DOI: https://doi.org/10.21203/rs.3.rs-2111826/v1
The relationship between exposure to a single heavy metal and liver function has been reported. However, the effect of strontium exposure on liver function has not been examined. A cross-sectional study involving 1,021 residents from a mining area in Hunan Province, China, was conducted to explore the single and combined effects of strontium exposure on liver function. Liver function was determined by detecting the level of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin (TBIL) in the serum. The plasma concentrations of strontium (Sr), lead (Pb), cadmium (Cd), manganese (Mn), Zinc (Zn), and copper (Cu) in residents were measured using inductively coupled plasma-mass spectrometry (ICP-MS). The sociodemographic characteristics, lifestyle factors, and history of disease were assessed by questionnaire. Then the association between plasma Sr and liver function was analyzed by multiple linear regression and BKMR.
A positive correlation was found between Sr and ALT and Sr and AST (ALT: β = 14.86, 95% CI: 8.23, 21.50, P < 0.05 and AST: β = 9.67, 95%CI: 3.54, 15.80, P < 0.05) by multiple linear regression model. According to the BKMR, Sr and Pb and Sr and Cu had a synergistic effect on liver function. A single exposure or combined exposure to Sr are associated with liver function, which is influenced by age and gender. Sr and Pb and Sr and Cu have a synergistic effect on liver function. We reveal that Sr was an independent risk factor for ALT and AST based on the results of BKMR and GAMS.
Liver diseases are a significant cause of morbidity and mortality worldwide, with about 2 million people dying from them each year (Asrani, Devarbhavi et al. 2019). In 2017, more than one-fifth of Chinese people were reported to have liver diseases. In addition to viruses and bacteria, harmful chemical toxins are garnering attention as potential causes, especially environmental pollutants. In recent years, increasing evidence has shown that heavy metal exposure is associated with liver diseases (Huang, Pan et al. 2021).
Heavy metals are widely distributed in the environment. With the development of industrialization, heavy metal pollution has become an increasingly severe public health problem. Some studies found that toxic heavy metals were harmful to human health. The disease "Itai-Itai" caused by cadmium-polluted drinking water was first reported in Japan (Ishihara, Kobayashi et al. 2001, Kobayashi, Okubo et al. 2002, Wang, Chen et al. 2021). Many heavy metals influence liver function, such as Cd, Cu, and Pb (Skibniewski, Skibniewska et al. 2017). However, the association between some heavy metals, such as Sr, and liver functions is unknown. Sr has toxic and therapeutic effects. Hendrych found that Sr exposure could increase the risk of cardiovascular disease (Hendrych, Olejnickova et al. 2016) and diabetes (Chen, Guo et al. 2020). The U.S. Environmental Protection Agency (USEPA) proposed that drinking water with a strontium (II) concentration above 0.3 µg/L should be reported (Guelfo and Adamson 2018). Sr is not a restrictive indicator in China because there is no strong evidence of the effect of Sr exposure on human health.
Studies have shown that multiple heavy metal exposure could pose a risk to human health, with Pb and Mn being synergistically toxic for infant development (Claus Henn, Ettinger et al. 2010). Liu found that co-exposure to multiple heavy metals could lead to elevated fasting blood glucose levels. The Hunan province is located in southern China and is famous for its rich heavy metal mineral resources (Liu, Chen et al. 2019). It has the second-largest manganese mine and the third-largest lead-zinc mine in China. In the vicinity of the mining area, heavy metals are abundant in water and vegetables,(Zeng, Wei et al. 2015, Wang, Bao et al. 2021) with Pb, Mn, Cu, Cd, and Zn levels in the soil far exceeding background levels (Luo, Ren et al. 2020).
Hengyang is located in the southern part of Hunan province, China, and has severe heavy metal pollution. We selected it as our study site to investigate the associations between Sr plasma levels and liver function and to explore the joint effect of co-exposure to Sr and other heavy metals on liver function.
