BACKGROUND: Alzheimer's Disease Neuroimaging Initiatives ("ADNI") aim to validate neuroimaging and biochemical markers of Alzheimer's disease (AD). Data of the pilot European-ADNI (E-ADNI) biological marker programme of cerebrospinal fluid (CSF) and plasma candidate biomarkers are reported. METHODS: Six academic EADC centres recruited 49 subjects (healthy controls, subjects with mild cognitive impairment (MCI) and AD). We measured CSF beta-amyloid 42 (CSF Abeta42), total tau-protein (t-tau), phosphorylated tau-proteins (P-tau181, P-tau231), plasma beta-amyloid 40 and 42 (Abeta40/Abeta42). Immediate fresh shipment was compared to freezing and later shipment on dry ice. RESULTS: CSF T-tau (fresh samples) was increased in AD versus controls (p=0.049), CSF Abeta42 (frozen samples) was decreased in MCI and AD (p=0.02), as well as plasma Abeta40 (fresh and frozen samples) in AD (p=0.049 and p=0.016). Pooled values of neurochemical parameters and ratios thereof were different between centres (p<0.005). Analysis of frozen samples yielded higher diagnostic accuracy than immediate fresh shipment with 100% (fresh: 100%) correctly classified in control subjects, 100% (78%) in MCI, 91% (91%) in AD. CONCLUSION: The use of frozen rather than fresh samples renders higher diagnostic accuracy within a multicentre context. We confirmed the feasibility of a multicentre AD biomarker programme for future clinical trials.