Interaction of benzo[a]pyrene with other risk factors in hepatocellular carcinoma: a case-control study in Xiamen, China

Interaction of benzo[a]pyrene with other risk factors in hepatocellular carcinoma: a case-control study in Xiamen, China

Annals of Epidemiology 24 (2014) 98e103 Contents lists available at ScienceDirect Annals of Epidemiology journal homepage: www.annalsofepidemiology...

227KB Sizes 0 Downloads 20 Views

Annals of Epidemiology 24 (2014) 98e103

Contents lists available at ScienceDirect

Annals of Epidemiology journal homepage: www.annalsofepidemiology.org

Original article

Interaction of benzo[a]pyrene with other risk factors in hepatocellular carcinoma: a case-control study in Xiamen, China Yanhua Su a, Benhua Zhao a, *,1, Fei Guo a, Zhao Bin b, Yue Yang a, Sheng Liu a, Yaofeng Han a, Jianjun Niu c, Xiayi Ke d, Ning Wang e, Xuesi Geng f, Changnan Jin f, Yichen Dai g, Yuanyuan Lin g a

State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics, School of Public Health, Xiamen University, Fujian, China Organ Transplantation Institute of Xiamen University, Fujian, China Xiamen Center for Disease Control and Prevention, Fujian, China d Institute of Child Health, University College London, London, UK e Department of Hematology, First Affiliated Hospital of Xiamen University, Fujian, China f Xiamen Hospital of Traditional Chinese Medicine, Fujian, China g Digestive System Department of the 174th Hospital of People’s Liberation Army, Xiamen, Fujian, China b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 24 July 2013 Accepted 29 October 2013 Available online 9 November 2013

Purpose: Large epidemiologic studies about the relationship between benzo[a]pyrene (B[a]P) and hepatocellular carcinoma (HCC) have been limited. B[a]P diol epoxide (BPDE) is a highly reactive metabolite of B[a]P that binds covalently to form DNA adducts. We evaluated the interaction between B[a]P exposure with other risk factors in HCC, in a case-control study of 345 HCC and 961 healthy controls. Methods: Concentration of BPDE-DNA adducts in blood was determined by enzyme-linked immunosorbent assay. The interaction between BPDE-DNA adducts and other risk factors on HCC were evaluated by multivariate logistic regression analysis. Results: Mean concentration of BPDE-DNA adducts in blood of cases was significantly higher than that of the controls. The risk of HCC increased with elevated concentration of BPDE-DNA adducts (x2 ¼ 203.57, Ptrend < .001) and the odds ratio was 7.44 (95% confidence interval, 5.29e10.45) for the first versus fourth quartile of adduct levels. The relative excess risk due to interaction between BPDE-DNA adducts and hepatitis B virus surface antigen and drinking was 34.71 and 54.92, and the attributable proportion due to interaction was 41.53% and 75.59%, respectively. Conclusions: The high level of BPDE-DNA adducts in blood is associated with HCC and that environmental exposure to B[a]P may increase the risk of HCC, especially among drinkers and populations with hepatitis B virus infection. Ó 2014 Elsevier Inc. All rights reserved.

Keywords: Hepatocellular carcinoma Benzo[a]pyrene DNA adduct Case-control study Risk factor

Introduction Hepatocellular carcinoma (HCC) is one of the most common malignant neoplasms worldwide, with rates that show considerable geographic variation. Along the southeastern coastline of China, the provinces of Jiangsu, Zhejiang, Fujan, Guangdong, and Guangxi Autonomous Region are well known to have a high incidence of HCC [1], and Chinese cases of HCC contribute to about 55% of the total number worldwide [2,3]. Over the past few decades, epidemiologic studies have suggested that the well-documented multiple etiologic factors associated with the development of HCC

Conflict of interest statement: The authors declare no conflict of interest. * Corresponding author. School of Public Health, Xiamen University, Xiamen 361005, Fujian, China. Tel.: þ86-592-2188682; fax: þ86-592-2181578. E-mail address: [email protected] (B. Zhao). 1 Joint first author. 1047-2797/$ e see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.annepidem.2013.10.019

are chronic hepatitis B virus (HBV) infection, aflatoxin B1 (AFB1), alcohol abuse, smoking, diabetes mellitus, oral contraceptives, sex, and genetic susceptibility [3e15]. However, viral infection cannot explain the overall risk of developing HCC, and AFB1-lysine albumin adducts (biomarkers of AFB1 exposure) can only be detected in the serum of a small number of HCC [16e18]. In general, 15%e50% of HCC cases remain idiopathic, which suggests that additional unidentified factors may be responsible for an increased risk for HCC. Studies have reported that an elevated HCC risk is associated with exposure to a class of ubiquitous environmental contaminants known as polycyclic aromatic hydrocarbons (PAHs) [19,20]. PAHs are formed during the incomplete combustion of coal, oil, gas, wood, or other carbon-containing organic substances in gasburning motor vehicle, wood-burning furnaces, cigarette smoke, industrial smoke or soot, and charcoal-broiled foods and are thus widely present in polluted air, water and soil, and the diet [21e23]. PAHs comprise a wide range of different compounds with different

