ABSTRACT
Parkinson's disease (PD) is an oxidative stress-linked neurodegenerative disorder, with the highest prevalence among seniors. The objective of this study were: (1) to analyse levels of following oxidative stress parameters: total antioxidant capacity (TAC), uric acid (UA), total glutathione (tGSH), bilirubin (Bil) and albumin (Alb), in blood of PD patients and healthy controls; (2) to find possible associations of examined oxidative stress parameters with PD subtypes and levodopa treatment status; and (3) to evaluate power and relevance of the aforementioned oxidative stress parameter for the prediction of onset and progression of PD by utilizing Random Forest machine learning (RFML). Oxidative stress parameters were determined in 125 PD patients and 55 healthy controls. Evaluated with frequentist statistics, our data revealed that UA is the only oxidative stress parameter associated with PD. However, when the PD cohort was divided in gender-dependent manner, tGSH and Bil were also significantly associated with PD in subgroup of female patients. RFML rendered no predictive power of any of the tested oxidative stress parameters in respect to PD, its subtypes, and/or status of levodopa treatment. In conclusion, despite the positive association of UA with PD (in complete cohort of PD patients) and of tGSH and Bil with PD but only in female patients, these oxidative stress parameters are of no use in clinical practice due to the lack of the predictive/diagnostic power.
