Predictive ability of the IQCODE items for detection of cognitive decline and dementia in aging adults with Down syndrome

Abstract

AbstractBackgroundThere is evidence of a genetic association between Down Syndrome (DS) and early progression of Alzheimer’s disease (AD). Few brief instruments have been validated for the screening of dementia in this population. Previously, we showed that the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) may be useful for cognitive screening in adults with DS1. Nonetheless, the adaptation of certain items to account for the presence of intellectual deficiency may increase diagnostic accuracy. The aim of the present study is to evaluate the predictive ability of IQCODE items for adults with DS.MethodUnivariate and multivariate logistic regressions were used to assess the effects of the 26 IQCODE items on the diagnosis of cognitive decline (prodromal dementia or dementia versus stable cognition) according to the Cambridge Examination for Mental Disorders of Older People with DS and Others with Intellectual Disabilities, CAMDEX‐DS. The study comprised 92 participants with DS aged ≥35 years. The accuracy of the final multivariate model was evaluated using receiver operator characteristics (ROC) curve.ResultAssociations were found between all IQCODE items and diagnosis, except for two items (handling money for shopping and financial matters). Multivariate logistic regression using the CAMDEX‐DS diagnosis as dependent and the items of IQCODE as predictors, showed that four items remained in the model (remembering: members of family and friends, day and month, things they learned when they were younger, and adjusting to change in routine). Participants who presented more difficult with these items were, respectively, 84.1, 86.1, 10.1 and 23.6 times more likely to have a diagnosis of prodromal dementia or dementia. Using the estimated probability of these 4 items, we found an area under the ROC curve of 0.961.ConclusionThe IQCODE items showed good ability to discriminate subjects with cognitive decline as indicated by the CAMDEX‐DS. The use of 4 of the IQCODE items proved to suffice to discriminate between diagnostic groups. We will next perform an adaptation and validation study to warrant the “IQCODE‐DS” as a reliable, brief test for the cognitive screening in this population.