Long-Term Exposure to Fine Particulate Matter and the Risk of Chronic Liver Diseases: A Meta-Analysis of Observational Studies

Although fine particulate matter (PM2.5) is a known carcinogen, evidence of the association between PM2.5 and chronic liver disease is controversial. In the present meta-analysis study, we reviewed epidemiological studies to strengthen evidence for the association between PM2.5 and chronic liver disease. We searched three online databases from 1990 up to 2022. The random-effect model was applied for detection of overall risk estimates. Sixteen eligible studies, including one cross-sectional study, one retrospective cohort study, and 14 prospective cohort studies, fulfilled inclusion criteria with more than 330 thousand participants from 13 countries. Overall risk estimates of chronic liver disease for 10 μg/m3 increase in PM2.5 was 1.27 (95% confidence interval (CI): 1.19–1.35, p < 0.001). We further analyzed the relationship between PM2.5 exposure and different chronic liver diseases. The results showed that increments in PM2.5 exposure significantly increased the risk of liver cancer, liver cirrhosis, and fatty liver disease (hazard ratio (HR) = 1.23, 95% CI: 1.14–1.33; HR = 1.17, 95% CI: 1.06–1.29; HR = 1.51, 95% CI: 1.09–2.08, respectively). Our meta-analysis indicated long-term exposure to PM2.5 was associated with increased risk of chronic liver disease. Moreover, future researches should be focused on investigating subtypes of chronic liver diseases and specific components of PM2.5.


Introduction
Fine particulate matter (PM2.5), responsible for most air pollution, has been increasingly affecting human health with exploding urbanization [1]. PM2.5 is defined as an ambient particulate matter with an aerodynamic equivalent diameter of ≤2.5 µm, mainly including organic matter, carbon, various metal compounds, nitrates, and sulfates [2]. Moreover, PM2.5 is with small particle size and large surface area, which results in absorbing toxins easily [3]. PM2.5 enters the circulation and travels to several organs after infiltrating the alveoli of lungs [4]. The International Cancer Research Center (IARC) has classified inhalable PM2.5 as the first class of carcinogen [5].
Chronic liver disease has become one of the leading causes of death worldwide with the increasing number of cases [6]. Chronic liver disease consists of various liver diseases, such as nonalcoholic fatty liver disease (NAFLD), alcoholic liver disease, liver cancer, cirrhosis and so on [7]. In vivo and in vitro studies have reported that inflammation, oxidative stress, gut microbiota dysbiosis and insulin resistance may be underlying pathogenesis of chronic liver disease [8][9][10][11].
An epidemiological study by Copeland et al. demonstrated that inflammation was identified as a signature of liver fibrosis and elevated risk for liver disease progression [12].
In a randomized study, antioxidant supplementation improved insulin sensitivity and reduced anthropometric parameters of NAFLD patients [13]. PM2.5 may affect the occurrence and progression of chronic liver disease through the mechanisms outlined above.
Current meta-analyses of PM2.5 and chronic diseases primarily focus on the morbidity and mortality of respiratory system diseases, cancer, cardiovascular and cerebrovascular disease [14][15][16][17][18]. In the literature, there are four meta-analyses [19][20][21][22] which have examined the association between PM2.5 and the incidence or mortality of liver cancer in nine epidemiological studies. In the above meta-analysis studies, only Wu et al. reported that PM2.5 was significantly associated with the incidence and mortality of liver cancer (hazard ratio (HR) = 1.28, 95% confidence interval (CI): 1.15-1.41; HR = 1.21, 95% CI: 1.13-1.29, respectively) [22], while others only examined the mortality of liver cancer [19][20][21]. The association between chronic liver disease and PM2.5, other than liver cancer, has been detected in epidemiological studies. Recently, a cross-sectional study showed that a 10 µg/m 3 increase in PM2.5 increased metabolic dysfunction-associated fatty liver disease with odds ratio (OR) (95% CI) of 1.29 (1.25-1.34) [23]. Another prospective cohort study with 58,026 participants over 22 years presented a 1 µg/m 3 increase in PM2.5 levels increased NAFLD with HR (95% CI) of 1.06 (1.04-1.07) [24]. Therefore, our present meta-analysis aims to investigate the association between PM2.5 and the risk of chronic liver disease based on observational epidemiological studies and various subgroup analyses to strengthen results based on new evidence.