A mining area with heavy metal pollution in Hengyang was selected as our study area, located in the southern part of Hunan province, China. This area is rich in non-ferrous metal resources. All residents living in this mining area for more than one year who were over 18 years of age were enrolled as study subjects. We excluded those who refused to fill out questionnaires, were positive for protein in their urine, suffered from cardiovascular disease, cancer, and significant liver diseases (liver cancer, hepatitis, etc.), or used radiational medicine or antibiotics in the previous two weeks. In total, 1,021 subjects were included in our study. This study was approved by the Central South University Hospital's Ethics Committee.
We surveyed all subjects by questionnaire and physical examination. All investigators received rigorous training before the formal survey and then conducted face-to-face interviews to collect basic information regarding sociodemographic characteristics and lifestyle factors, such as smoking, drinking, and weekly exercise. A physical examination was performed to collect health information, such as body mass index (BMI) (calculated according to the formula weight (kg)/height(m)2), blood pressure, and dyslipidemia. Hypertension was defined as blood pressure levels of more than 140/90 mmHg or having a history of taking medication for hypertension. Dyslipidemia was diagnosed based on the participant's lipid levels (total cholesterol > 5.72 mmol/L or triglycerides > 1.70 mmol/L ).
Liver functions were detected by conventional liver function tests measuring serum levels of aminotransferases (Lozano-Paniagua, Parrón et al. 2021). We measured ALT, AST, and TBIL levels. The normal range of ALT, AST, and TBIL were 5–40 U/L, 8–40 U/L, and 3.4–20.6 µmol/L, respectively.
The method of detecting plasma heavy metals was based on the published literature on experimental procedures(Li, Xu et al. 2019, He, Chen et al. 2020). After the participant's plasma samples were freeze-thawed at room temperature, 200 µl of plasma was collected and placed in a 5 ml polypropylene centrifuge tube with 1.0% nitric acid. This was placed into the ultrasonic device, shaken for 15 minutes, and tested after fully mixed. An Inductively Coupled Plasma Mass Spectrometry Instrument (ICP-MS) (Agilent 7700X, USA Agilent Technologies Co, Ltd.) was used to detect the content-heavy metals in plasma samples, including six heavy metals (Sr, Mn, Cu, Zn, Cd, and Pb). For quality control, three replicates were used to improve the recovery of collected plasma samples. The plasma sample concentration at which metal detection was below the limit of detection (LOQ) was expressed as LOQ/2.
The demographic characteristics of the participants were descriptively analyzed. Quantitative variables were presented as a mean ± standard deviation and categorical variables were presented as frequencies. We used the Student's t-test, ANOVA, or Chi-square test. Spearman correlation was used to analyze the correlation coefficients among the six heavy metals.
The association between plasma Sr and liver function was analyzed using a restricted cubic spline (RCS) curve (Johannesen, Langsted, et al. 2020). We then used stratified analysis to investigate the association between Sr and liver function by age and gender. Multiple metals linear regression models were employed to estimate the relationship between co-exposure to the six metals as a mixture and indicators of liver function. A Bayesian kernel machine regression (BKMR) model (Bobb, Valeri et al. 2015) was used to analyze the association between heavy metal mixtures with liver function and the possible interaction between Sr and other heavy metals. The posterior inclusion probability (PIP), which was a measure of variable importance with a higher value indicating greater importance according to the BKMR was estimated for each heavy metal. To assess the reproducibility of the BKMR results and to detect potential interactions among the mixture components, we subsequently fitted multivariable generalized additive models (GAMs) to evaluate the validity of prior BKMR assumptions in lower-dimensional settings. All of the data was analyzed with SPSS 25.0 software and R 3.5.0.
Based on inclusion and exclusion criteria, a total of 1,021 residents from the mining areas were included in this study. Their average age was 44.6\(\pm\)9.4 years. There were 403 individuals with abnormal ALT, 182 individuals with abnormal AST, and 183 individuals with abnormal TBIL. Abnormal ALT significantly differed by t in age, sex, education level, and history of diabetes. Abnormal TBIL was significantly different by sex only, and abnormal AST was significantly different by sex, educational status, BMI, and history of diabetes (Table 1).
