Premium
P2–215: Is the laboratory evaluation performed for dementia useful in mild cognitive impairment?
Author(s) -
Pereira Ana Filipa,
Couto Frederico Simões,
de Mendonça Alexandre
Publication year - 2006
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2006.05.1054
Subject(s) - dementia , medicine , cognitive impairment , cognition , pediatrics , memory clinic , cognitive test , psychiatry , disease
the accuracy of CAMCOG for the identification of cases of Mild Cognitive Impairment (MCI) and dementia in a cohort of Brazilian elderly of high educational level. Methods: 140 individuals were included (mean age: 69 7 years). They had at least 8 years of formal education (mean schooling 14.4 3.8 years). Subjects were divided into 3 groups according to cognitive profile: 54 were controls, 61 had MCI and 25 had mild or moderate dementia (DSM-III-R). The cut-off values were established through ROC curve analysis. Results: Comparing controls with subjects with dementia, the cut-off value was 93 (AUC 0.99: sensitivity: 100%, specificity: 93%) and the subscales most relevant to differentiation were orientation, (AUC 0.86), language (AUC 0.96), memory (AUC 0.89) and abstraction (AUC 0.85). Comparing subjects with MCI and dementia: CAMCOG 86 (AUC 0.89, sensitivity: 80%, specificity: 80%) and the subscales most relevant were orientation (AUC 0.83), language (AUC 0.82) and memory (AUC 0.92). Comparing controls with MCI: CAMCOG 96 (AUC 0.83, sensitivity: 67%, specificity: 87%) and no subscale gave a good discrimination. Conclusions: The best cut-off value to differentiate controls and demented retests in this sample (93 points) was considerably higher than the usually accepted (79/80). One possible reason was that the demented subjects had mostly mild or initial symptoms of Alzheimer’s disease. The CAMCOG had limited value to discriminate subjects with MCI from controls. The discrimination of MCI and dementia can be optimized using not only the total CAMCOG score but also the subtests memory and language. Studies with subjects with low schooling are necessary to permit broader use of this test in Brazil.