Study Selection and Eligibility
The selected inclusion criteria were as follows: (1) observational epidemiological studies, (2) investigation of the association between PM2.5 and the incidence and mortality of chronic liver diseases, (3) data of PM2.5 exposure levels were collected from monitoring station, horizontal-vertical locations, or satellite, and (4) outcome measures with adjusted relative risk (RR) or hazard ratio (HR) and 95% CI reported. We included comprehensive analysis to re-calculate duplicated or shared data appeared in two or more analyses [19,20]. The exclusion criteria were as follows: (1) reviews, letters, and reports, (2) animal-or cellrelated data, (3) studies without subdivision of the types of particulate matter, (4) studies without PM2.5 increment data, (5) data collection time less than one year, and (6) studies on acute liver disease but not chronic liver disease [19][20][21][22]. Based on the inclusion criteria, J.S. and H.X. independently assessed the eligibility of studies. Any disagreements between evaluators were solved by mutual decision.

Data Extraction
A further collection from eligible articles were performed by J.S. and Q.Z. The collected data in each article were as follows: the publication year, first author name, type of study, study period, location of the study, number of participants, type of liver disease, adjustment variables, and adjusted RR/HR, 95% CI, etc.

Literature Quality Assessment
We estimated the quality of eligible studies with Newcastle-Ottawa Scale (NOS) [25]. We defined low-quality or high-quality studies according to the average score. The literature with a score higher than the average score was defined as high-quality.

Statistical Analyses
We standardized the PM2.5 increment to a 10 µg/m 3 increment, and recalculated the RR/HR with the following Formula [26]: where RR is the relative risk, and Ln is the log to base e. Cochran's Q and I 2 were performed to evaluate heterogeneity in selected studies [27]. If p value for heterogeneity was less than 0.10 or I 2 value was <50%, fixed effects model (FEM) was applied for the null hypothesis. Otherwise, we applied the random effects model (REM). Begg's funnel plot and Egger's test were used to calculate publication bias [28]. If the funnel plot was asymmetry or the p value for Egger's test was less than 0.05, it indicates that there is publication bias in enrolled studies. We used Stata statistical software 11.0 to conduct all statistical analyses.

Characteristics of Studies Included in the Meta-Analysis
The characteristics of 16 studies included in the meta-analysis were shown in Table 1. All were published between 2016 and 2022, which included more than 331,114 participants. I above studies which reported age, the age range of the participants was more than 18 yea garding the types of liver diseases, 13 studies were related to liver cancer [29-32,34-38,40 study was related to liver cirrhosis [33], and two studies were related to fatty liver disease bolic dysfunction-associated fatty liver disease and NAFLD, respectively) [23,24]. Among  Years Enrolled, data collection time; NOS, Newcastle-Ottawa Scale.Studies were conducted in multiple countries, mainly in Asia (n = 7), Europe (n = 4), and the Americas (n = 5). The number of participants in these studies ranged from less than 300 to more than 180 thousand. All studies included both male and female participants. The NOS score for methodological quality of the included studies ranged from 8 to 9 points, with an average of 8.9 points, and 16 studies were of high quality (NOS score ≥ 8).

Overall Meta-Estimates and Publication Bias
To explore the association between PM2.5 exposure and chronic liver diseases, the adjusted HRs and 95% CIs, which were extracted from included researches, were used to calculate the overall effects. The results were I 2 = 73.9%, p < 0.001 under the random effects model. The results showed that a 10 µg/m 3 increase in PM2.5 concentration was significantly correlated with chronic liver diseases, and HR was 1.27 (95% CI: 1.19-1.35, p < 0.001), indicating that maternal exposure to PM2.5 was positively correlated with chronic liver diseases. The result of the Egger's test for asymmetry showed p = 0.28, indicating no publication bias (shown in Figure 2). In addition, the sensitivity analysis showed that the summary results did not substantially change when excluding any single study. Years Enrolled, data collection time; NOS, Newcastle-Ottawa Scale.Studies were conducted in multiple countries, mainly in Asia (n = 7), Europe (n = 4), and the Americas (n = 5). The number of participants in these studies ranged from less than 300 to more than 180 thousand. All studies included both male and female participants. The NOS score for methodological quality of the included studies ranged from 8 to 9 points, with an average of 8.9 points, and 16 studies were of high quality (NOS score ≥ 8).

