Objective: Examine the utility of latent class analysis (LCA) and growth mixture modeling (GMM) to elucidate cognitive phenotypes and longitudinal trajectories of cognitive change in non-demented Parkinson’s disease (PD) patients.

Background: Methods to detect early cognitive decline in PD and their value in predicting progression are needed. Quantitative methods such as LCA offer an objective approach to identify homogenous subgroups of impairment and GMM may elucidate unique longitudinal trajectories of change over time.

Methods: LCA was applied to eight neuropsychological measures to identify cognitive subtypes in 199 non-demented PD patients’ baseline assessments. Two measures from four cognitive domains were analyzed: executive function, memory, visuospatial, and language. Group differences in demographics, everyday cognitive functioning, motor symptom severity, and expert consensus cognitive diagnosis were examined. GMM will be utilized to examine distinct longitudinal trajectories up to a 5-year period.

Results: LCA identified 3 distinct groups at baseline: (1) intact cognition group (n=109; 54.8%); (2) amnestic group (n=64; 32.1%); and (3) mixed impairment group (n=26; 13.1%). The amnestic group showed impaired recall and recognition on a verbal memory task. The mixed impairment group had difficulty on measures of verbal fluency, visuoconstruction, and delayed free recall on a memory task, but intact recognition memory. Both impaired groups had significantly lower scores on ratings of everyday cognitive functioning and greater motor symptoms than the cognitively intact group. Of those with a consensus diagnosis of cognitively normal (n=151), LCA classified 35 (23.2%) patients as amnestic and 15 (9.9%) as mixed impairment. Participants have longitudinal data up to a 5-year period, and GMM will be utilized to examine longitudinal trajectories of change among the three LCA-derived subgroups.

Conclusions: Non-demented PD patients exhibit distinct neuropsychological profiles. One-third of patients with impairment as determined by LCA were diagnosed as cognitively intact by expert consensus, indicating that classification using a statistical algorithm may assist in detection of very early, subtle changes. Use of GMM will further inform the utility of these LCA-derived groups in predicting long-term cognitive decline.

« Back to 2017 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/use-of-latent-class-analysis-and-growth-mixture-modeling-to-examine-longitudinal-neurocognitive-trajectories-in-non-demented-parkinsons-disease/