Cerebrovascular accident and essential and toxic metals: cluster analysis and principal component analysis

Abstract

Background

Cerebrovascular accidents are known as a great cause of morbidity and mortality worldwide. Although there are known risk factors for ischemic stroke, the cases that cannot be justified with these risk factors are increasing. Toxic metals as a potential risk factor for other diseases in humans are assessed in this study in the CVA group and compared to controls.

Method

70 participants (35 each group) have been selected for this study. The group with recent medical history of documented CVA and a control non-CVA group. The serum level of several metals has been assessed using Inductively coupled plasma mass spectrometry (ICP-MS) method. principal components and cluster analyses were employed to compare toxic metal toxicity between the groups.

Results

Cu (p < 0.001) and Pb (p = 0.002) levels were significantly higher in the CVA group whereas Ni (0.003) were significantly lower. There was no significant difference between the smoking (p = 0.56) and opium (p = 0.46) use between these groups. Most of the essential metals were positively correlated with each other in both groups (Ni with Fe, Zn; Fe with Zn with r over 0.6). there was also PCA and CA are crucial in and cluster analysis in which Ni, Fe, and Zn were most similarly correlated in both groups.

Conclusion

we found a complex interaction between toxic metals in the healthy and CVA human body. Due to the lack of data on in vivo interaction of these metals even in healthy individuals, further investigation is needed to evaluate the exact mechanism of such relations.

Introduction

Also known as stroke, a Cerebrovascular accident (CVA) is a clinical emergency that consists of clinical symptoms that have vascular origin and can lead to loss of brain function [1]. Depending on the location the signs and prognosis of CVA can be varied from weakness of extremities to loss of memory, speech, or movement function [2]. CVA is further divided into ischemic and hemorrhagic stroke in which ischemic is more dominant by accounting for 80% of CVAs [1].

CVA is considered as most common cause of disability worldwide and the fourth cause of death globally accounting for 8% of deaths internationally [1, 3]. Based on the World’s Stroke Organization, 13.7 million new cases of stroke are diagnosed and the prevalence of CVA is estimated between 2 and 15% depending on the age [4, 5].

Based on The TOAST (Trial of ORG 10172 in Acute Stroke Treatment) system, cardio-embolism followed by other determined pathogenesis and undetermined pathogenesis or more common in young patients and small-vessel occlusion is more common in the older population [6]. The risk factors of ischemic stroke are well-known and widely discussed as they are similar to cardiovascular disease risk factors in which hypertension, dyslipidemia, diabetes, smoking, and obesity can be mentioned [1]. Unlike other causes, these risk factors are modifiable and can be prevented by the patient. studies have suggested other causes for ischemic stroke that are not as widely discussed and need more investigation.

Toxic metals as a potential pollutant can be found in almost every aspect of human life such as air, water, soil, or even food. Elements weighing more than 5 g cm³ are considered toxic metals. toxic metals are further divided into essential and non-essential elements. Essentials are Molybdenum (Mo), Manganese (Mn), Copper (Cu), Nickle (Ni), iron (Fe), and Zinc (Zn) which are responsible for maintaining the human body function and are needed, and non-essentials like Cadmium (Cd), Arsenic (As), lead (Pb), mercury (Hg) which the smallest amount are toxic. Recently, toxic metals have been proposed to be a worldwide concern due to the increase in urbanization, and the use of materials that contain such metals and pesticides [7].

Studies indicate various effects of these metals on the cerebrovascular system. For instance, Hg showed to increase oxidative stress, and production of reactive oxygen species (ROS) [8]. Pb can damage the vessels’’ endothelium and interfere with cerebral brain flow [9]. Cd causes the proliferation of arteries’ walls and is associated with an increased risk of cardiovascular diseases and CVA [10, 11].

On the other hand, in vivo, metal-metal interactions have been underestimated and not been investigated adequately. Limited findings state Cd administration caused alteration in Zn and Cu levels, Fe is increased in several neurodegenerative diseases and Cu-induced DNA damage can be antagonized by other metals such as Ni, Cd, Zn, and Pb [12, 13].

Given the risks and side effects associated with toxic metals, our objective is to analyze the correlation of some important and available toxic metal levels (Fe, Co, Ni, Cu, Zn, Mn, Pb, Hg) in cerebrovascular accident patients using cluster analysis. Additionally, we explored the interaction between toxic and essential elements within these groups through cluster analysis. This attempt can help better understanding of the role of metals in the CVA and prevention of their sources. With cluster analysis we tend to better comprehend the relation between these metals and their interactions in human body in both control and CVA patients.

