The first Aβ immunotherapy protocol, active immunisation with Aβ42 (Elan Pharmaceuticals), was effective in removing plaques but encountered side effects associated with focal white matter abnormalities, inflammation and neurological dysfunction, putatively due to cerebrovascular pathophysiology. Subsequent passive immunotherapy has shown similar changes especially in APOE ∊4 carriers (Bapineuzimab, Elan Pharmaceuticals) raising the possibility that it may be a complication universally encountered with Aβ immunotherapy. We hypothesise that immunization prompts solubilization of Aβ plaques with relocation of solubilized Aβ to the cerebral vasculature, resulting in increased severity of CAA. In addition, the formation of immune complexes in the brain may exacerbate these problems. We have performed a clinical and neuropathological follow up of 16 patients immunized with Aβ42. The relative distribution of Aβ in plaques and in the cerebral vasculature (cerebral amyloid angiopathy - CAA) was assessed and compared with unimmunized AD cases. We have generated two experimental models to mimic the human situation. In the first model, dextran is injected intracerebrally (i.c.) to mimic drainage in physiologic conditions. In the second model, to mimic the effects of immune complexes on the drainage of soluble protein present in the cerebral extracellular space, mice are immunized against ovalbumin, injected i.c. with ovalbumin to form immune complexes and subsequently injected i.c. with dextran. In the first experimental model, soluble dextran is localised to the walls of cerebral blood vessels 10-15 min after the injection. The second model shows that (i) immune complexes form in the walls of cerebral blood vessels and (ii) this exacerbates accumulation of dextran in the vessel walls and perivascular spaces. In the human study, plaque reduction is associated with increased severity of CAA and inflammation. The models indicate that IgG immune complex formation in the blood vessel walls leads to enhanced accumulation of substances solubilized in the extracellular space. This pathophysiological process may play a role in the increased CAA observed in immunized AD patients and the side effects of the immunotherapy trials. With better understanding of these processes it may be possible for them to be circumvented.