The ideal biomarker for Alzheimer's disease (AD) should detect a fundamental feature of neuropathology and be validated in neuropathologically-confirmed cases; it should have a sensitivity >80% for detecting AD and a specificity of >80% for distinguishing other dementias; it should be reliable, reproducible, non-invasive, simple to perform, and inexpensive. Recommended steps to establish a biomarker include confirmation by at least two independent studies conducted by qualified investigators with the results published in peer-reviewed journals. Our review of current candidate markers indicates that for suspected early-onset familial AD, it is appropriate to search for mutations in the presenilin 1, presenilin 2, and amyloid precursor protein genes. Individuals with these mutations typically have increased levels of the amyloid Abeta42 peptide in plasma and decreased levels of APPs in cerebrospinal fluid. In late-onset and sporadic AD, these measures are not useful, but detecting an apolipoprotein E e4 allele can add confidence to the clinical diagnosis. Among the other proposed molecular and biochemical markers for sporadic AD, cerebrospinal fluid assays showing low levels of Abeta42 and high levels of tau come closest to fulfilling criteria for a useful biomarker.