Method

Study population

In this case-control study, 35 Cerebrovascular Accidents with medical record and 35 individuals with no history of Cerebrovascular Accident controls were enrolled as the case group. The subjects of the case group were selected from among the hospitalized patients with Cerebrovascular Accidents in Vali Asr Hospital in Birjand. The subjects of the control group were selected from people who had no history of stroke, heart attack, high blood pressure, and diabetes based on both medical records and personal interview. The attack were in cases were three months or less prior to the sample collection. The two groups were matched for age, gender, and occupation. Once the study objectives and confidentiality of the obtained data were explained to the participants, they signed written consent forms for participation and the consent form were taken. Demographic data, including age, gender, education, marital status, occupation, opium use, and history of smoking were collected from the subjects.

Baseline examination

In this study, CVA and non-CVA individuals after consent for participating and donating blood samples, an individual’s blood was taken from an antecubital vein by venipuncture. 20 milliliters of blood were taken from each patient according to the protocol. The samples were centrifuged in room temperature within the first 30 min, and serum was collected.

The serum blood samples underwent digestion through a combination of nitric acid and perchloric acid, mixed in a volumetric ratio of 2:1. For the acid digestion procedure, 5 cc of each serum blood sample was transferred into 25-ml glass test tubes. Subsequently, 2 cc of 65% pure nitric acid (Merck, Germany) was added to each serum blood sample, and the mixture was allowed to digest gradually at room temperature overnight. Following this initial step, 1 cc of 72% perchloric acid (Merck, Germany) was introduced into the blended specimens. The resulting mixture underwent a 4-hour digestion process in a hot water bath (Bain-Marie) set at 98 °C until complete digestion was achieved. After the digestion process was completed, the samples were cooled to ambient temperature, and subsequent dilution was carried out using 25 ml of deionized water. Finally, the samples, prepared for toxic metal analysis, were measured utilizing an Agilent 7900 ICP-MS.

All standard solutions employed for metal analysis in this investigation were derived from Merck standards with a concentration of 1000 ppm. The concentrations of toxic metals (Pb, Cu, Ni, Cr, Co, Fe, Hg, Mn, and Zn) in this study were expressed in micrograms per deciliter. The ICP-MS performance parameters were set as follows: radiofrequency power—1.5 kW; plasma gas flow rate—15 l per minute; carrier gas flow—1.01 l per minute; constituent gas—0.15 l per minute; sample absorption rate—1.7; sample depth—10 mm; detector mode—auto; scan type—peak hopping (three sweeps per reading and three readings per repetition); and scan number—3.

The linearity of the calibration curve was assessed by examining the coefficient of determination (R²) in the linear regression model, with a value greater than 0.99 indicating acceptable linearity. The limits of quantification (LOQ) and limits of detection (LOD) were determined using the following formulas:

$${\text{LOQ = }}\left( {{\text{Standard Deviation / Slope}}} \right){\text{ * 10}}$$

$${\text{LOD = }}\left( {{\text{Standard Deviation / Slope}}} \right){\text{ * 3}}{\text{.3}}$$

For metals, the determined LOD was 1.3 µg/L, and the LOQ was 3.8 µg/L, with an accuracy of 90%. To ensure accuracy, ten repeated measurements of control samples were performed within a short time frame while maintaining consistent experimental conditions. The relative standard deviation (RSD) of these repeated measurements was 3.82%.

Statistical analysis

The characteristics of the study population were presented using means (standard deviation: SD) and numbers (%) for categorical variables. For covariates with a skewed distribution, medians (interquartile range: IQR) were reported. A comparison of baseline characteristics between individuals with and without CVA utilized the Student’s t-test for continuous variables, the chi-square test for categorical variables, and the Mann-Whitney test for skewed variables.

Pearson and Spearman correlation coefficients were utilized to examine the relationships between toxic metal concentrations within each group. Additionally, cluster analysis and principal component analysis (PCA) were conducted to compare the toxicity of toxic metals across the groups. PCA was further employed to identify potential sources of metals across various matrices. The validity of PCA was confirmed through Bartlett’s test of sphericity (P < 0.001) and Kaiser-Meyer-Olkin (KMO) values greater than 0.5. Comparisons of toxic metal toxicity between smokers and non-smokers were carried out using PCA with varimax rotation. Furthermore, Ward’s method was applied to categorize the data into three distinct clusters, revealing different associations among toxic metal concentrations.

