The leading hypothesis of the pathophysiology of Alzheimer's disease holds that the pivotal event is cleavage of the amyloid precursor protein to release intact the 42-amino-acid amyloid-beta peptide (Abeta); this hypothesis best explains the known genetic causes of Alzheimer's disease. If this theory is correct, optimal strategies for altering the disease process should be directed toward modifying the generation, clearance and/or toxicity of Abeta. Abeta is highly aggregable, spontaneously assuming a beta-sheet conformation and polymerising into oligomers, protofibrils, fibrils and plaques. The relative contribution of the various forms of Abeta to neuronal dysfunction in Alzheimer's disease remains uncertain; however, recent evidence implicates diffusible oligomeric species. This article reviews the range of strategies that have been investigated to target Abeta to slow the progression of Alzheimer's disease, from secretase modulators to anti-polymerisation agents. One amyloid-binding drug, tramiprosate (3-amino-1-propanesulfonic acid; Alzhemed), which is effective in reducing polymerisation in vitro and plaque deposition in animals, has now reached phase III clinical trials. Thus, it is plausible that an effective anti-amyloid strategy will become available for the treatment of Alzheimer's disease within the next few years.