Parameter | N | ALT | AST | TBIL | ||||||
---|---|---|---|---|---|---|---|---|---|---|
normal | abnormal | P | normal | abnormal | P | normal | abnormal | P | ||
Gender(n) | ||||||||||
Male | 894 | 559 | 335 | 0.010 | 720 | 174 | 0.000 | 746 | 148 | 0.020 |
Female | 127 | 59 | 68 | 119 | 8 | 92 | 35 | |||
Age (mean ± SD) | 44.69.4 | 43.57.4 | 46.411.7 | 0.003 | 44.89.6 | 43.38.7 | 0.272 | 44.58.9 | 45.311.7 | 0.598 |
BMI(mean ± SD) | 24.03.5 | 23.33.1 | 25.13.8 | 0.000 | 24.03.5 | 24.03.4 | 0.967 | 23.73.2 | 25.44.3 | 0.000 |
Smoking(n) | 0.146 | 0.139 | 0.426 | |||||||
Non-smoker | 406 | 231 | 175 | 324 | 82 | 326 | 80 | |||
Passive smoker | 156 | 100 | 56 | 126 | 30 | 132 | 24 | |||
Smoker | 459 | 287 | 172 | 389 | 70 | 380 | 79 | |||
Drinking(n) | 0.078 | 0.190 | 0.098 | |||||||
Non drinker | 713 | 425 | 288 | 596 | 117 | 596 | 117 | |||
Passive drink | 59 | 30 | 29 | 48 | 11 | 43 | 16 | |||
drinker | 249 | 163 | 86 | 195 | 54 | 199 | 50 | |||
Education level(n) | 0.000 | 0.777 | 0.004 | |||||||
Junior high school and below | 374 | 193 | 181 | 309 | 65 | 290 | 84 | |||
Senior high school and above | 647 | 425 | 222 | 531 | 116 | 548 | 99 | |||
Diabetes(n) | 0.000 | 0.833 | 0.000 | |||||||
Yes | 871 | 553 | 318 | 715 | 156 | 733 | 138 | |||
No | 150 | 65 | 85 | 124 | 26 | 105 | 45 | |||
Hypertension(n) | 0.064 | 0.458 | 0.272 | |||||||
Yes | 994 | 606 | 388 | 816 | 178 | 818 | 176 | |||
No | 27 | 12 | 15 | 23 | 4 | 20 | 7 |
The concentration ranges of Zn, Cu, Mn, Pb, Sr, and Cd in the plasma were 388.46-1800.59 µg/L, 144.33-1896.56 µg/L, 0.60-153.51 µg/L, 0.0057–33.84 µg/L, 0.60-153.51 µg/L, and 0.0008–59.67 µg/L, respectively (Table 2). The arithmetic means were 988.46, 854.43, 4.35, 3.46, 4.35, and 0.304 µg/L, respectively. The correlation between the six heavy metals is shown in Fig. 1. The correlations among these six heavy metals were statistically significant (P < 0.05). The correlations between Sr and the other five heavy metals were weak, and the correlation coefficient range was 0.17–0.29. However, there were stronger correlations between Pb and Cd and Pb and Mn, and the correlation coefficients were 0.55 and 0.58, respectively.
