Abstract
BACKGROUND: The Alzheimer's Disease Assessment Scale cognitive subscale (ADAS-Cog) is widely used in AD, but may be less responsive to change when used in people with mild cognitive impairment (MCI). METHODS: Participants from the Alzheimer's Disease Neuroimaging Initiative were administered a neuropsychological battery and 1.5 T MRI scans over 2-3 years. Informants were queried regarding functional impairments. Some participants had lumbar punctures to obtain cerebrospinal fluid (CSF). We added executive functioning (EF) and functional ability (FA) items to the ADAS-Cog to generate candidate augmented measures. We calibrated these candidates using baseline data (n = 811) and selected the best candidate that added EF items alone and that added EF and FA items. We selected candidates based on their responsiveness over three years in a training sample of participants with MCI (n = 160). We compared traditional ADAS-Cog scores with the two candidates based on their responsiveness in a validation sample of participants with MCI (n = 234), ability to predict conversion to dementia (n = 394), strength of association with baseline MRI (n = 394) and CSF biomarkers (n = 193). RESULTS: The selected EF candidate added category fluency (ADAS Plus EF), and the selected EF and FA candidate added category fluency, Digit Symbol, Trail Making, and five items from the Functional Assessment Questionnaire (ADAS Plus EF&FA). The ADAS Plus EF& FA performed as well as or better than traditional ADAS-Cog scores. CONCLUSION: Adding EF and FA items to the ADAS-Cog may improve responsiveness among people with MCI without impairing validity.