Association between Ambient Particulate Matter 2.5 Exposure and Mortality in Patients with Hepatocellular Carcinoma

Air pollution is a severe public health problem in Taiwan. Moreover, Taiwan is an endemic area for hepatocellular carcinoma (HCC). This study examined the effect of particulate matter 2.5 (PM2.5) exposure on mortality in this population. A total of 1003 patients with HCC treated at Chang Gung Memorial Hospital between 2000 and 2009 were included in this study. At the end of the analysis, 288 (28.7%) patients had died. Patients with HCC living in environments with PM2.5 concentrations of ≥36 µg/m3 had a higher mortality rate than patients living in environments with PM2.5 concentrations of <36 µg/m3 (36.8% versus 27.5%, p = 0.034). The multivariate Cox regression analysis confirmed that PM2.5 ≥ 36 µg/m3 was a significant risk factor for mortality (1.584 (1.162–2.160), p = 0.004). A nonlinear relationship was observed between the odds ratio and PM2.5. The odds ratio was 1.137 (1.015–1.264) for each increment of 5 µg/m3 in PM2.5 or 1.292 (1.030–1.598) for each increment of 10 µg/m3 in PM2.5. Therefore, patients with HCC exposed to ambient PM2.5 concentrations of ≥36 µg/m3 had a 1.584-fold higher risk of death than those exposed to PM2.5 concentrations of <36 µg/m3. Further studies are warranted.


Introduction
In 2013, the International Agency for Research on Cancer [1] classified outdoor air pollution and particulate matter from outdoor air pollution as carcinogenic to human beings (Group 1), according to sufficient evidence of carcinogenicity in humans and experimental animals, and strong mechanistic evidence. Long-term exposure to particulate matter air pollution is a well-known risk factor for cardiopulmonary and pulmonary neoplasm mortality [2].
In a study conducted in Taiwan (Table 1), Pan et al. [4] reported that PM 2.5 exposure was positively associated with risks of HCC, and that an elevated blood alanine aminotransferase concentration could be a mediator for the association between PM 2.5 and HCC. In a European study, Pedersen et al. [5] demonstrated that the HR associated with each 10 µg/m 3 increase in nitrogen dioxide was 1.10 (95% CI 0.93-1.30) and 1.34 (95% CI 0.76-2.35) for each 5-µg/m 3 increase in PM 2. 5 . In an American study, Deng et al. [6] revealed that the all-cause mortality HR associated with a standard deviation (5.0 µg/m 3 ) increase in PM 2.5 was 1.18 (95% CI 1.16-1.20): 1.31 (95% CI 1.26-1.35) at a local stage, 1.19 (95% CI 1.14-1.23) at a regional stage, and 1.05 (95% CI 1.01-1.10) at a distant stage. The associations were nonlinear, with substantially larger HRs at higher exposures [5]. In another American study, VoPham et al. [7] revealed that higher concentrations of ambient PM 2.5 exposure were associated with a statistically significant increased risk for HCC (HR 1.26 associated with a 10 µg/m 3 increase, 95% CI 1.08-1.47). Nevertheless, the exact mechanisms of PM 2.5 -mediated HCC migration and invasion remain unclear. In a laboratory study using HCC cell lines, Zhang et al. [8] revealed that PM 2.5 treatment not only stimulated the migration and invasion of HCC cells, but also increased the levels of matrix metalloproteinase (MMP)-13. Furthermore, PM 2.5 increased oxidative stress by induction of intracellular reactive oxygen species formation in HCC cells. The phosphorylation of RAC-alpha serine/threonine-protein kinase (AKT) increased in response to PM 2.5 . High concentrations of PM 2.5 decreased the proliferation of normal HL7702 hepatocyte-like cells and promoted apoptosis. Therefore, the activation of AKT by PM 2.5 resulted in MMP-13 overexpression, and stimulated HCC cell migration and invasion [8]. A previous study also indicated that the carcinogenicity of PM 2.5 might act through its collective effect on the suppression of DNA repair and augmentation of DNA replication errors [9].
Air pollution is a severe public health problem in Taiwan. Moreover, Taiwan is an endemic area for liver disease and HCC [10,11]. Therefore, the objective of this study was to examine the long-term effect of ambient PM 2.5 exposure on mortality in this sensitive population.

