ABSTRACT
BACKGROUND: Individuals with Down syndrome (DS) are at high risk for dementia, specifically Alzheimer's disease. However, many measures regularly used for the detection of dementia in the general population are not suitable for individuals with DS due in part to floor effects. Some measures, including the Severe Impairment Battery (SIB), Brief Praxis Test (BPT) and Dementia Scale for People with Learning Disabilities (DLD), have been used in clinical trials and other research with this population. Validity research is limited, particularly regarding the use of such tools for detection of prodromal dementia in the DS population. The current project presents baseline cross-sectional SIB, BPT and DLD performance in order to characterise their predictive utility in discriminating normal cognition, possible dementia and probable dementia in adult DS. METHOD: Baseline SIB, BPT and DLD performances from 100 individuals (no dementia = 68, possible dementia = 16 & probable dementia = 16) were examined from a longitudinal cohort of aging individuals with DS. Receiver operating characteristic curves investigated the accuracy of these measures in relation to consensus dementia diagnoses, diagnoses which demonstrated high percent agreement with the examining neurologist's independent diagnostic impression. RESULTS: The SIB and BPT exhibited fair discrimination ability for differentiating no/possible versus probable dementia [area under the curve (AUC) = 0.61 and 0.66, respectively]. The DLD exhibited good discrimination ability for differentiating no versus possible/probable dementia (AUC = 0.75) and further demonstrated better performance of the DLD Cognitive subscale compared with the DLD Social subscale (AUC = 0.77 and 0.67, respectively). CONCLUSIONS: Results suggest that the SIB, BPT and DLD are able to reasonably discriminate consensus dementia diagnoses in individuals with DS, supporting their continued use in the clinical assessment of dementia in DS. The general performance of these measures suggests that further work in the area of test development is needed to improve on the AUCs for dementia status discrimination in this unique population. At present, however, the current findings suggest that the DLD may be the best option for reliable identification of prodromal dementia in this population, reinforcing the importance of including informant behaviour ratings in assessment of cognition for adults with DS.