Abstract
The Mattis Dementia Rating Scale (DRS) is widely used to assess cognition and screen for dementia in Parkinson's disease (PD). In 2008, Llebaria and colleagues reported excellent sensitivity and specificity detecting dementia in PD using a DRS total score cutoff of 123. However, this study used a sample with rather low mean educational achievement (8.9 years). The 123 cutoff score was cross-validated in a sample of 51 PD patients with a rather high mean educational achievement (14.84 years). A total score cutoff of 123 correctly classified 60.7% of the sample, with 20% sensitivity and 100% specificity. Use of age- and education-corrected scaled scores improved classification accuracy, but these cutoffs corresponded to the 25th percentile. Optimal classification of 82.4% accuracy was achieved using a raw total score cutoff of 133, corresponding roughly to the 16th percentile. Neither age nor education evinced a strong correlation to DRS total score in our sample. Results suggest caution in applying the DRS-2 total score cutoff of 123 and highlight the importance of comprehensive neuropsychological evaluations for clinical diagnosis of dementia in PD.