Assessing plasma levels of selenium, copper, iron and zinc in patients of Parkinson's disease

Abstract

Trace elements have been recognized to play an important role in the development of Parkinson's disease (PD). However, it is difficult to precisely identify the relationship between these elements and the progression of PD because of an insufficient number of patients. In this study, quantifications of selenium (Se), copper (Cu), iron (Fe) and zinc (Zn) by atomic absorption spectrophotometry were performed in plasma from 238 PD patients and 302 controls recruited from eastern China, which is so far the largest cohort of PD patients and controls for measuring plasma levels of these elements. We found that plasma Se and Fe concentrations were significantly increased whereas Cu and Zn concentrations decreased in PD patients as compared with controls. Meanwhile, these four elements displayed differential changes with regard to age. Linear and logistic regression analyses revealed that both Fe and Zn were negatively correlated with age in PD patients. Association analysis suggests that lower plasma Se and Fe levels may reduce the risk for PD, whereas lower plasma Zn is probably a PD risk factor. Finally, a model was generated to predict PD patients based on the plasma concentrations of these four trace elements as well as other features such as sex and age, which achieved an accuracy of 80.97±1.34% using 10-fold cross-validation. In summary, our data provide new insights into the roles of Se, Cu, Fe and Zn in PD progression.

Figure 1. Plasma trace element concentrations in three age groups of PD patients and controls.

The number of subjects with age ≥55, 55~65, ≤65 were 38, 53, 147 respectively for PD patients, and 71, 59, 172 respectively for controls. Box charts represent the distribution of plasma Se (A), Cu (B), Fe (C) and Zn (D) concentrations. Data were analyzed with Mann-Whitney U-test and significant difference was marked with * for p <0.05.

Figure 2. Relationship between plasma trace element levels and age.

The scatters represent the correlation of plasma Se (A and B), Cu (C and D), Fe (E and F) and Zn (G and H) levels with age in PD patients and controls. The parameters showed in the figure represent the performance of the linear regression (red dash line).

Figure 3. Element-element ratios for age groups and clinical subtypes.

Element-element ratios for three age groups in PD patients (age ≤55, 55~65 and ≥65) were compared with their respective age-matched controls (A), and for clinical subtypes in PD patients were compared with total controls (B). Histograms represent the fold changes of each element-element ratio compared to the mean value of their respecitve controls which is set as 1. Data were analyzed with Mann-Whitney U-test and significant difference was marked with * for p <0.05.

Figure 4. Performance of 10-fold cross-validation SVM model.

The generalized performance of the SVM model. We rebuilt the model for 100 times for the validation using AUC, accuracy, sensitivity and specificity.