Exposure | mean | Range | Percentile | |||
---|---|---|---|---|---|---|
Mix | Max | 25th | 50th | 75th | ||
Zn(µg/L) | 988.46 | 388.19 | 1800.59 | 857.93 | 970.05 | 1099.86 |
Cu(µg/L) | 854.46 | 144.33 | 1896.56 | 716.40 | 832.78 | 982.81 |
Mn(µg/L) | 4.35 | 0.60 | 153.51 | 1.83 | 2.85 | 4.85 |
Pb(µg/L) | 3.46 | 0.0057 | 33.84 | 1.04 | 2.15 | 3.60 |
Sr(µg/L) | 4.35 | 0.60 | 153.51 | 21.83 | 26.84 | 28.12 |
Cd(µg/L) | 0.304 | 0.0008 | 59.62 | 0.12 | 0.19 | 0.27 |
According to Table 3, Sr (β = 15.47, 95% CI: 8.92, 22.01, P < 0.05) was significantly associated with ALT and significantly associated with AST in model 1(β = 10.57, 95% CI: 4.55, 16.59, P < 0.05). In model 2, after adjusting for covariates and other heavy metals, Sr was associated with ALT (β = 14.86, 95% CI: 8.23, 21.50, P < 0.05) and AST (β = 9.67, 95% CI: 3.54, 15.80, P < 0.05). However, it was not significantly associated with TBIL (P > 0.05). Cu was negatively associated with ALT and TBIL (P < 0.05). Zn was positively associated with TBIL (P < 0.05), and Pb was negatively associated with ALT (P < 0.05). We analyzed the collinear relationship among the six heavy metals (Table S1). The variance inflation factors (VIF) showed no obvious collinearity among Zn (1.267), Cu (1.308), Pb (1.082), Mn (1.071), Sr ( 1.094), and Cd (1.022).
exposure | ALT(β) | P | AST(β) | P | TBIL(β) | P |
---|---|---|---|---|---|---|
Model1 | ||||||
log(Sr) | 15.47(8.92, 22.01) | < 0.001 | 10.57(4.55, 16.59) | < 0.0001 | -1.62(-3.59, 0.34) | 0.10 |
log(Mn) | 2.90(-0.06, 5.87) | 0.06 | -0.05(4.55, 16.59) | 0.96 | 0.23(-0.65, 1.12) | 0.60 |
log(Pb) | -2.70(-5.08, -0.32 | 0.02 | -2.20(-4.39, -0.01) | 0.04 | 0.45(-0.26, 1.17) | 0.21 |
log(Cd) | 0.92(-2.07, 3.92) | 0.54 | -1.09(-3.85, 1.65) | 0.43 | -0.36(-1.26, 0.54) | 0.43 |
log(Cu) | -10.01(-18.29,-1.73) | 0.01 | 3.26(-4.35, 10.88) | 0.40 | -3.13(-5.62, -0.64) | 0.01 |
log(Zn) | -6.00(-16.51, 4.50) | 0.26 | -6.51(-16.18, 3.15) | 0.18 | 4.13(0.97, 7.29) | 0.01 |
Model2 | ||||||
log(Sr) | 14.86(8.23,21.50) | < 0.001 | 9.67(3.54, 15.80) | < 0.001 | -1.43(-3.4, 0.56) | 0.15 |
log(Mn) | 2.98(0.004,5.96) | 0.04 | 0.10(-2.65, 2.85) | 0.94 | 0.22(-0.67, 1.12) | 0.62 |
log(Pb) | -2.92(-5.32,-0.53) | 0.01 | -2.14(-4.35, 0.07) | 0.06 | 0.41(-0.30, 1.13) | 0.26 |
log(Cd) | 0.81(-2.19-3.82) | 0.59 | -0.96(-3.74, 1.81) | 0.49 | -0.42(-1.32, 0.49) | 0.36 |
log(Cu) | -10.67(-19.56,-1.78) | 0.02 | 0.94(-7.26, 9.15) | 0.82 | -2.76(-5.45, -0.08) | 0.04 |
log(Zn) | 2.98(-15.96, 6.24) | 0.10 | -3.51(-13.77, 6.74) | 0.50 | 3.74(0.39, 7.09) | 0.02 |
Restricted cubic spline was used to flexibly model and visualize the relationship between Sr and liver function. ALT had an increasing trend until around 26.85 ug/L of predicted plasma strontium concentration, and then it started to increase rapidly (P > 0.05). As the concentration of Sr in the plasma increased, the rising trend of AST slowed. When the plasma strontium concentration reached 26.63 µg/L, the trend gradually flattened (P > 0.05). However, there was a negative association between Sr and TBIL (P > 0.05) (Table S2). We then analyzed the association between Sr and ALT and Sr and AST in different age groups and genders. A plasma strontium concentration higher than 26.85 ug/L was more harmful to those under the age of 45 and males.
