Antibodies against Aβ have been suggested as potential therapeutic strategies for the treatment of Alzheimer disease (AD) for nearly 8 years. Animal studies have been very encouraging in that both active and passive immunization of transgenic mice can reduce amyloid load and reverse memory deficits found in these mice. Three mechanisms have been proposed to explain these results: (a) catalytic conversion of fibrillar Aβ to less toxic forms, (b) opsonization of Aβ deposits leading to microglial phagocytosis, or (c) promote the efflux of Aβ from the brain to the circulation. Evidence exists supporting all three mechanisms, which, it should be noted, are not mutually exclusive. Phase 2 clinical trials of active immunization with vaccines against human Aβ1-42 were halted due to an unacceptable incidence of meningoencephalitic reactions (6% of patients treated). However, a recent report from a fraction of the patients in this trial found that those patients developing antibodies which reacted with brain amyloid deposits had a significantly slower progression of cognitive loss over a period of 12 months. This supports the continued cautious testing of passive immunization and, possibly even active immunization against the Aβ peptide using preparations less likely to cause autoimmune reactions in the central nervous system.