BACKGROUND: There is no consensus on when and how to discontinue cholinesterase inhibitors (ChEI). Predictors of non-persistency of antidementia drugs have been poorly investigated, mostly during short-term periods and using administrative data. OBJECTIVE: The aim of this study was to investigate the incidence and predictors of ChEI switch and discontinuation among subjects with ascertained Alzheimer's disease (AD). METHODS: A total of 557 community-dwelling, mild-to-moderate AD subjects initiating ChEIs in 29 European clinic centres were assessed twice-yearly for 2 years. Antidementia drug exposure was recorded through a physician-administered structured questionnaire to document any change in drug therapy (start and stop dates, reasons). Discontinuation was defined as >35 days without any antidementia drug. Switch was defined as a change for any antidementia drug strategy within 35 days after ChEI cessation. Two separate time-dependent multivariate Cox survival analyses were conducted to identify predictors of discontinuation and switch. RESULTS: The incidences of discontinuation and switch were 9.65 and 12.47/100 person-years, respectively. Behavioural disturbances, low body mass index, falls, decline in Mini-Mental State Examination (MMSE) score, and AD-related hospitalization predicted discontinuation. MMSE score, decline in activities of daily living score, aberrant motor behaviour, shorter AD duration and higher nurse resource use predicted a switch. An ineffective ChEI dose and clinic specialty predicted both outcomes. Sensitivity analyses using a 60-day cut-off provided stable results. CONCLUSION: Several predictors were identified: adverse drug events and their predisposing factors, perceived loss of efficacy or disease progression on cognitive or functional scales, behavioural disturbances, hospitalization and professional practices. The latter implies a need for harmonization in AD drug prescription practice.