DATA-DRIVEN IDENTIFICATION OF DISTINCT COGNITIVE SUBTYPES IN PARKINSON’S DISEASE DEMENTIA: ASSOCIATIONS WITH CLINICAL, GENETIC, AND BIOMARKER CHARACTERISTICS
- Michel J. Grothe, Spain
Abstract
Aims
To conduct a data-driven exploration of distinct cognitive profiles among patients with Parkinson’s disease dementia (PDD), and characterize these cognitive subtypes with respect to clinical and genetic characteristics as well as biomarkers of Alzheimer’s disease (AD).
Methods
From the Parkinson’s Progression Markers Initiative (PPMI) cohort we identified all individuals with Parkinson’s disease and dementia (N=74). Of these, 51 (69%) were initially enrolled as de novo idiopathic PD and 23 (31%) had a confirmed LRRK2, GBA, or SNCA mutation. Ward’s hierarchical clustering was applied to domain-specific neuropsychological test data, and the derived PDD subgroups were then compared on clinical characteristics, prevalence of genetic mutations, as well as APOE4 genotype and CSF AD biomarkers (available for 44 and 59 patients, respectively).
Results
Clustering identified two distinct PDD subgroups with differing cognitive profiles (Fig. 1). Subgroups did not differ in overall dementia severity (MoCA: 19.7 vs 21.4, p=0.15) and were equally impaired in attention/executive functions, but Cluster-A (N=52, 70%) showed a more pronounced memory impairment, whereas Cluster-B (N=22, 30%) was more selectively impaired in visuospatial skills. Subgroups did not differ in duration or severity of motor symptoms, but Cluster-B trended towards an earlier age at PD onset (60.5 vs 65.3 years, p=0.07) and had a higher proportion of GBA mutation carriers (27% vs 4%, p=0.007). Subgroups did not differ in APOE4 prevalence or CSF amyloid-β or tau levels.
Conclusions
We identified distinct memory-predominant and visuospatial-predominant dementia presentations among PDD patients. These distinct cognitive profiles do not appear to be driven by comorbid AD pathology.