Results

Comparison of demographic information between CVA and non-CVA groups

Among 70 participants people (62 were male) with a mean age of 48.25 ± 12.45 years. The results of the demographic information reported in Table 1 indicate that the two groups exhibit similarities. No significant variable was found in this table as opium use and smoking were not significantly different in the two groups. This fact is considerable because it shows that differences in metal levels in CVA and non-CVA groups are less likely to be associated to be due to smoking or opium use. Occupation status was also not statistically different between the two groups.

Table 1 Comparison of demographic information between CVA and non-CVA groups

Comparison of toxic metal levels between CVA and non-CVA groups

The comparison data illustrating the levels of toxic metals in two distinct groups is presented in Table 2. When comparing the fundamental characteristics of the two groups, we observed that individuals with CVA, who were also older, exhibited higher levels of Cu and Pb and lower levels of Ni compared to those without CVA with Cu having the strongest difference between the two groups.

Table 2 Comparison of toxic metal levels between CVA and non-CVA groups

The association between toxic metals CVA and non-CVA groups

In the CVA group, there was a significant positive correlation between Fe level and Ni (r = 0.83), Zn (r = 0.94), Mn (r = 0.91), and Cu levels (r = 0.88) (Fig. 1A). positive correlation was also found in Mn and Ni (r = 0.91), Cu (r = 0.7), and Zn (r = 0.8). also, Zn and Cu (r = 0.97), and Ni (r = 0.76) were positively correlated. Lastly, Ni and Cu (r = 0.69) were positively associated (Fig. 1A).

Fig. 1

A The relationship between toxic metals in the CVA group. B The relationship between toxic metals in the control group

In the control group, there was a significant positive correlation between serum Fe and Hg (r = 0.64), Zn (r = 0.76), and Ni (r = 0.83). Ni was also positively associated with Hg (r = 0.67), Zn (r = 0.81), and negatively with Mn (r=-0.55) (Fig. 1B).

Principal component analysis (PCA)

The results of the CVA group indicated that the three components of PC1-PC3 cumulatively accounted for 83.40% of the total variance, reflecting the major part of the results. The value of the KMO index in PCA in the CVA group was 81.0%.

In the first component (PC1; Eigenvalue: 4.41), which accounted for 55.1% of the total variance, the main factors for PC1 with significant loading were Fe, Ni, Cu, Zn, and Mn. The second component, which explained 15.6% of the total variance, included Co (0.87) and Hg (0.87) with the Eigenvalue of 1.25. The third component (12.7% variance explained) included Pb (7.99%) (Fig. 2A).

Fig. 2

A Principal component analysis for toxic metals in the CVA group. B Principal component analysis for toxic metals in the control group

Cluster analysis

In the CVA group, cluster analysis was performed on the elemental concentrations, the results of which are presented in dendrograms. The degree of association between the metals is shown by the distance between clusters. Our analysis yielded three clusters, including A1 (Cu, Zn), A2 (Ni), A3(Fe, Mn), B (Pb) and C (Co and Hg). Furthermore, cluster analysis produced similar results as PCA regarding toxic metals (i.e., Cd, Cr, and Pb), indicating a common source. Cluster A included Mn and Cu, which are essential metals for the body and were placed in one stratum (Fig. 3A).

Fig. 3

A Cluster analysis for toxic metals CVA group. B Cluster analysis for toxic metals in the control group

Factor analysis was performed for dimensionality reduction, that is, to decrease multiple variables to fewer ones. The results of the control group indicated that three components accounted for 86.52% of the total variance. In other words, the first three components had the most significant impact on the results. The value of the KMO index in PCA in the control group was 81.0%.

Two principal components (eigenvalue > 1) emerged, which explained more than 77.7% of the cumulative variance. The PC1 included Fe, Zn, Ni, and Hg, which explained 43.61% of the total variance. PC2 consisted of Co, Cu, and Pb accounting for 23.65% of the total variance. The element Mn belonged to PC3, explaining 10.49% of the variance (Fig. 2B).

In addition, Fig. 3B shows that the elements in the control group samples were grouped into Four main clusters, including A (Fe, Zn, Ni, and Hg), A1 (Hg), A2 (Zn), A3(Fe, Ni), B (Mn), C (Cu, Pb) and D (Co). Although there were some differences in the results of CA and PCA, similar distribution patterns were observed between PC 1 and cluster A, and PC3 and cluster B. In addition, Cu and Cr in cluster C1 and Cd in cluster C2 were highly associated with PCA consisting of Cu, Cr, and Cd.

