Clinical heterogeneity in Parkinson's disease revisited: a latent profile analysis.

ABSTRACT

OBJECTIVE: The heterogeneity of Parkinson's disease (PD) is increasingly recognized, and several attempts have been made to subclassify subjects on clinical or cognitive features. We explored the utility of latent profile analysis (LPA) as a means of classifying patients with PD on clinical features and test validity of these subclasses against neuropsychological data. METHODS: LPA utilizing clinical variables while controlling for age was applied to a cohort of 71 outpatients with PD. The resultant subgroups were validated via comparison to 30 control subjects on neuropsychological tests of executive, memory, and visuospatial functions. RESULTS: The LPA resulted in a three-class solution identifying a 'younger onset, mild motor impairment group', a 'moderate motor impairment group', and an 'old onset, fast progression group'. The groups were distinguishable on cognitive variables with the 'younger onset mild motor impairment subgroup' displaying deficits pertaining verbal acquisition, visuospatial construction, and set maintenance. The 'moderate motor impairment group' exhibited widespread cognitive impairment, and the 'old onset, fast disease progression group' had extensive cognitive impairment but outperformed the former group on verbal acquisition and visuospatial function. CONCLUSION: LPA holds promise in PD research as it uncovered three PD subtypes distinguished by motor symptoms and disease progression and validated by cognitive variables.