670/671NL) mice. Select presenilin mutations also lead to a prominent CAA, and importantly, presenilin mutations are shown to have varied effects on the production of Abeta40, the predominant amyloid found in CAA. Conversely, APP mutations such as Austrian APP (T714I) drastically decrease Abeta40 production and are deficient in CAA. Apolipoprotein E-epsilon4 is also shown to be a risk factor for CAA, and this might be because of its specific role in the aggregation of Abeta40. Recent data also suggest that dense-core senile plaques in humans and dense plaques in transgenic mice, composed predominantly of Abeta40, associate with vessels. This review highlights some of these aspects of genetics and biochemistry of CAA and pathological descriptions linked to a prominent CAA and/or dense plaques in humans and relevant mouse models and discusses how this knowledge has led to a better understanding of the processes involved in vascular amyloidosis, and in causing dementia, and thus has important therapeutic implications.