Alzheimer's disease (AD) is a chronic neurodegenerative disease causing progressive impairment of memory and cognitive function. The amyloid cascade hypothesis suggests that mismetabolism of the beta-amyloid (A beta) precursor protein (APP) followed by subsequent formation of non-fibrillar and fibrillar A beta deposits leads to glial activation and eventually to neurotoxicity, causing cognitive impairment. Several lines of evidence indicate that an inflammatory process contributes to the pathology of AD. First, inflammatory proteins have been identified as being associated with neuritic plaques and in glial cells surrounding these plaques. Second, certain polymorphisms of acute-phase proteins and cytokines associated with AD plaques increase the risk or predispose for earlier onset of developing AD. Third, epidemiological studies indicate that anti-inflammatory drugs can retard the development of AD. Several steps in the pathological cascade of AD have been identified as possible targets for actions of nonsteroidal anti-inflammatory drugs. For instance, microglia are considered a target because this cell type is closely involved in AD pathology through secretion of neurotoxic substances and by modulating a positive feedback loop of the inflammatory mechanism that may be involved in the pathological cascade in AD. On the basis of studies in APP transgenic mice, immunisation with A beta was recently suggested as a novel immunological approach for the treatment of AD. Immunisation elicits A beta-specific antibodies that could affect several early steps of the amyloid-driven cascade. Antibodies could prevent A beta from aggregating into fibrils and accelerate clearance of A beta by stimulating its removal by microglial cells. This review outlines the pathological and genetic evidence that an inflammatory mechanism is involved in AD and the therapeutic approaches based on inhibition or mediation of inflammation.