Predicting Progression to Parkinson's Disease Dementia Using Multivariate Normative Comparisons

Objective: Parkinson’s disease with mild cognitive impairment (PD-MCI) is a risk factor for progression to PD dementia (PDD) at a later stage of the disease. The consensus criteria of PD-MCI use a traditional test-by-test normative comparison. The aim of this study was to investigate whether a new multivariate statistical method provides a more sensitive tool for predicting dementia status at 3- and 5-year follow-ups. This method allows a formal evaluation of a patient’s profile of test scores given a large aggregated database with regression-based norms. Method: The cognitive test results of 123 newly diagnosed PD patients from a previously published longitudinal study were analyzed with three different methods. First, the PD-MCI criteria were applied in the traditional way. Second, the PD-MCI criteria were applied using the large aggregated normative database. Last, multivariate normative comparisons (MNCs) were made using the same aggregated normative database. The outcome variable was progression to dementia within 3 and 5 years. Results: The MNC was characterized by higher sensitivity and higher specificity in predicting progression to PDD at follow-up than the two PD-MCI criteria methods, although the difference in classification accuracy did not reach statistical significance. Conclusion: We conclude that MNCs could allow for a more accurate prediction of PDD than the traditional PD-MCI criteria, because there are encouraging trends in both increased sensitivity and increased specificity