Discussion

In this study, we achieved notable findings that have not been discussed before. Pb and Cu serum levels were significantly higher in the CVA group and Ni was significantly lower. Pb is a considered as toxic metal that cannot be found normally in human cell structures there so any concentration of Pb is external [14]. As one of the oldest occupational toxins has wide effects on the human body but mainly discussed here the effect of Pb on the cerebrovascular system is significant. Peripheral motor neuropathy and increased systolic blood pressure were correlated with higher levels of blood Pb [15]. Cu on the other hand is an essential metal with potential or known toxicity that is necessary for human body function. Cu is a cofactor for several enzymes and a structural component and energy metabolism, Fe metabolism, and antioxidative defense rely on Cu [16]. The blood-brain barrier (BBB) has been known as the major route of Cu transport and blood- the cerebrospinal fluid barrier (BCB) contributes to the maintenance of the Cu concertation. after entering the parenchyma (in which grey matter has a higher level of Cu), Cu is released into the cerebrospinal fluid (CSF) [16]. However, the white matter is more prone to experience ischemic stroke. Additionally, glial cells are known to be the storage of Cu in the human brain and play a crucial role in its metabolism [16]. Also during cerebral reperfusion after ischemic stroke, it has been found that there is an elevation in copper ion levels and increase in copper-induced cell death in neurons [12]. Concerning Ni, most of the findings suggest that a higher level of Ni is associated with an increase in oxidative stress, brain tissue damage, and altered neuronal microarchitecture [17]. Though Our findings state a lower level of Ni in the CVA group the cause of this finding needs further evaluation and confirmation.

In the second part, the interaction between metals is discussed. Though it’s a complex network of relations some relations stand out. For instance, Zn and Ni were shown to be positively correlated in both CVA and non-CVA groups. Previous findings indicated that Ni-induced glutathione and lipid peroxidation in brain cells of mice can be treated by Zn administration [18]. This antioxidant and stabilizer of biomolecule’s property might be the case in that protects the body from toxic metals [18].

Most of the relations were between essential metal with potential or known toxicity (Zn, Cu, Fe, Ni, and Mn). This fact indicates the delicate interaction between such metals and the effect of intake of these metals. Usually, these metals come together in our diet and therefore these positive correlations are available in both groups but their positive effect on CVA is not that significant as we found out the only element that is significantly lower in the CVA group is Ni. As previous findings indicated Ni deficiency decreases rat’s growth and decreases the hematocrits [19].

Cluster analysis and PCA are methods by which we can simply understand the positive or negative effects of metals on the human body. These analyses indicate which metals are more similar in the human body in a matter of protecting or harming the systems. In this study, we found Fe, Zn, and Ni have similar protective effects on the human body. All of these are well-known essential metal with potential or known toxicity that are crucial for cell survival [19, 20].

Pb effect on the control group was similar to Cu as they both were significantly lower in the control group. Previous studies have also stated the significant negative effect of Pb on ischemic stroke [21, 22]. Another meta-analysis found that serum Cu can be known as a risk factor for ischemic stroke can induce an atherosclerotic process and is available in superoxidase dismutase and monoamine oxidase [23]. It Is hypnotized in the state of a CVA due to the increase of superoxide and inflammation, or the increase of catecholamine-by-catecholamine synthesis which contains copper might be the case [23]. This finding is crucial and a pivotal fact, because it shows that an increase in some metals in the state of pathologies such as CVA should not always be considered a risk factor till the exact mechanism is understood. In this example, the Cu level has been increased by an increased level of an enzyme that contains it and it’s not associated which the accordance of the pathophysiology of CVA. Although there are various mechanisms known for Cu that make this metal a risk factor for CVA [23].

These findings of metal interactions are notable and each of them needs further investigation and confirmation as they are very complex to interpret. Here in our study, we gave a glimpse of what these interactions are capable of but due to the lack of previous data on these metals’ interactions even in healthy human beings rather than pathologic states such as CVA, we were limited on interpretation and justification of such findings.

Conclusion

In this study, we found Cu and Pb levels significantly higher in the CVA group. We also found strong positive correlation Zn, Cu, Fe, Mn and Ni with each other in pair in CVA group. These correlations were not consisted in control group as in control group Ni and Mn were negatively corelated and Hg showed positive correlation with Zn, Ni and Fe. Also, by cluster analysis we found Cu level in CVA group were more related to other essential metals than in control group. Hg as toxic metals were in a same cluster as essential metals such as Zn, Fe and Ni in control group despite the distinct place to these metals in CVA group. principal component analysis also indicated a same PCA Ni, Fe and Zn were in same PCA in both CVA and control groups. The relation between these measured metals in both control participants and CVA patients give us a new insight of the interplay of these metals in both normal and pathological physiologies.