Abstract Several drug candidates for Alzheimer's disease are being evaluated in clinical trials, with the goal of finding a drug with disease-modifying effects. When such a drug finally reaches the market, there will be a demand for accurate diagnostic tools useful for early detection of disease and for monitoring biochemical effects. The core cerebrospinal fluid (CSF) biomarkers amyloid peptides (Abeta42), total-tau and phospo-tau are promising in this respect. However, inter-center variation (caused by pre-analytical, analytical and post-analytical factors), and inter-laboratory variation (caused by analytical factors), particularly for CSF Abeta42, lowers their utility in multicenter studies. Here, we discuss the causes of these variations, and present a global quality control program to overcome them.