Alzheimer's disease is associated with a substantial economic impact on patients, their caregivers and society. Due to the current rise in the aging population, the prevalence and impact of Alzheimer's disease are expected to increase greatly. The cost of caring for someone with Alzheimer's disease is magnified in the more severe stages of the disease. There are four cognitive enhancers commonly used for the treatment of Alzheimer's disease: three cholinesterase inhibitors (donepezil, rivastigmine and galantamine) and one NMDA receptor antagonist (memantine). Of these, donepezil and memantine have been approved in many countries as pharmacological treatments for moderate to severe Alzheimer's disease, while donepezil, rivastigmine and galantamine are approved treatments for mild to moderate Alzheimer's disease. While cost effectiveness has been well studied in mild to moderate Alzheimer's disease, the cost-benefit information for drug therapy in moderate to severe Alzheimer's disease is less clear. This article reviews the pharmacoeconomic data available on these four drugs, with a specific focus on moderate to severe Alzheimer's disease, including economic burden, cost drivers, clinical outcomes and pharmacoeconomic studies. A key driver of the cost of Alzheimer's disease is the severity of the disease, indicating that the ability to stabilize the disease state is a potential source of cost savings. Drug therapies that can limit increases in behavioural problems and cognitive and functional impairment, and postpone institutionalization without an increase in longevity may serve to reduce the economic burden on Alzheimer's disease patients. The data suggest that, while the available, approved agents offer only modest improvements in clinical outcomes, they could be cost-effective treatments for moderate to severe Alzheimer's disease when viewed from the societal perspective. For memantine and donepezil, data are available that suggest that the cost of these drugs is offset by the clinical and societal benefits provided by slowing the progression of Alzheimer's disease. While there are few head-to-head comparison trials, the similarity in costs of the treatments and efficacy against placebo suggest that cost effectiveness will not be substantially different among treatments. More studies that examine longitudinal resource utilization and its relationship to drug treatment in the moderate to severe stages are needed to clarify cost benefit in this population and possibly differentiate between individual medications.