Risk factors associated with severe progression of Parkinson’s disease: random forest and logistic regression models

Frontiers in Neurology, Journal Year: 2025, Volume and Issue: 16

Published: April 7, 2025

Parkinson's disease (PD) is a neurodegenerative disorder with significant variability in progression. Identifying clinical and environmental risk factors associated severe progression essential for early diagnosis personalized treatment. This study evaluates the performance of Random Forest (RF) Logistic Regression (LR) models forecasting major PD We performed retrospective analysis 378 patients (aged 40-75 years) at 2 years follow-up. The dataset included patient demographics, features, medication history, comorbidities, exposures. data were randomly split into training group (70%) validation (30%). Both RF LR trained on set, was assessed through accuracy, sensitivity, specificity, Area Under Curve (AUC) derived from ROC analysis. identified similar progression, including older age, tremor-dominant motor subtype, long-term levodopa use, comorbid depression, occupational pesticide exposure. model outperformed model, achieving an AUC 0.85, accuracy 82%, sensitivity 79%, specificity 85%. In comparison, had 0.78, 76%, 74%, 79%. showed that while both could distinguish between slow rapid stronger discriminatory power, particularly identifying high-risk patients. provides better predictive power compared to highlights potential machine learning techniques like stratification management PD.

Language: Английский

Synergistic impact of air pollution and artificial light at night on memory disorders: a nationwide cohort analysis

BMC Public Health, Journal Year: 2025, Volume and Issue: 25(1)

Published: April 30, 2025

Air pollutants and outdoor artificial light at night (ALAN) are known health risks, with established effects on respiratory cardiovascular health. However, their impact cognitive function, particularly neurodegenerative diseases like Alzheimer's, remains poorly understood. Using data from the China Health Retirement Longitudinal Study (CHARLS) Family Panel Studies (CFPS), including 44,689 participants, memory impairment (Memrye) was defined by self-reported memory-related diseases. Cox regression models were applied to assess relationship between pollutants, ALAN exposure, Memrye. Interaction analyses evaluated combined using relative excess risk due interaction (RERI), attributable proportion (AP), synergy index (S). Biomarker stepwise causal mediation examined underlying mechanisms. significantly associated Memrye (p < 0.05), hazard ratios (HR) ranging 1.010 1.343. Synergistic observed, such as for PM2.5 ALAN, RERI, AP, S values of 0.65 (0.33, 0.97), 0.30 (0.26, 0.34), 1.43 (1.21, 1.65), respectively. showed significant correlations glucose, cholesterol, uric acid, while negatively glucose acid. Mediation indicated that PM2.5, NO2, indirectly affected through biomarkers, accounting 1.07-8.28% total effects. pollution exposure linked impairment, potentially amplifying risk. Biomarkers play a key role in mediating these effects, suggesting need targeted public measures mitigate environmental risks.

Language: Английский