Au stade précoce, le bilan neuropsychologique est l'un des éléments essentiels de l'examen clinique pour le dépistage précoce de la maladie d'Alzheimer. Le but de cette étude est de valider en langue française une batterie d'évaluation courte (B2C) composée de 4 tests : le test d'orientation temporelle, l'épreuve des 5 mots, le test de l'horloge, le test de fluence verbale sémantique. La B2C a été administrée à 123 sujets. Le diagnostic de chaque patient a été porté au terme d'un bilan réalisé dans un centre Mémoire ; 49 sujets avaient une maladie d'Alzheimer (MA) d'intensité légère, 27 sujets présentaient une symptomatologie dépressive et 47 étaient sujets témoins. Le temps moyen pour l'administration de la batterie était de 11,2 minutes dans le groupe MA, 8,2 dans le groupe dépression et 7,6 dans le groupe contrôle (p < 0,001). L'analyse de la courbe ROC (Response Operating Characteristics) met en évidence, pour la comparaison entre patients MA et sujets contrôle, une sensibilité de 93,8 % et une spécificité de 85 %. Pour la comparaison entre MA et patients déprimés, la sensibilité est de 63 % et la spécificité de 96 %.

Alzheimer's disease (AD) is a major healthcare challenge due to the increasing longevity of the population. Clinically prominent neuropsychological and neurological impairments, together with behavioral disorders characterize Alzheimer's disease (AD). In the past, behavioural and emotional disturbances received less attention than cognitive symptoms in studies of dementia. The association between cognitive and behavioural symptoms is complicated by the fact that such association could also occur with different patterns during depressive episode without dementia. Because Alzhei[shy]mer's disease (AD) tends to be under diagnosed, there is an increasing need for accurate neuropsychological screening tools that are easy to administer by psychiatrists. The aim of the present study was to validate, in French, a sensitive and specific screening battery (B2C) designed to improve the discrimination between patients with AD, patients with depression, and healthy elderly subjects. Population and method – The B2C was administered to 123 ambulatory subjects (mean age 76.4 ± 2.3 years) : divided in three groups of subjects. AD subjects were included (n = 49) with a Mini-Mental Status Examination (MMSE) score of between 18 and 26, and a confirmed diagnosis (DSM IV) of mild to moderate AD. Subjects were not included if they were receiving treatment with an acetylcholinesterase inhibitor. The depressive group comprised elderly subjects (n = 27) with at least two DSM IV criteria for a major depressive episode including the depressive mood criterion and a score of more than 17 on the Montgomery-Åsberg Depression Rating Scale (MADRS). The healthy control group (n = 47) comprised age-matched subjects with no neurological or psychiatric patho[shy]logy. The B2C consists of four individual tasks derived from classical neuropsychological tests. Tasks were presented in the following order : temporal orientation test (knowledge of month, date, year, day of the week and time of day), 5 word test (task is originally derived from the Enhanced Cued Recall test), clock drawing test (In this widely used test, the subject had to draw a clock with all the numbers and then draw the clock hands at twenty minutes to four), and the semantic verbal fluency test (the subject was asked to generate as many words as possible from a given category in a fixed time period of 60 seconds). During the pre-study investigator meeting, the test procedure was adapted to ensure uniformity of practice in all centres. The B2C was administered one week to one month after the study inclusion date by a psychologist blinded to the patient groups and who had not participated in the subject's inclusion. Multivariate analysis was performed using a forced model of all four tests. Due to the nature of the dependant variable (AD vs controls and depressive vs control), the chosen discrimination model was a binary logistical regression model. Explanatory varia[shy]bles were limited to the variables of the test battery, and the dependent variable was the subject's status (AD, depressive or control). Results – The mean results for each test are presented in Table II . The time taken to perform the tests was significantly higher (p = 0.0001) for the AD group (11.2 minutes) when compared with both the control (7.6 minutes) and depressive group (8.2 minutes). In each of the four subtests, the AD subjects were significantly more impaired than the two other groups. Multivariate analysis was performed using a forced model of all four tests which provided correct classification of a high percentage of subjects (88.5 %). The analysis followed a normal distribution and demonstrated that the AD patients were significantly impaired in all four tests of the B2C compared with controls. Depressive, elderly subjects were only impaired in verbal fluency. Multivariate analysis showed that, compared with controls, patients with mild AD were significantly impaired for all four tests. Response ope[shy]rating characteristics (ROC) analysis of the B2C showed : 93.8 % sensitivity and 85 % specificity for discriminating AD from control patients (table III) , and 63 % sensitivity and 96 % specificity for discriminating AD from depressive patients (table IV) . Discussion – The main objective of this study was to demonstrate that the Short Cognitive Evaluation Battery developed in the French language is able to discriminate between patients suffering from AD and healthy elderly subjects. The results clearly demonstrate that AD patients were significantly impaired in all four tests of the B2C compared with the control group. The present study also supports the use of the screening battery for discriminating between AD and depressive subjects. The SCEB was less discriminatory for AD versus depressive patients than for AD versus controls. This could be due to the limited size of the depressive group. The verbal fluency test was the most sensitive for discriminating between AD and depression but this was at the expense of specificity. Other brief screening tests have already been developed in English speaking countries, In French langage, the B2C appears to be a highly sensitive and specific tool for discriminating between patients with mild AD and healthy elderly individuals. Furthermore, in combination with clinical evaluation, the B2C could improve the specificity of the difficult discrimination between mild AD and depression. The next step of the validation process will include concurrent validity study and inclusion of a higher number of subjects with depressive symptoms.