OBJECTIVES: The literature documents that patients with diabetes have greater risks of developing dementia, suggesting that antidiabetic agents might lower the risk of dementia. However, it remains unclear about the roles of specific anti-diabetic drug classes in developing dementia in the elderly.  Furthermore, traditional methodology of capturing dementia using claims data alone in observational studies are subject to significant underestimation of dementia. The objective of this study is to explore the association between specific anti-diabetic drugs and the risk of dementia or Alzheimer's diseases (AD) identified by an improved methodology among the elderly. METHODS: A retrospective cohort study was conducted using data from the Medicare Current Beneficiary Survey (MCBS) linked to the Medicare claims from 2001 to 2010. The study sample consisted of Medicare beneficiaries aged> 65 years and were diagnosed with diabetes.  Anti-diabetic medications, both insulins and oral anti-diabetic agents, were identified from the prescription medication file. Dementia or AD was determined based on the patients’ self-reports, diagnosis codes in Medicare claims, or dementia-targeted medications during 2-year follow up period to identify incident dementia. Multivariable logistic regressions were used to estimate the association between anti-diabetic drugs and dementia. Cross-sectional weights were used to adjust for the complex survey design. RESULTS: Overall, after adjusting for demographics, socioeconomics, comorbidities and service year, patients receiving anti-diabetic medication were not associated with higher risk of dementia compared to non-users (Odds Ratio [OR], 1.23;  95% confidence interval [CI], 0.94-1.62). We also found that exposure to specific anti-diabetic drugs were not related to the risk of developing dementia:metformin only (OR, 1.06; 95% CI, 0.69-1.64), sulfonylurea only (OR, 1.04; 95% CI, 0.78-1.40), insulins only (OR, 0.77; 95% CI, 0.40-1.48), thiazolidinedione only (OR,0.77; 95% CI, 0.41-1.45). CONCLUSIONS: Use of anti-diabetic drugs, such as insulins, metformin, sulfonylurea and thiazolidinedione, are not associated with an altered risk of dementia among elderly.