Utility of the Clinical Dementia Rating Scale in Detecting Autopsy-Proven Dementia in Patients with Low Education
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
BACKGROUND AND OBJECTIVE: The Clinical Dementia Rating scale (CDR) assesses impairment in 6 cognitive and functional domains to stage cognitive decline and dementia. Each domain is scored from 0 (no impairment) to 3 (severe impairment), and these scores are summed to a sum-of-boxes (CDR-SB) score ranging from 0 to 18. The CDR-SB score has shown high reliability in staging dementia. However, no studies have determined whether the CDR remains effective when applied to less-educated individuals. This study investigated the sensitivity and specificity of the CDR-SB score in detecting dementia associated with autopsy-proven AD in patients with less than 12 years of education. HYPOTHESIS: Using the validated CDR-SB cut score for mild dementia (4.5) to detect autopsy-proven AD in this population was hypothesized to yield low sensitivity and/or specificity (i.e. <70%). A higher cut score was expected to be required for optimal sensitivity/specificity. METHODS: Participants from the National Alzheimerメs Coordinating Center Uniform Data Set with less than 12 years of education were divided into two cohorts (autopsy-proven AD and normal age-related brain changes), matched for age and sex, and excluded if other major neurological diseases were present (n = 34; 17 per cohort). Receiver Operating Characteristic (ROC) analysis was performed to determine the sensitivity and specificity of CDR-SB scores in discriminating between subjects with autopsy-proven AD and those with normal age-related brain changes. RESULTS: The validated CDR-SB cut score for mild dementia (4.5) correctly classified 10 of 17 patients with normal age-related brain changes and 16 of 17 with autopsy-proven AD (sensitivity = .941, specificity = .588). These data reflect the unexpected presence of 7 patients with clinically-diagnosed dementia in the normal cohort and 1 patient without clinically-diagnosed dementia in the autopsy-proven AD cohort. The optimal cut score was found to be 9.5, correctly classifying 15 of 17 patients with normal age-related brain changes and 14 of 17 with autopsy-proven AD (sensitivity = .824, specificity = .882). DISCUSSION: In patients with <12 years of education, the optimal CDR-SB cut score to detect AD-related dementia (9.5) is in a range associated with moderate dementia, which may be too high for clinical utility. Although numerous neurological syndromes were excluded, factors other than education may have contributed to high CDR-SB scores in the comparison group. Further research in larger samples is needed to validate the results of this preliminary investigation.