Memantine is approved by the US Food and Drug Administration for the treatment of moderate to severe Alzheimer disease (AD). We investigated the frequency and variables associated with its use in mild to moderate/severe AD as defined by criteria involving the Mini-Mental Status Examination (MMSE) and Clinical Dementia Rating (CDR) scale. Consecutive possible and probable AD patients seen at our research center from November 2003 to December 2006 were included. Individuals were classified as mild dementia either by CDR=1 or MMSE >or=15, using criteria derived in part from the pivotal trials of memantine used for its approval by the Food and Drug Administration. Of 117 patients, 37% of those with mild AD by MMSE criterion (total N=94), and 38% of those with mild AD by CDR criterion (total N=86) used memantine. Logistic regression was used to simultaneously estimate the odds ratios (ORs) of the likelihood of memantine usage associated with a set of predictor variables. Lower MMSE was associated with a greater likelihood of using memantine independent of CDR [ORMMSE=7.45, 95% confidence interval (CI)=1.50-37.05]; CDR was not significantly related to memantine use. Controlling both MMSE and CDR, Whites were more likely to use memantine than African Americans (OR=6.47, 95% CI=1.25-33.39). Patients who used other antidementia medications were more likely to use memantine than those who did not (OR=3.15, 95% CI=0.995-9.97). Eight other patient characteristics were not significant predictors. Use of memantine in mild AD was common. Patterns of memantine usage are complex and deserve further study in a larger sample because of their implications for medical system cost, equitable access to care, and risk of drug interactions.