To analyze the interaction between Sr and the other heavy metals, BKMR was used to estimate the combined effect of the metal mixture. The posterior inclusion probability (PIP) is shown in Table S3. The conditional PIPs of the six heavy metals were higher than 0.3; Sr had the lowest PIP with TBIL, which indicates a lower association with TBIL. The model adjusted for covariates that were statistically different for ALT, such as age, gender, education level, and history of diabetes. Figure 3A shows the overall effect of the six heavy metals on ALT. When the plasma concentration of all of the heavy metals was above the 50th percentile, there was a negative correlation compared to when the plasma concentration of all of the heavy metals was at the 50th percentile. When the plasma concentration of all of the heavy metals was below the 50th percentile, there was an overall decreasing trend.
However, the overall effect of all heavy metals on AST was a generally increasing trend. When the plasma concentrations of all of the heavy metals were above the 50th percentile, there was a positive correlation compared to when the plasma concentration of all of the heavy metals was at the 50th percentile (and the 95% CI contained 0). Figure 3B shows the effect of single heavy metal exposure when the plasma concentration of one heavy metal was changed from the 25th to the 75th percentile when the plasma concentration of the other heavy metal was fixed at a given percentile. When the plasma concentrations of the other five heavy metals were fixed at the 50th percentile, Sr was positively correlated with ALT, an effect that declined from the 25th to the 75th percentile. When the Cu concentration increased from the 25th to the 75th percentile, Cu was negatively correlated with ALT at the 75th percentile. Pb, Zn Mn, and Cd did not correlate with ALT when their plasma concentrations increased from the 25th to the 75th percentiles because the confidence interval for the four heavy metals contained 0. Sr was positively correlated with AST; however, its effect decreased when the plasma Sr concentration increased from the 25th to the 75th percentile. When the plasma Cu concentration increased from the 25th to the 75th percentile, Cu was negatively correlated with AST. Cu, Zn Mn, Sr, Pb, and Cd did not correlate with TBIL when their plasma concentrations increased from the 25th to the 75th percentile because the confidence intervals of the six heavy metals contained 0.
Bivariate exposure-response analysis was used to explore the interaction between any two metals when the plasma concentration of each increased from the 25th to 75th percentiles (Fig. 3C). For ALT, there was an interaction between Sr and Pb, Sr and Cu, and Pb and Cu. For AST, there was an interaction between Sr and Pb and Sr and Cu. For TBIL, there was no significant interaction between any two heavy metals.
To verify the reliability of the results, we fitted the generalized additive mixed model (GAMM) to adjust for various potential confounding factors. As shown in Table S4, Sr-Cu and Sr-Pb interacted (P < 0.05).
This study investigated the associations between the independent and joint effects of Sr exposure on the liver function of mining residents in China. According to the RCS and univariate analysis, we found that Sr was positively correlated with ALT, especially in participants under the age of 45 (OR: 2.21, 95% CI: 1.50–3.29) or in males (OR: 1.60, 95% CI: 1.22–2.15), and positively correlated with AST. In a multiple linear regression model, we found that Sr has a positive association between ALT (β = 14.86, 95% CI: 8.23, 21.50, P < 0.05) and AST (β = 9.67, 95% CI: 3.54, 15.80, P < 0.05) and that Sr had a critical impact on liver function. The results of the multiple linear regression were consistent with the results of the RCS and univariate analysis. Liu found that Sr2+ caused hepatotoxicity in zebrafish embryos, mainly manifested by hepatic macrophages and delayed yolk sac uptake (Liu, Chen et al. 2019). Another study found that urinary Sr might reduce male sperm count(Miao, Liu et al. 2021) and excessive Sr could lead to osteoporosis (Bauman, Valinietse et al. 1988). However, the mechanism of action of Sr on ALT and AST is still unclear. Possibly, excess strontium may interfere with the uptake and metabolism of calcium by hepatocytes (Pors Nielsen 2004). Our research is the first to investigate the effect of Sr on liver function in a Chinese population.
