Background: AD is characterized by cerebral deposition of β-amyloid plaques with amyloid β-peptide (Aβ) 42 as the major peptide constituent, along with neurofibrillary tangles and neuronal loss. In transgenic mice, active immunization against Aβ42 removes these plaques and improves cognitive function. A Phase I study in AD patients demonstrated good safety and tolerability of multiple injections of aggregated Aβ42 (AN1792) with QS-21 as adjuvant.

Methods: Three hundred seventy-two patients with mild to moderate AD were randomized to receive IM injections of AN1792 or placebo (4:1) at baseline and at months 1, 3, 6, 9, and 12 in a multicenter Phase II safety, tolerability, and pilot efficacy study. Dosing was terminated after four early reports of meningoencephalitis, but follow-up continued. The study remains blinded, and further results will be reported after its termination.

Results: Symptoms and laboratory findings consistent with meningoencephalitis occurred in 18 of 298 (6%) patients treated with AN1792 compared with 0 of 74 on placebo (p = 0.020). Sixteen of the 18 had received two doses, one had received one dose, and one had received three doses of the study drug before symptoms occurred. The median latency from the first and last injections to symptoms was 75 and 40 days. No case occurred later than 6 months after the first immunization. Anti-Aβ42 antibody titers were not correlated with the occurrence or severity of symptoms or relapses. Twelve patients recovered to or close to baseline within weeks, whereas six remain with disabling cognitive or neurologic sequelae. All 18 patients remain alive to date (December 31, 2002), 6 months to >1 year after symptom onset.

Conclusions: Postvaccination meningoencephalitis occurred without clear relation to serum anti-Aβ42 antibody titers. Potential mechanisms such as T-cell and microglial activation may be responsible and are under consideration to develop a safer anti-Aβ immunotherapy for AD.