Predicting decline in mild cognitive impairment: a prospective cognitive study.
|Title||Predicting decline in mild cognitive impairment: a prospective cognitive study.|
|Publication Type||Journal Article|
|Year of Publication||2014|
|Authors||Belleville S, Gauthier S, Lepage E, Kergoat M-J, Gilbert B|
|Date Published||2014 Jul|
|Keywords||Aged, Aged, 80 and over, Analysis of Variance, Dementia, Disease Progression, Executive Function, Female, Humans, Language, Logistic Models, Male, Memory Disorders, Memory, Episodic, Memory, Short-Term, Mental Status Schedule, Middle Aged, Mild Cognitive Impairment, Neuropsychological Tests, Predictive Value of Tests, Prospective Studies, Visual Perception|
OBJECTIVE: The primary aim of this study was to identify cognitive tests that differentiate between persons with mild cognitive impairment (MCI) who later develop cognitive decline and those who remain stable.METHOD: This study used a prospective longitudinal design. One hundred twenty-two older adults with single-domain or multiple-domain amnestic MCI were recruited from memory clinics. They completed tests to measure baseline episodic memory, working memory, executive functions, perception, and language. They were then followed annually to determine with criteria independent from those tests whether they had remained stable or had developed dementia or significant cognitive decline. This was used as the reference standard to measure diagnostic test accuracy value.RESULTS: ANOVAs indicated that participants with progressive MCI showed more impaired performance than those with stable MCI at baseline on episodic memory (word and story recall), the Brown-Peterson working memory test, object naming, object decision, and position of gap test. Logistic regression derived a significant model with 87.8% overall predictive value. The model included delayed text memory, free recall, naming, orientation match, object decision, and alpha span. Its sensitivity was 86.2% and its specificity was 88.9%. Positive predictive value was 83.3%, and negative predictive value was particularly high at 90.9%.CONCLUSIONS: Identifying individuals with MCI who will progress to dementia or more severe cognitive impairment is a challenge. This study shows that cognitive measures provide valuable information regarding the predictive diagnosis of persons with MCI. Predictive accuracy of a cognitive battery might be optimized by selecting both memory and nonmemory measures.
|Grant List||/ / Canadian Institutes of Health Research / Canada|