The liver was more susceptible to damage as it is the organ that accumulates the heaviest metals (Fang, Yin et al. 2021). Many studies have reported the effects of heavy metal exposure on human health (Bhagat, Nishimura et al. 2021). Fan found that many toxic metals may increase the risk of acute and chronic diseases, such as arsenic-induced diabetes (Fan, Zhu et al. 2017). Farkas found that the liver was the most important storage organ for Cu, Fe, Mn, and Zn (Farkas, Bidló et al. 2021). A study found that four Taiwanese toothed cetaceans had the highest content of Zn and Cu in the liver (Chen, Lin et al. 2020). There was a positive correlation between Mn and ALT (β = 2.98, 95% CI: 0.004, 5.96, P < 0.05). Wang found that long-term, low-dose Mn exposure could cause pathological liver changes and might be related to the inhibition of Nrf2 expression (Wang, Bao et al. 2021). While a negative correlation between Pb (β = -2.92, 95% CI: -5.32, -0.53, P < 0.05), Cu (β = -10.67, 95% CI: -19.56, -1.78, P < 0.05), and ALT. Although the hepatotoxic mechanisms of Pb remain unclear, one explanation is that Pb could stimulate the reactive oxygen species (ROS) responsible for oxidative stress and the destruction of the antioxidant defense system (Lee, Choi et al. 2019).
Yuan also found that low doses of Pb could significantly modulate hepatic superoxide dismutase (Yuan, Dai et al. 2014). TBIL had a negative correlation with Cu (β = -2.76, 95% CI: -5.45, -0.08, P < 0.05) and a positive correlation with Zn (β = 3.74, 95% CI: 0.39, 7.09, P < 0.05). Ge found that plasma copper levels were associated with a decrease in liver function, which was similar to our results (Ge, Liu et al. 2020). Clinical studies found that Cu was positively correlated with the serum levels of gamma-glutamyl transpeptidase (GGTP) in liver cirrhosis (Poznański, Sołdacki et al. 2021). Hyder found an association between urinary cadmium concentrations and liver function (Hyder, Chung et al. 2013). However, our results showed that Cd and liver function have no significant correlation, which may be due to the lower concentration of cadmium in the plasma. A low-pollution environment study showed no significant association between low concentrations of Cd and ALT, which was consistent with our results (Cave, Appana et al. 2010).
The scientific community focuses on the disease risk caused by simultaneous exposure to multiple metals, which is how the metals present in the environment. The BKMR model describes the overall effect and independent and joint associations between co-exposure to six metals and liver function. When other heavy metals were fixed at a percentile, there was a positive correlation between Sr and ALT and Sr and AST. The multiple linear regression results were consistent with the univariate analysis. Then we found that Sr, Pb, and Cu have a synergistic effect. The results of the GAMs model were consistent with those of BKMR. Howe found that simultaneous exposure to multiple metals, including As, Cd, Co, Ni, Hg, Pb, and Ti, synergistically increased health risk (Howe, Claus Henn et al. 2020). However, they found no interaction between Cd and Pb, which was in line with our study. A case-control study of trace elements and the risk of goiter showed an interaction between Pb and Sr (He, Li et al. 2021). Huang found that Pb was negatively correlated with liver function, and co-exposure would aggravate liver damage, but no interaction between Pb and Cd was found in the population survey (Huang, Pan et al. 2021). Bauman found that combined exposure to more than three heavy metals might increase health risks (Bauman, Valinietse et al. 1988).
In summary, this study analyzed the association of single Sr exposure and Sr combined exposure with liver function. The results were verified by GAMM and BKMR and provide important clues regarding Sr's health risk. However, our study sample size is limited and the cause-effect relationship between Sr exposure and liver function needs further investigation.
Single or combined exposure to Sr is associated with liver function, which is influenced by age and gender. Sr and Pb as well as Sr and Cu have a synergistic effect on liver function.
Acknowledgments
The study were supported by Hunan Outstanding Youth Fund (2020JJ3053), National Natural Science Foundation of China (81773393, 81502787), Hunan Key R&D Project (2019SK2041, 2016YFC0900802) and Central South University Innovation Drive Project (20170027010004).
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare the following financial interests/personal relationships which may be considered as potential competing interests.