Validation de la batterie rapide de dénomination (BARD) chez 382 témoins et 1004 patients d’une consultation mémoire

Abstract

The Rapid BAttery of Denomination (BARD) is a short 10-item naming test derived from the 60-item Boston Naming Test. It is easily performed in less than 15 seconds by normal controls independently of age, gender and education (Croisile, 2005,2007,2008). Our aim was to evaluate the BARD in various conditions seen in a memory clinic. The BARD was used in 382 normal subjects (165 men and 217 women, aged from 20 to 97 years) and 1004 patients attending a memory clinic. Three groups of 505 patients with Alzheimer's disease (AD) were compared: mild patients (n=402), moderate patients (n=84) and moderately severe patients (n=19). The BARD was also used in 499 patients with a Mini Mental Status (MMSE)>or=20: 173 patients with amnestic Mild Cognitive Impairment (aMCI), 56 patients with frontotemporal dementia (FTD), 41 patients with Lewy Body dementia (LBD), 36 patients with nonfluent primary progressive aphasia (NFPPA), 27 patients with semantic dementia (SD), 16 patients with posterior cortical atrophy (PCA), 150 patients with anxiety or depression (ADD). The performance of the patients was not affected by age, gender or education. aMCI had a score of 9.97+/-0.18, ADD a score of 9.97+/-0.2. A mild anomia was observed in three groups: mild AD (9.78+/-0.5), FTD (9.79+/-0.65) et LBD (9.98+/-0.16). A more pronounced anomia was present in moderate AD (9.10+/-1.06), moderately severe AD (8.05+/-1.27), PCA (8.12+/-3.28) and NFPPA (8.44+/-1.61). The anomia was severe in SD (5.85+/-2.46). The 10 items were perfectly named by 98 % of ADD, 96.53 % of aMCI, 82.09 % of mild AD, 87.5 % of FTD patients, 97.56 % of LBD patients, 68.75 % of PCA patients, but only 45.24 % moderate AD, 5.26 % of moderately severe AD, 27.78 % of NFPPA, and 3.7 % of SD. In the patients with MMS>or=20, Anova showed that the BARD scores of the ADD, aMCI, mild AD, FTD and LBD groups were significantly greater than the BARD scores of NFPPA, SD and PCA. PCA and NFPPA groups did not differ for BARD scores whereas they were significantly better than SD. A ROC curve comparing the 822 mild anomic patients (AD, FTD, LBD, aMCI, ADD) with the 79 more anomic patients (NFPPA, SD, PCA) showed that for a BARD score of 10, sensitivity was 72.2 %, specificity was 89.2 %, and 87.7 % of the patients were correctly classified. The BARD is a quick and useful tool for identifying naming disorders in a memory clinic. In patients with MMSE>or=20, making one error at the BARD is highly abnormal and significantly characteristic of cognitive disorders: the more frequent the errors are, the more probable is the presence of a visual agnosia (PCA), an aphasia (NFPPA), or a semantic disorder (SD).