Overall Meta-Estimates and Publication Bias
To explore the association between PM2.5 exposure and chronic liver diseases, the adjusted HRs and 95% CIs, which were extracted from included researches, were used to calculate the overall effects. The results were I 2 = 73.9%, p < 0.001 under the random effects model. The results showed that a 10 μg/m 3 increase in PM2.5 concentration was significantly correlated with chronic liver diseases, and HR was 1.27 (95% CI: 1.19-1.35, p < 0.001), indicating that maternal exposure to PM2.5 was positively correlated with chronic liver diseases. The result of the Egger's test for asymmetry showed p = 0.28, indicating no publication bias (shown in Figure 2). In addition, the sensitivity analysis showed that the summary results did not substantially change when excluding any single study.

Subgroup Analyses of Particulate Matter on Risk of Liver Diseases
Then, we presented the association between PM2.5 and risk of chronic liver diseases in subgroup analyses by incidence and mortality. Across all selected studies, in total, PM2.5 significantly increased both the incidence and mortality of chronic liver diseases (pooled HR = 1.33, 95% CI: 1.20-1.46; I 2 = 71.5% and pooled HR = 1.21, 95% CI: 1.09-1.35; I 2 = 78.2%, respectively) in Figure 2. The association between PM2.5 and risk of chronic liver diseases in subgroup analyses by region and types of chronic liver disease is shown in Table 2. Regarding region, PM2.5 had a significant effect on non-lung risk of chronic

Subgroup Analyses of Particulate Matter on Risk of Liver Diseases
Then, we presented the association between PM2.5 and risk of chronic liver diseases in subgroup analyses by incidence and mortality. Across all selected studies, in total, PM2.5 significantly increased both the incidence and mortality of chronic liver diseases (pooled HR = 1.33, 95% CI: 1.20-1.46; I 2 = 71.5% and pooled HR = 1.21, 95% CI: 1.09-1.35; I 2 = 78.2%, respectively) in Figure 2. The association between PM2.5 and risk of chronic liver diseases in subgroup analyses by region and types of chronic liver disease is shown in Table 2. Regarding region, PM2.5 had a significant effect on non-lung risk of chronic liver diseases in Asia, Europe, and North America (Table 2). Moreover, the meta-analyses of higher exposure to PM2.5 and risk of liver cancer, liver cirrhosis, and fatty liver disease showed significant correlation (Table 2). Regarding geographic region, PM2.5 had a significant effect on both incidence and mortality of chronic liver diseases in Asia, Europe, and North America. The meta-analysis illustrated the association between PM2.5 exposure and risk of incidence of liver cancer, liver cirrhosis, and fatty liver disease (pooled HR = 1.28, 95% CI: 1.15-1.42; I 2 = 0, pooled HR = 1.17, 95% CI: 1.06-1.29; I 2 = NA, and pooled HR = 1.51, 95% CI: 1.09-2.08; I 2 = 94.7%, respectively; Table 2).

