Neuropsychologic variables in sporadic late-onset Alzheimer disease in patients of spanish nationality: a fast and efficient assessment procedure.

Objective: To learn which neuropsychologic tasks are affected in Spanish patients with sporadic late-onset Alzheimer Disease and to determine which are more relevant to make a diagnosis with minimum cost but without losing accuracy.

Background: The sample used in this study comprised of 88 subjects: 42 subjects diagnosed with SLOAD and 46 subjects who, with no family history of dementia at the time of examination, showed no type of neurologic or psychiatric disorder; neither did they have any record of alcoholism or drug addiction.

Methods: The evaluation of each subject in the clinical group consisted of a full neurologic examination, a neuropsychologic evaluation with a battery of tests designed for this study, and an assessment of mood with the Beck Depression Inventory. The battery of neuropsychologic tests used in diagnosing AD was subjected to discriminant analysis by the stepwise method, using diagnosis (1 = AD, 2 = control) as a dependent variable. The objective of the stepwise method of this technique is to select from the set of independent variables those that best discriminate between the two groups of the dependent variable, thus, satisfying the criterion of parsimony with which greater classifying specificity is obtained with a minimum number of variables.

Results: The analysis selected the following tests as optimum and sufficient: Babcock Story Recall (Delay), Categories Completed (Wisconsin), and Remote Memory.

Conclusions: The results of discriminant analysis made it possible to obtain a highly significant discriminant function. This analysis reduced the initial battery to three tests, yielding a sensitivity of 96.6% in the correct diagnostic classification of the subjects.