Recent clinical and neuropathological data show that the AN 1792 vaccine enhanced the production of Abeta antibodies in the sera of Alzheimer's disease (AD) patients, but it appears to have been ineffective at stimulating the removal of Abeta deposits from the brain or at slowing the rate of cognitive decline. The 19 cases of meningoencephalitis were not linked in an obvious way to serum antibody titre, but they may have been linked to infiltration of the brain by antibodies and/or T-cells. Brain imaging indicated that oedema associated with the neuroinflammation did not reflect the typical distribution of neuritic plaques in AD. These outcomes were not anticipated by experiments on transgenic mice because compared to humans, these mice have less genetic variability, and their plaques have a different chemical composition, making them far more soluble and easier to remove. Furthermore, the consequences of vaccination are different. Vaccination of transgenic mice removes superfluous human Abeta while leaving endogenous mouse Abeta intact, whereas in humans the immune response is directed against an endogenous target that occurs naturally and is present in healthy brain tissue. The most important lesson to be learned from the AN 1792 trials is that new strategies for treating AD should not be tested on humans until they have been extensively tested on non-murine species.