Résumé
Objective: To validate the Rowland Universal Dementia Assessment Scale for use in Brazil (RUDAS-BR). Methods: We first completed an English-Brazilian Portuguese translation and back-translation of the RUDAS. A total of 135 subjects over 60 years of age were included: 65 cognitively healthy and 70 with Alzheimer's disease (AD) according to the DSM-IV and Neurological and Communicative Disorders and Stroke/Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria. All participants completed an interview and were screened for depression. The receiver operating characteristic curves of the RUDAS were compared with those of the Mini Mental State Examination (MMSE) regarding the sensitivity and specificity of cutoffs, taking education into consideration. Results: The areas under the curve were similar for the RUDAS-BR (0.87 [95%CI 0.82-0.93]) and the MMSE (0.84 [95%CI 0.7-0.90]). RUDAS-BR scores < 23 indicated dementia, with sensitivity of 81.5% and specificity of 76.1%. MMSE < 24 indicated dementia, with sensitivity of 72.3% and specificity of 78.9%. The cutoff score was influenced by years of education on the MMSE, but not on the RUDAS-BR. Conclusions: The RUDAS-BR is as accurate as the MMSE in screening for dementia. RUDAS-BR scores were not influenced by education. The RUDAS-BR may improve the cognitive assessment of older persons who are illiterate or of lower educational attainment.