Neuropsychology, neuroimaging or motor phenotype in diagnosis of Parkinson's disease-dementia: which matters most?

Parkinson's disease (PD) is a neurodegenerative disorder affecting not only the motor system but also the cognitive and behavioral domains. Although there are many studies addressing the issue of cognition, a universally recognized method to diagnose patients with dementia is still lacking. The aim of this study was to determine which neuropsychological test is the most reliable in the diagnosis of dementia in PD and to establish if mini mental state examination (MMSE) is enough to detect this condition. We studied 200 consecutive PD patients through an extensive neuropsychological battery, clinical evaluation and brain magnetic resonance imaging over a period of 4 years. A logistic regression model was used to evaluate the interplay between possible risk factors and the accuracy of different neuropsychological tests. PD patients with dementia performed significantly worse in all the tests as compared to patients with PD alone: phonological verbal fluency, attentional matrices, Rey auditory verbal learning test and digit span were the most useful tools. Age and disease duration were correlated with cognitive impairment. No relevant differences were detected in phenotype, affected body side at onset, levodopa equivalent daily dose or neuroimaging findings (except for the occurrence of atrophy). Despite reasonable accuracy of MMSE (90%), its positive predictive value is only 74%. Using at least 3 neuropsychological tests, among those more significant detected with logistic regression analysis, the positive predictive value rises to 91%. In conclusion, the use of an extensive neuropsychological battery is still recommended in the diagnosis of dementia in PD.