Ethical Statement
This retrospective cohort study complied with the guidelines of the Declaration of Helsinki and was approved by the Medical Ethics Committee of Chang Gung Memorial Hospital, Linkou, Taiwan. Since this study included retrospective evaluation of existing data, the Institutional Review Board approval (Institutional Review Board No. 201700631B0) was acquired, but without specific informed consent from patients. However, all individual data was protected (by delinking identifying information from main data set) and accessible to researchers only. Additionally, all the data were examined namelessly. The Institutional Review Board of Chang Gung Memorial Hospital has waived the need for consent. Lastly, all primary data were gathered according to epidemiology guidelines aimed at strengthening the reporting of observational studies.

Inclusion and Exclusion Criteria
A total of 1003 patients with HCC who were treated at Chang Gung Memorial Hospital between 2000 and 2009 were included in this study. Patients were grouped according to their yearly average ambient PM 2.5 exposure as <36 µg/m 3 (N = 870) or ≥36 µg/m 3 (N = 133). The choice of this PM 2.5 cutoff value was based on the study of Pan et al. [4]. The medical records were reviewed to obtain information, such as gender, age, presence of liver cirrhosis, alcohol usage, number of tumors, largest tumor size, presence of ascites upon surgery, alpha fetoprotein, albumin, bilirubin, prothrombin time, creatinine, aspartate aminotransferase, alanine aminotransferase, date of surgical resection, date of local recurrence, living place PM 2.5 concentrations, and date of the last follow-up or mortality. Patients with HCC aged less than 18 years, whose HCC was not primary, whose residential addresses were missing (without PM 2.5 data), or who died within 1 month were excluded from this study. The air quality monitoring data was presented real-time and archived as historical data on the web site. The PM 2.5 data were normally acquired from monitoring stations in the same area. If a patient lived between two monitoring stations, the PM 2.5 data from nearest station was chosen for analysis. If there was no monitoring station, the PM 2.5 data from the nearest station (within <15 km) was chosen.

Diagnosis of HCC
HCC was diagnosed clinically by testing for alpha fetoprotein, through imaging studies such as ultrasonography, radiocontrast-enhanced triphasic dynamic computed tomography, magnetic resonance imaging, angiography, and/or documented tissue histopathology [13]. The pretreatment diagnosis of HCC was made on the basis of dynamic imaging studies and biopsy, according to the guidelines of the American Association for the Study of Liver Diseases [14]. A biopsy was performed only if the HCC was not typical or if it was equivocal.

Barcelona Clinic Liver Cancer Staging
The HCC was staged according to the Barcelona Clinic Liver Cancer criteria [15] .

Follow-Up
Patients were followed up with clinic visits every 2-3 months during the first 2 years and every 3-6 months thereafter. At each follow-up visit, a complete history and physical examination were performed, a blood sample was drawn to test alpha fetoprotein levels and liver function, and the liver tumor was monitored using ultrasonography and chest radiographs.

Statistical Analysis
Continuous variables were expressed as mean and standard deviation for the number of observations, whereas categorical variables were expressed as numbers and percentages in brackets. Student's t-test was used for quantitative variables, whereas the chi-squared or Fisher's exact test were used for categorical variables. A multivariate Cox proportional hazards model was used for the analysis of mortality risk. Survival data were analyzed using the Kaplan-Meier method, and their significance was tested using the log-rank test. Comparisons of survival durations were made using the log-rank test. The predictive performance was evaluated using the area under the receiver operating characteristic (ROC) curve. A P value of less than 0.05 was considered statistically significant. The data was analyzed using R software.

