Clinical predictors of progression to Alzheimer disease in amnestic mild cognitive impairment

Objective: To investigate the neurocognitive measures that best predict progression from amnestic mild cognitive impairment (aMCI) to Alzheimer disease (AD). Methods: We evaluated 539 participants with aMCI from the Alzheimer's Disease Cooperative Study clinical drug trial of donepezil, vitamin E, or placebo. During the study period of 36 months, 212 aMCI participants progressed to AD. Using progression from aMCI to AD within 36 months as the dependent variable, a generalized linear model was fit to the data using the least absolute shrinkage and selection operator. Independent variables included in this analysis were age, sex, education, APOE-e4 (APOE4) status, family history of dementia, Mini-Mental State Examination score, Digits Backwards (Wechsler Memory Scale), Maze Time and Errors, Number Cancellation, Delayed Recall of Alzheimer's Disease Assessment Scale Word List, New York University Paragraph Recall Test (Immediate and Delayed), Boston Naming Test, Category Fluency, Clock Drawing Test, and the Alzheimer's Disease Assessment Scale-Cognitive subscale (ADAS-cog). Results: The model that best predicted progression from aMCI to AD over 36 months included APOE4 status, the Symbol Digit Modalities Test, Delayed 10-Word List Recall, New York University Paragraph Recall Test (Delayed), and the ADAS-cog total score. When APOE4 was removed from the analysis the resulting model had a similar estimated predictive accuracy as the full model. As determined by cross-validation, the estimated predictive accuracy of the final model was 80%. Conclusion: Progression from amnestic mild cognitive impairment to Alzheimer disease in this cohort was best determined by combining four common, easily administered, cognitive measures.