CERAD neuropsychological compound scores are accurate in detecting prodromal alzheimer's disease: a prospective AddNeuroMed study.

The Consortium to Establish a Registry for Alzheimer's Disease Neuropsychological Battery (CERAD-NB) is a widely used tool in screening for Alzheimer's disease (AD), however, it does not include a validated total score for delayed memory. Our aim was to develop clinically applicable memory compound scores for CERAD-NB and examine whether they and global cognitive total scores could detect prodromal AD and cognitive progression in MCI. One year follow-up data of 201 subjects with a baseline diagnosis of MCI (46 progressed to AD; 155 remained stable) and 212 controls in the European multicenter AddNeuroMed study were analyzed. Two previously described cognitive total scores and two memory compound scores were tabulated for CERAD-NB. Receiver Operating Characteristic analysis was applied in the group discrimination at baseline and the annual change for different compound scores was examined. Normative cut-offs for CERAD compound scores were tabulated in the Finnish CERAD sample of 306 controls. Country adjusted CERAD compound scores (AUC 0.95-0.96) were more accurate than Word List Recall (AUC 0.93) and Mini-Mental State Examination (AUC 0.90) in discriminating progressive mild cognitive impairment (MCI) subjects from controls. With normative cut-off values CERAD total scores yielded to 87-89% sensitivity and 84-86% specificity in screening for prodromal AD in a separate multinational population. The annual deterioration in all CERAD compound scores was significant in the progressive (p ≤ 0.001) but not in the stable MCI group (p > 0.08). CERAD total scores are a practical way of screening for prodromal AD and assessing cognitive progression in MCI. The new memory compound scores were more accurate than CERAD subtests in predicting AD conversion.