Red blood cell (RBC) acetylcholinesterase (AChE) inhibition has been used as a peripheral surrogate marker for the activity of centrally acting AChE inhibitors (AChEIs) in the treatment of Alzheimer disease. As a valid peripheral surrogate marker, RBC AChE inhibition should reflect the central pharmacodynamic activity of the compound and should demonstrate a relation with cognitive or global improvement in patients with Alzheimer disease. As a useful clinical tool, RBC AChE inhibition should also provide an advantage in dose optimization. However, the application of surrogate markers in research and clinical use is controversial (Prentice, 1989; Gotzsche, 1996; Colburn, 1997; De Gruttola et al., 1997). For instance, surrogate markers that have been identified or applied inappropriately can lead to erroneous conclusions, slowing the drug development process (Colburn, 1997). Also, the validation of surrogate markers for the pharmacodynamic activity of central nervous system drugs is not always possible because samples of brain tissue cannot be analyzed in humans. Finally, although validation of peripheral markers for central nervous system drugs has been approached via analysis of cerebrospinal fluid (Cutler et al., 1998a), few markers have been subjected to such rigorous evaluation in clinical studies. The extent to which measures of peripheral AChE inhibition accurately model central drug activity and therapeutic effectiveness of AChEIs, both as individual agents and as a drug class, is the focus of this review. AChEIs comprise a group of structurally diverse compounds with a wide range of relative specificities for the various molecular species of cholinesterase found in plasma, RBCs, and the brain. Studies of RBC AChE inhibition after administration of AChEIs in animals are of limited utility because of the differential sensitivity of AChEIs for human versus animal forms of AChE, the poor correlation between effective doses in animals and humans, and the lack of standardized measurements of effectiveness. Although clinical studies of donepezil, metrifonate, and eptastigmine have suggested the potential use of RBC AChE inhibition as a predictor of clinical response, the degree of inhibition yielding maximum cognitive improvements was highly variable from compound to compound (30-80%). Further, investigators did not prove a relation between central and peripheral pharmacodynamics or demonstrate an advantage over dose in the ability of RBC AChE inhibition to predict clinical response. A study of rivastigmine in patients with Alzheimer disease revealed that cerebrospinal fluid AChE inhibition correlated well with cognitive performance, whereas peripheral inhibition did not. Therefore, RBC cholinesterase inhibition is not a reliable surrogate marker for the activity of AChEIs as a class of drugs, and its usefulness as a dose optimization tool for individual agents has yet to be demonstrated clearly.