Discussion
As presented in Table 1, the literature on the health effects of PM 2.5 on patients with HCC has been limited. Not only did the present study investigate a large patient population (N = 1003), but it also indicated that PM 2.5 ≥ 36 µg/m 3 [1.584 (1.162-2.160), p = 0.004] was a significant risk factor for mortality. The patients with HCC exposed to ambient PM 2.5 concentrations of ≥36 µg/m 3 had a 1.584-fold higher risk of death than those exposed to PM 2.5 concentrations of <36 µg/m 3 .
The positive association between ambient PM 2.5 exposure and mortality is consistent with the data reported by other groups (Table 1). Unlike in the study of Pan et al. [4], the blood alanine aminotransferase concentration was not a significant predictor for mortality according to the Cox regression analysis (Tables 3 and 4). Although the mean blood concentrations of alanine aminotransferase were slightly higher in patients living in environments with PM 2.5 concentrations of ≥36 µg/m 3 than in those living in environments with PM 2.5 concentrations of <36 µg/m 3 , the difference was not significant (71.98 (134.46) versus 68.13 (86.50), p = 0.66, Table 2).
According to Wong et al. [3], PM 2.5 -associated mortality could possibly be [16] a result of oxidative stress induced by PM 2.5 on epithelial cells creating reactive oxygen species that can injure DNA, proteins, and lipids. Another explanation is [17] that PM 2.5 -induced inflammation leads to the production of chemokines and cytokines that activate angiogenesis, enabling the epithelial invasion of metastatic cells and the persistence of attacking cells in distant tissues.
The multivariate Cox regression model revealed that PM 2.5 ≥ 36 µg/m 3 (p = 0.004), Child-Pugh score (p < 0.001), albumin (p < 0.001), macrovascular invasion (p < 0.001), tumor number (p < 0.001), and tumor size (p < 0.001) were critical risk factors for mortality. The positive association between the Child-Pugh score and mortality as well as the negative association between the blood albumin level and mortality were reasonable because both variables reflect the status of hepatic reserve. The Child-Pugh score has been widely used to assess the severity of liver dysfunction in clinical practice [18].
In a pioneering study, Deng et al. [6] reported that exposure to elevated PM 2.5 after the diagnosis of HCC shortens patient survival, with larger effects at higher concentrations. Notably, the study demonstrated that the associations between PM 2.5 exposure and mortality were nonlinear, with substantially larger risks at higher exposures. Similarly, a nonlinear relationship between odds ratio and PM 2.5 , with larger risks at higher exposures, was also observed in our study. The odds ratio was 1.137 (1.015-1.264) for each increment of 5 µg/m 3 in PM 2.5 or 1.292 (1.030-1.598) for each increment of 10 µg/m 3 in PM 2.5 .
A positive association was observed between mortality and macrovascular invasion (p < 0.001), tumor number (p < 0.001), and tumor size (p < 0.001). In a study of 104 patients with HCC Barcelona Clinic Liver Cancer stage B after hepatectomy [19], microvessel invasion, lymph node metastasis, and multiple lesions were risk factors for mortality (p < 0.05). In another study [20], the amount of vascular invasion based on the presence of microvascular invasion and gross invasion was associated with tumor recurrence and mortality. HCC growing larger than 3 cm in diameter is considered [21] a key turning point in the transformation of a tumor from having relatively benign features to more aggressive behaviors. Clinical evidence [22] also indicated that patients with HCC with tumors measuring >3 cm in diameter have an increased risk of microvascular invasion and satellite nodules. Furthermore, Asaoka et al. [23] reported that the median survival time after the development of vascular invasion was only 6 months. In a retrospective study, Montasser et al. [24] surveyed 105 patients with HCC, representing 138 lesions, who underwent radiofrequency ablation and were followed up for at least 1 year. Intrahepatic distant recurrence developed in 62 (59.0%) of the patients. Both a tumor size of >2.8 cm and multinodular tumors were significant risk factors for intrahepatic distant recurrence within 1 year. In a study of 554 patients with HCC, Tsuchiya et al. [25] reported that a higher rate of recurrence was noted in patients who had a larger tumor size (>2 cm) and/or a higher serum alpha fetoprotein concentration (>100 ng/mL) after radiofrequency ablation.

Conclusions
In conclusion, this retrospective cohort study revealed that patients with HCC living in environments with PM 2.5 concentrations of ≥36 µg/m 3 had a higher mortality rate than patients living in environments with PM 2.5 concentrations of <36 µg/m 3 (p = 0.034), and PM 2.5 ≥ 36 µg/m 3 was a significant risk factor for mortality (p = 0.004). The patients with HCC living in environments with PM 2.5 concentrations of ≥36 µg/m 3 had a 1.584-fold higher risk of death than those living in environments with PM 2.5 concentrations of <36 µg/m 3 . A nonlinear relationship was observed between the odds ratio and PM 2.5 . The odds ratio was 1.137 (1.015-1.264) for each increment of 5 µg/m 3 in PM 2.5 or 1.292 (1.030-1.598) for each increment of 10 µg/m 3 in PM 2.5 . The limitations of this study include its retrospective nature, short follow-up duration, and small sample size. Further studies are warranted.