Increased serum levels of cadmium are associated with an elevated risk of cardiovascular disease in adults

Previous studies have determined the effects of exposure to certain heavy metals on cardiovascular disease (CVD); however, the association between cadmium exposure and CVD in adults remains unclear. The relationship between serum levels of cadmium and the risk of CVD was studied by analyzing available data from 38,223 different participants of the National Health and Nutrition Examination Survey (NHANES) from 1999 to 2016. After adjusting for all covariates, we found that higher serum cadmium concentrations were positively related to both the overall risk of CVD (odds ratio (OR): 1.45; 95% confidence interval (CI): 1.22, 1.72; p for trend <0.001) and the risks of its subtypes, including congestive heart failure, coronary heart disease, heart attack, and stroke. Elevated cadmium levels were associated with increased levels of lipids and inflammatory factors, including blood triglycerides, total cholesterol, white blood cells (WBCs), and C-reactive protein (CRP). Our study provided epidemiological evidence that cadmium may increase the risk of CVD by elevating blood lipids and inflammation.


Introduction
Cardiovascular disease (CVD), which includes heart and vascular diseases such as coronary heart disease (CHD), angina, heart attack (HA), heart failure (HF), and stroke, is a leading cause of death worldwide (Go et al. 2013;Lloyd-Jones et al. 2010;Xu et al. 2020a). According to a report from the American Heart Association (AHA) in 2010, CVD mortality accounted for nearly 33% of all mortality in the USA, and one person died from CVD every 38 s. Moreover, more than 785,000 people are estimated to have new or recurrent CVD every year (Lloyd-Jones et al. 2010). Therefore, determining and controlling the risk factors for CVD are critical (Phillips andGuazzi 2015, Thiara 2015). Some new risk factors were recently identified in addition to traditional factors, such as smoking, high cholesterol, physical inactivity, obesity, and diabetes (Aw et al. 2020, Koller and Agyemang 2020, Lloyd-Jones et al. 2010. In particular, environmental pollution was found to contribute to the development of CVD and its risk factors Li et al. 2020c;So et al. 2020;Xu et al. 2020a). Of these factors, heavy metal pollutants (methylmercury, lead, chromium), which constitute one type of environmental pollutant, are associated with CVD and its risk factors (Ali et al. 2020;Cao et al. 2020;Duan et al. 2020;Orisakwe et al. 2020). However, few studies have addressed the correlation between cadmium (Cd) and CVD.
Many regions have reported levels of Cd exceeding the maximum permissible limit of 0.3 mg kg −1 established by the World Health Organization (WHO) (Li et al. 2020a;Orisakwe et al. 2015;Orisakwe et al. 2020;Pan et al. 2016;Wang et al. 2018). In addition to natural sources, various human activities can increase Cd levels, including smoking, traffic emissions, metallurgical processes, nuclear energy production, mining, coal combustion, and chemical manufacturing (Li et al. 2019;Li et al. 2018;Sall et al. 2020;Wu et al. 2019). Furthermore, similar to other heavy metals, the stability and permeation of Cd lead to its persistence and accumulation in vivo (Wang et al. 2015;Wu et al. 2019). Therefore, the relationships between Cd and many diseases have attracted considerable attention. Cd was found to increase not only the risk of carcinogenesis but also noncancer-related mortality (Al Amin et al. 2020;Amadou et al. 2020;Suwazono et al. 2020); exposure to Cd was shown to be associated with kidney function decline and development of neurodevelopmental disorders and airway inflammation (Ijomone et al. 2020;Klein et al. 2020;Sotomayor et al. 2020). In addition, Cd was associated with elevated lipid levels and atherogenic indices, which might induce CVD in susceptible people (Igharo et al. 2020;Xu et al. 2020b). Several experimental studies also showed that Cd could participate in a variety of physiopathological cardiovascular processes, such as cardiovascular system development, cellular metabolic pathway alterations, and vascular and endothelial lesions (Diaz et al. 2021;Hudson et al. 2019;Ronco et al. 2011).
However, few studies have identified the relationship between Cd and CVD. A study in a Korean population showed that Cd was associated with the risk of stroke in people under the age of 60 years. An investigation in a larger and more representative population is needed to determine the correlation between Cd levels and CVD (Jeong et al. 2020). Herein, based on a large population from the National Health and Nutrition Examination Survey (NHANES) in 1999-2016, our study investigated the serum concentrations of three heavy metals, the association between these metals and CVD, and the underlying mechanism.

