The amyloid cascade hypothesis of Alzheimer's disease (AD) is testable: it implies that interference with Abeta aggregation and plaque formation may be therapeutically useful. Abeta42 immunisation of amyloid precursor protein (APP) transgenic mice prevented plaque formation and caused removal of existing plaques. The first clinical studies of Abeta immunisation in AD patients (AN1792, Elan Pharmaceuticals) were halted when some patients suffered side effects. Since our confirmation that Abeta immunisation can prompt plaque removal in human AD, we have performed a clinical and neuropathological follow up of AD patients in the initial Elan Abeta immunisation trial. In immunised AD patients, we found: a lower Abeta load, with evidence that plaques had been removed; a reduced tau load in neuronal processes, but not in cell bodies; and no evidence of a beneficial effect on synapses. There were pathological "side effects" including: increased microglial activation; increased cerebral amyloid angiopathy; and there is some evidence for increased soluble/oligomeric Abeta. A pathophysiological mechanism involving effects on the cerebral vasculature is proposed for the clinical side effects observed with some active and passive vaccine protocols. Our current knowledge of the effects of Abeta immunotherapy is based on functional information from the early clinical trials and a few post mortem cases. Several further clinical studies are underway using a variety of protocols and important clinical, imaging and neuropathological data will become available in the near future. The information obtained will be important in helping to understand the pathogenesis not only of AD but also of other neurodegenerative disorders associated with protein aggregation.