Discussion
In the present meta-analysis, 16 studies involving more than 330 thousand participants in 13 countries were included. Our study showed a significant positive association between PM2.5 exposure and risk of chronic liver diseases. Such association was also found when studies were separated into incidence and mortality of chronic liver disease. Among each type of chronic liver disease, increments in PM2.5 exposure had a harmful impact on the risk of liver cancer, liver cirrhosis, and fatty liver disease. Therefore, we believe that the data obtained from our study show a strong correlation between PM2.5 exposure and the risk of chronic liver diseases.
Overexposure to PM2.5 pollution promoted the incidence and mortality of serious multi-system diseases including the respiratory system, cardiovascular systems, and digestive system. An increasing number of studies have proved that PM2.5 enters the blood circulation and deposits in liver, brain, and other organs through gas exchange in the alveolus [43,44]. Recent studies also reported that inhalation of PM2.5 disordered gut microbiota, leading to abnormal serum metabolome and insulin resistance [45,46]. However, the role of PM2.5 in the occurrence, development and mechanism of chronic liver diseases remains unclear.
Another key potential mechanism was caused by oxidative stress. An in vivo study showed that continuous exposure to PM2.5 activated hepatic stellate cells through oxidative stress and increased the risk of liver fibrosis [51]. Ding et al. [52] explored that chronic PM2.5 exposure disordered redox homeostasis, and induced hepatic steatosis in mice by increasing expression of hepatic Nrf2 and Nrf2-regulated antioxidant enzyme gene, the sterol regulatory element binding protein-1 c, and fatty acid synthase in the liver. More and more evidence showed that the development of liver cancer was associated with oxidative stress, which induced DNA damage, cancer-related gene mutations, and dysregulation [53]. In addition, a randomized controlled trial study from China, including 107 cases of liver cancer and 178 healthy controls, confirmed the significant association between oxidative stress and the risk of liver cancer [54].
Gut microbiota dysbiosis and insulin resistance might be one of possible mechanisms. Both in vivo and prospective studies have indicated that chronic exposure to PM2.5 caused dysbiosis of the gut microbiota and may subsequently contribute to the development of abnormal glucose metabolism and insulin resistance [55][56][57][58]. Previous studies presented that gut microbiota affected bioprocessing of bile acids and production of pro-inflammatory intermediate metabolites resulting in hepatic disorders [59,60]. Another in vivo study showed that 12 weeks of PM2.5 exposure activated the c-Jun n-terminal kinase signaling pathway, increased insulin receptor substrate-1 phosphorylation, and caused significant liver damage with hepatic insulin resistance [61]. Moreover, insulin resistance was strongly associated with fat accumulation in the liver [62]. Thus, gut microbiota dysbiosis and insulin resistance caused by exposure to PM2.5 may result in chronic liver diseases.
Studies have reported that exposure to some chemical compounds, carcinogenic compounds, and metals also increased the risk of chronic liver disease. Polychlorinated biphenyls, as environmental endocrine and metabolism disrupting chemicals, were associated with the genesis and progression of steatohepatitis and liver cancer induced by insulin resistance, dyslipidemia, proinflammatory cytokines [63,64]. Both childhood and adulthood passive smoking contributed to higher risk of fatty liver [65]. Exposure to even low doses of bisphenol A (BPA) may adversely affect liver function of children in their later life [66]. Environmental exposure to cadmium was associated with the risk of suspected NAFLD [67]. To increase the accuracy and validity of the study, more studies are urgently needed to analyze the association between different environmental pollutants and specific chronic liver diseases.
The main strengths of our present meta-analysis are the inclusion of various of chronic liver diseases, separation incidence from mortality of chronic liver diseases, and the coverage of more participants and countries than previous studies. Previous meta-analysis focused on the relationship between PM2.5 exposure and liver cancer. However, in the present meta-analysis, we included the studies to comprehensively analyze the relationship between PM2.5 exposure and other chronic liver diseases as well. All included data were adjusted for multiple hypothesized confounders in the present study, which was also free from publication bias. Overall, as far as we know, this is the first study that has comprehensively demonstrated the relationship between different types of chronic liver diseases and PM2.5 exposure.
Additionally, our study also has several limitations. First, each included study has different adjusted confounders. Thus, there may be information bias in our study, which reduced the accuracy. Second, there was a lack of data on other types of chronic liver disease except liver cancer. Third, we did not distinguish between indoor and outdoor PM2.5 pollution, and did not explore the source and composition of PM2.5 pollution. Fourth, heterogeneity might be generated from difference between individuals and between various observational studies to influence the result. Finally, the number of included studies is not enough; therefore, more studies are needed.

Conclusions
Our results suggested that PM2.5 exposure was associated with an increased risk of different chronic liver diseases. Although data for specific liver diseases were restricted, our results suggested that PM2.5 can increase the risk of fatty liver disease, liver cirrhosis, and liver cancer. The current meta-analysis indicated that diverse biological mechanisms might play a role in the numerous subtypes of chronic liver disease. More and more studies focus on the effects of various ambient air pollution on the risk of chronic liver diseases. Therefore, future research is needed to strengthen the association between certain types of air pollutants and specific chronic liver diseases.