Neuropsychological assessment as a predictor of dementia conversion in Alzheimer’s disease

Abstract

AbstractBackgroundProgressive cognitive decline is a characteristic of Alzheimer’s disease (AD). Identifying reliable predictors of conversion to dementia in older adults, particularly those with mild cognitive impairment (MCI), is critical for clinicians, patients, and patient families.Method1083 adult participants (Age: M = 73.20, SD = 6.92) were included in this study from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). The average follow up was over 5 years (M = 65.25, SD = 40.16). Cox proportional hazards regression models were implemented to estimate the risk of dementia. Models were built sequentially to examine changes in predictive contributions when including biomarkers and cognitive predictors.Result57% of participants were classified as MCI and 43% as cognitively normal (CN). By the end of study, there is an overall conversion rate of 24% from the participants with cognitive normal and MCI to dementia. Participants with MCI demonstrated significantly higher rates of conversion than those with CN (F(1,258) = 92.51, p < .001, η2 = .26). The initial model indicated greater age (HR = 1.03, p = .004), fewer years of education (HR = 0.95, p < .027), the presence of at least one APOE ε4 gene (HR = 3.23, p < .001), and higher levels of neurofilament light chain (NFL) (HR = 1.28, p < .001) predicted conversion to dementia. A second model indicated higher levels of Ab42 was predictive of lower risk (HR = 0.41, p < .001), whereas higher levels of P‐tau was predictive of greater risk (HR = 1.48, p < .001). APOE ε4 presence and NfL levels remained significant predictors. Finally, composites of memory and executive scores were added. Memory performance at baseline was highly associated with dementia risk (HR = 0.26, p < .001), with poorer performance at baseline predicting greater likelihood of conversion. Similarly, poorer performance on executive functioning tasks was predictive of dementia risk (HR = 0.57, p < .001).ConclusionComposite scores of memory and executive functions are strong predictors of conversion to dementia. Such composites are approximately as predictive as Ab42, and more predictive than demographics, APOE ε4 presence, baseline P‐tau and NFL.