Data-driven disease progression model of Parkinson's disease and effect of sex and genetic variants

Abstract

As Parkinson's disease (PD) progresses, there are multiple biomarker changes, and sex and genetic variants may influence the rate of progression. Data-driven, long-term disease progression model analysis may provide precise knowledge of the relationships between these risk factors and progression and would allow for the selection of appropriate diagnosis and treatment according to disease progression. To construct a long-term disease progression model of PD based on multiple biomarkers and evaluate the effects of sex and leucine-rich repeat kinase 2 (LRRK2) mutations, a technique derived from the nonlinear mixed-effects model (Statistical Restoration of Fragmented Time course [SReFT]) was applied to datasets of patients provided by the Parkinson's Progression Markers Initiative. Four biomarkers, including the Unified PD Rating Scale, were used, and a covariate analysis was performed to investigate the effects of sex and LRRK2-related mutations. A model of disease progression over ~30 years was successfully developed using patient data with a median of 6 years. Covariate analysis suggested that female sex and LRRK2 G2019S mutations were associated with 21.6% and 25.4% significantly slower progression, respectively. LRRK2 rs76904798 mutation also tended to delay disease progression by 10.4% but the difference was not significant. In conclusion, a long-term PD progression model was successfully constructed using SReFT from relatively short-term individual patient observations and depicted nonlinear changes in relevant biomarkers and their covariates, including sex and genetic variants.