A simplified algorithm may lead to overestimate dementia in PD. A clinical and epidemiological study using criteria for PD-D proposed by the Movement Disorders Task Force.

Making an accurate diagnosis of dementia in Parkinson's disease (PD-D) patients is a challenge that neurologists will have to face in the coming years. In 2007, a Task force of the Movement Disorders Society proposed operational diagnostic criteria for the diagnosis of PD-D, consisting of step I and step II. We assessed the validity of step I with reference to the diagnosis made after a formal neuropsychological evaluation and by applying the current gold standard for the diagnosis of PD-D (DSM IV). Step I had a sensitivity of 78% and a specificity of 95.5%. Step I displayed a positive predictive value of 70%, a negative predictive value of 97%, and an accuracy of 93.4%. The clinimetric properties observed in our setting suggest that step I may be considered as a good screening tool (negative predictive value of 97%); however, using step I alone to make a diagnosis of PD-D may lead to an overestimation of dementia in PD, particularly in patients with considerable dysexecutive deficits (positive predictive value of 70%). In conclusion, formal neuropsychology and longitudinal follow-up are still required for the diagnosis and categorization of dementia in PD.