OBJECTIVE: To assess the deveopment of costs for dementia in Germany until the year 2050 under different demographic scenarios and the possible impact of treatment interventions with AChE inhibitors from various perspectives. METHODS: A model to estimate cost differences in dementia treatment with AChE inhibitors compared to a placebo scenario was developed. The model projects the number of dementia patients in Germany until 2050 under the assumption of extended life expectancy and immigration. Markov modelling allows to document the progression of dementia patients into more severe stages. Data of the population development were supplied by the Federal Bureau of Statistics and the German Institute for Economic Research. Number of patients were calculated from demographic prognoses considering prevalence of disease data from a published meta-analyses of the epidemiology of dementia. Data about transition probabilities between different disease stages were derived from a randomized clinical trial of an AChE inhibitor. Disease stage specific costs were taken from a cost of illness study previously published. RESULTS: The number of dementia patients will increase from 1 million to 2.5 million over the next 50 years in Germany. With treatment there will be 30% less patients in the most severe, and hence most costly disease stage. Costs for treatment are offset by cost savings through a shift of patients to less severe disease stages, if indirect family and caring costs are included. Reduction of direct costs due to treatment increase from 0.5 to 1.5 billion Euro per year and indirect cost reductions are 4-fold these figures. For 2000 net savings of 1.4 billion Euro are demonstrated for the societal perspective. Results are stable for extensive sensitivity analysis. CONCLUSION: Results document that the dimension of cost for care should make dementia a health policy priority. In an ageing society dementia treatment is also economically highly attractive.