The cytotoxic β-amyloid peptide (Aβ) of Alzheimer's disease (AD) occurs in both plasma and platelets and may modulate platelet function. Its biological activity may relate to its fibril content and factors that promote Aβ fibrillogenesis, e.g., plasma lipoproteins could, therefore, have implications for Aβ action. We undertook a study in which structure–activity relationships were considered with respect to the actions of Aβ 1–40 on platelet function. Thus, the influence of soluble Aβ and various fibrillar Aβ preparations (0.1–10 μM) on platelet aggregation and endogenous 5-hydroxytryptamine (5-HT) efflux was investigated. Soluble Aβ 1–40 only enhanced platelet aggregation (+30%, P<0.05) and 5-HT release (+28%) stimulated by ADP (1 μM) at the highest concentration tested (10 μM). By contrast, fibrillar Aβ 1–40 at 1, 5 and 10 μM potentiated aggregation by 17.4%, 68.8% (P<0.05) and 99.5% (P<0.0001), respectively, and 5-HT efflux by 17.4%, 65% and 208% (P<0.001). Aβ 1–40 fibrils generated in the presence of native and oxidised very low-density lipoprotein (VLDL), low-density lipoprotein (LDL) and high-density lipoprotein (HDL) yielded platelet responses that did not differ from those seen with the lipoproteins alone. These responses were markedly lower than those obtained with homogeneous Aβ fibrils. Our data indicate that homogeneous Aβ 1–40 fibrils are more potent than soluble Aβ 1–40 in promoting platelet reactivity and that interactions with plasma lipoproteins result in the formation of Aβ fibrils that are ineffective. We suggest that lipoproteins may interfere with the recognition of Aβ by appropriate platelet receptors and/or cause Aβ to assume an “overaggregated” biologically inert state.