Abstract
BACKGROUND: Cognitive impairment is common in Parkinson's disease (PD), with 80% cumulatively developing dementia (PDD). OBJECTIVE: We sought to identify tests that are sensitive to change over time above normal ageing so as to refine the neuropsychological tests predictive of PDD. METHODS: Participants with newly diagnosed PD (n = 211) and age-matched controls (n = 99) completed a range of clinical and neuropsychological tests as part of the ICICLE-PD study at 18-month intervals over 72 months. Impairments on tests were determined using control means (<1-2SD) and median scores. Mild cognitive impairment (PD-MCI) was classified using 1-2SD below normative values. Linear mixed effects modelling assessed cognitive decline, while Cox regression identified baseline predictors of PDD. RESULTS: At 72 months, 46 (cumulative probability 33.9%) participants had developed PDD; these participants declined at a faster rate in tests of global cognition, verbal fluency, memory and attention (p < 0.05) compared to those who remained dementia-free. Impaired baseline global cognition, visual memory and attention using median cut-offs were the best predictors of early PDD (area under the curve [AUC] = 0.88, p < 0.001) compared to control-generated cut-offs (AUC = 0.76-0.84,p < 0.001) and PD-MCI (AUC = 0.64-0.81, p < 0.001). Impaired global cognition and semantic fluency were the most useful brief tests employable in a clinical setting (AUC = 0.79, p < 0.001). CONCLUSION: Verbal fluency, attention and memory were sensitive to change in early PDD and may be suitable tests to measure therapeutic response in future interventions. Impaired global cognition, attention and visual memory were the most accurate predictors for developing a PDD. Future studies could consider adopting these tests for patient clinical trial stratification.