Subjects
We included subjects who had participated in the NHANES, which is a series of studies involving members of the general, noninstitutionalized population in the USA. The detailed survey design, methods, and available data are accessible on the NHANES website (https://www.cdc.gov/nchs/nhanes/). Subjects who participated in the NHANES had available serum heavy metal concentrations and underwent CVD evaluation from 1999 to 2016 were enrolled in our study. In total, 92,062 adults completed the interviews and examinations. A total of 42,560 people who did not have CVD data were excluded, and those who were pregnant (n= 1,486) or had missing data on all blood metals (n=9,793) were also excluded. Figure 1 shows the participant selection process. Finally, our study enrolled 38,223 participants. The data consisted of one result from each participant.

Evaluation of outcomes
The participants were evaluated by both a standardized medical questionnaire and self-reported physician diagnoses. The participants reported whether doctors or other health professionals had ever diagnosed them with CHD, congestive HF, angina, stroke, or HA. A participant was considered to have CVD if a positive response was given to any of the relevant questions.

Exposure to Cd
Whole-blood samples were collected by a trained investigator and frozen before analysis. First, the blood samples were diluted. Then, the serum Cd levels were measured with an inductively coupled plasma mass spectrometer with dynamic reaction cell technology (ELAN® DRC II; PerkinElmer, Norwalk, CT, USA). The quality assurance and quality control protocols followed the 1988 Clinical Laboratory Improvement Act mandates (https://www.cdc.gov/nchs/ nhanes/index.htm).

Covariates
The covariates included age, sex, race, physical activity level, education level, poverty-to-income ratio (PIR), past-year alcohol consumption status, serum cotinine level category, body mass index (BMI), family history of CVD, and fish consumption; these covariates were selected based on factors that could affect the correlation between Cd levels and CVD risk (Lee et al. 2021;Li et al. 2021;Salem Ali Albar et al. 2020;Tanaka et al. 1998;Xu et al. 2020a). Age and the concentrations of lead, Cd, and mercury are shown as means and standard deviations (SDs). Other variables are showed as percentages and frequencies. Race categories included Mexican American, other Hispanic, Non-Hispanic White, non-Hispanic Black, and other race. Education levels were divided into five classes: less than 9th grade, 9th-11th grade, high school graduate/ GED or equivalent, some college or Associate's degree, and college graduate or above. Family PIR was divided into two classes. BMI category was divided into three classes: less than 25, 25 to 30, and greater than 30. Physical activity was divided into three classes: never, moderate, and vigorous. Past-year alcohol consumption, family history of CVD, and fish consumption were divided into two classes. The blood concentrations of lipids were measured by Roche Modular P and Roche Cobas 6000 chemistry analyzers and the Friedewald equation. The Beckman Coulter method was used to measure inflammation parameters.

Statistical analysis
Continuous variables are presented as means with SDs, and categorical variables are presented as frequencies and percentages. We compared continuous and categorical variables between groups with and without CVD with the Mann-Whitney U and χ 2 tests, respectively, and analyzed the correlations between the serum levels of heavy metals and the risk of CVD by survey-weighted multiple logistic regression analysis with three separate models. Model 1 was a crude model; model 2 was adjusted for age, sex, race, and education level; and model 3 was adjusted for the variables included in model 2 and BMI, PIR, physical activity, past-year alcohol consumption, serum cotinine category, history of CVD, and fish consumption. We further investigated the relationships between serum concentrations of heavy metals and five CVD subtypes. Finally, multivariate analysis was used to explore the associations between serum Cd levels and blood lipids and inflammation parameters. Sampling weights were adjusted in all statistical analyses using SAS (version 9.2) and R software (version 3.5.0). We considered a value of p<0.05 to indicate statistical significance in this study.

Results
The demographics of the participants are shown in Table 1. Significant differences were found between subjects with and without CVD for age, sex, race, education level, PIR, physical activity level, past-year alcohol consumption, family history of CVD, serum cotinine category, BMI category, and fish consumption. Table 2 shows the correlations between the serum concentration quartiles of three heavy metals and the risk of overall CVD according to the multivariate logistic regression model after adjustment for covariates. After adjusting for all covariates (model 3), we found that the risk of overall CVD was 1.45 times (95% CI: 1.22, 1.72; p for trend <0.001) higher in the group with the highest quartile of serum Cd concentrations than in the group with the lowest quartile of serum Cd concentrations. No significant association was found between the other heavy metals and the risk of CVD.

Discussion
Our large population-based study is the first to show a doseresponse relationship between Cd and CVD in adults. Furthermore, the serum Cd level was positively related to the overall risk of CVD and the risks of four CVD subtypes. The underlying mechanism may involve Few previous studies have focused on the relationship between Cd levels and CVD in adults. Although a study involving 15,624 US adults showed that urinary Cd levels might be associated with all-cause mortality, more than 30% of which was attributed to CVD, no significant association was observed between Cd levels and CVD mortality (Kim et al. 2019). Given that the urinary Cd level is sensitive to kidney function and physical activity, it cannot be used to accurately Mean ± SD. percentage (frequency) NHANES, National Health and Nutrition Examination Survey; n, number; BMI, body mass index; CVD, cardiovascular disease; PIR, poverty to income ratio; LOD, limit of detection reflect the exposure level Munoz et al. 2020). The association of the blood Cd level with CVD was studied in another population, and the results showed that elevated Cd levels were associated with an increased risk of CVD in adults younger than 60 years old (Jeong et al. 2020). Nevertheless, given that the investigated population consisted of a single  ethnicity, had few CVD subtypes, and was not adjusted for the confounding effects of smoking (Li et al. 2019), this finding lacks generalizability. Therefore, our study provided valid evidence of the relationship between Cd levels and CVD risk after overcoming the abovementioned limitations. Further analysis is needed to investigate the underlying mechanisms by which Cd affects CVD; these mechanisms may involve the relationships between Cd and the triglyceride, total cholesterol, and CRP levels and the WBC count.
Our results showed that the Cd level was positively correlated with the blood levels of triglycerides and total cholesterol, which indicated that elevated triglyceride and total cholesterol levels may play important mediating roles in Cd-related CVD. Consistent with our findings, a substantial amount of evidence suggests that exposure to Cd can result in dyslipidemia, which has been identified as a risk factor for CVD (Samarghandian et al. 2015;Xu et al. 2020b;Zhu et al. 2020). Indications of the possible underlying mechanisms can be found in the results of this and previous studies. An animal study showed that Cd could increase triglyceride levels by reducing lipid uptake receptors in the liver ). In addition, Cd exposure also initiated the endoplasmic reticulum (ER) stress process, which negatively affected lipid homeostasis and metabolic gene expression (Rajakumar et al. 2020). Furthermore, high Cd levels could also increase lipid production by markedly elevating serum lipase activity, reduce lipid degradation by reducing fatty acid β oxidation, and promote lipid synthesis by modifying many liver enzymes, such as hydroxyl-methyl-glutaryl CoA reductase (HMG-CoA) (Aja et al. 2020;Ali et al. 2020;Pawlak et al. 2015;Wu et al. 2017).
Our study also suggested that WBC counts and CRP levels were positively associated with Cd levels, which provided insight into another possible mechanism underlying Cd-related CVD. Although many studies have reported high WBC counts in patients with CVD, the reason for this phenomenon is unclear (Koren-Morag et al. 2005;Lassale et al. 2018;Xu et al. 2020a). Interestingly, some studies have shown that changes in WBC counts and CRP levels indicate systemic inflammation (Baek and Chung 2017;Fagerberg et al. 2017;Saggu et al. 2019). It is worth noting that reactive oxygen species (ROS), autophagy, and immune-related and apoptosis-related genes were found to be involved in this process, which possibly increased the risk of CVD by inducing cytotoxicity, vascular toxicity, nephrotoxicity, and cardiotoxicity Liu et al. 2020b;Reyes-Becerril et al. 2019;Roy et al. 2020;Wang et al. 2020).
Our study had some limitations. First, due to its long biological half-life and low excretion rate, it was difficult to determine whether the timing of Cd exposure influenced the CVD risk in our study (Bhardwaj et al. 2020;Kabamba and Tuakuila 2020). In addition, genetic factors also contributed to the risk of CVD; however, no genetic data were collected in the NHANES. Although the results also showed that mercury levels were positively related to the risks of congestive HF and stroke, our study mainly addressed the relationship between Cd levels and the risk of CVD. In addition, an individual's occupation might influence their living habits and the Cd exposure levels, but the occupational information of participants could not be effectively analyzed because a large amount of this information was missing. Finally, a cross-sectional study can only provide epidemiological evidence of a correlation between Cd levels and the risk of CVD. Further functional experiments and prospective cohort studies are needed to verify this correlation.

Conclusion
In our study, high serum levels of Cd were associated with increased risks of overall CVD and four CVD subtypes, and the Cd concentration was also positively related to the levels of lipids and inflammation parameters, which might provide insight into the possible mechanism underlying Cd-related CVD.