Y. Su et al. / Annals of Epidemiology 24 (2014) 98e103

aromatic ring numbers. B[a]P is the most thoroughly studied PAH and is a procarcinogen in animals and humans [21]. It exhibits its biological effects through metabolic activation by cytochrome P450 (CYP450) enzymes and epoxide hydrolase to the highly reactive metabolite, B[a]P diol epoxide (BPDE), which is an electrophilic species that can bind covalently to DNA and irreversibly damage DNA by the formation of DNA adducts [21]. So, BPDE-DNA adducts in human body reflect both exposure to B[a]P and the body’s metabolic capacity. In addition, the formation of DNA adducts is considered as a crucial step in the initiation phase of carcinogenesis [22]. Studies have shown that exposure to high levels of environmental B[a]P is associated with increased risk of lung, breast cancer, hepatic angiosarcoma, and HCC [23e29].However, most studies have been small epidemiologic studies with no more than 100 cases with BPDE-DNA adducts examined in specific tumor tissue or cells [24e29]. So far, no large case-control studies have been conducted on the association between the BPDE-DNA adducts in peripheral blood and HCC in humans. Xiamen City is located on the Southeastern coastline of China and is an area of high HCC incidence. The specific aims of our study were to assess whether B[a]P exposure, as determined by BPDEDNA adduct levels in peripheral blood of subjects, is related to HCC risk; assess the interaction between BPDE-DNA adducts and other risk factors on HCC; and shed light on mechanisms implicated in the etiology of PAH-related cancers. Materials and methods Subject enrollment and sample collection A total of 407 HCC cases were newly diagnosed histopathologically at the Department of Hepatobiliary Surgery, Xiamen Hospital of Traditional Chinese Medicine and the 174th Hospital of People’s Liberation Army from March 2007 to December 2009. All the HCC cases were clinical patients who had not suffered any other types of cancer and were recruited at the time of diagnosis. A total of 1173 controls were selected simultaneously from healthy non-blood relatives (especially the patient’s spouses) of all hospitalized patients except of those who suffered from smoking-related diseases (such as HCC, gastrointestinal, lung or head and neck cancers, and so forth). The controls and cases were matched by age (5 years), gender, and ethnicity (Han Chinese), and so forth. All participants had been living at least 10 years in Xiamen and signed the written consent form to voluntarily provide 10 mL blood sample at the time of enrollment and participate in a structured questionnaire survey that collected demographic information through in person interviews by trained interviewers. Finally, the participation rate of cases and controls were 84.8%(345/407) and 81.9% (961/1173), respectively. The other 62 new patients of HCC and 212 controls were failed to be complete the survey, due to the following reasons: refusing the interview accounted for 69%; language barrier accounted for 13%; and interview in improper time accounted for 18%. A total of 1306 subjects were enrolled and there were no significant difference between the subjects of participation and of nonparticipation in age, gender, education level, marital status, and ethnicity etc. The samples of whole blood with anticoagulant were separated and stored at 70 C for the determination of BPDE-DNA adducts. This study was approved by Institutional Review Board, School of Public Health, Xiamen University, Xiamen City, Fujian province, China. Variables and definitions The structured questionnaire included demographics (sex, age, income, educational attainment, occupation, and marital status),

99

lifestyle factors (cigarette smoking, alcohol, and tea and coffee intake), contaminated water drinking, and family history of cancer. The main definitions of risk categories were as follows: (1) a cigarette smoker was a person who had smoked 1 or more cigarette per day and for at least 6 months; (2) an alcohol drinker was a person who had consumed beer, wine, or hard liquor at least once weekly for at least 6 months during their lifetime; (3) a tea drinker was a person who had consumed tea twice weekly for at least 6 months during their lifetime; and (4) a drinker of contaminated water was a person who did not drink tap water, but did drink pond, ditch, or well water in daily life. Quantification of BPDE-DNA adducts in peripheral blood For detecting BPDE-DNA adducts in peripheral blood, the DNA was first extracted from 10 mL whole blood according to the protocol of the DNA isolation kit for mammalian blood (Roche, Penzberg, Germany). The extracted DNA was checked by agarose gel electrophoresis for integrity and quantified by ultraviolet spectrophotometer. PAH-DNA was measured by an immunoperoxidase method described previously [30]. Briefly, the detection of BPDEDNA adducts of extracted DNA was conducted using the BPDEDNA adduct enzyme-linked immunosorbent assay kit (OxiSelect; Cell Biolabs, San Diego, CA) following the recommended protocol. The quantity of BPDE adduct in DNA samples was determined by relative comparison of a known BPDE-DNA standard curve. BPDEDNA standards or unknown DNA samples were adsorbed onto a 96-well DNA high-binding plate. The BPDE-DNA adducts present in the sample or standard were probed with an anti-BPDE-I antibody, followed by a horseradish peroxidaseeconjugated secondary antibody. Statistical analysis SARS software version 9.1 (SAS Institute, Cary, NC) was used for the statistical analyses. t and c2 tests were conducted to analyze the distribution of variables between cases and controls and estimate the relative risk as the odds ratio (OR) of HCC in relation to BPDEDNA adducts. Based on the fourth quartile of BPDE-DNA adducts in the controls, quartile division was applied to divide the subjects into four subgroups (quartiles 1e4) for trend analysis and elucidation of the dose response for HCC risk. Unconditional logistic regression analysis was performed to evaluate each possible risk factor for HCC. The unconditional multivariate logistic regression was performed when the exposed factor was significantly associated with HCC in the univariate analysis. The attributable proportion due to interaction (API) and the relative excess risk due to interaction (RERI) were used to evaluate the interaction of risk factors. The Mann-Whitney U test was used to test the significance of RERI. The significant criteria in the tests were two tailed at significance level P ¼ .05. The adjusted OR of age, sex, education, marital status, and the other HCC risk factors and 95% confidence interval (CI) were calculated by the maximum likelihood approach. The assessment of two factors’ interaction was done using the following equations:

APIðABÞ ¼ ½RRðABÞ  RRðAB0 Þ  RRðA0 BÞ þ 1=RRðABÞRERI ¼ RRðABÞ  RRðAB0 Þ  RRðA0 BÞ þ 1; where API (AB) is the API of factor A and factor B, and RERI is the relative excess risk of interaction. RR is the relative risk. AB means exposure to factor A and factor B; AB0 means exposure to factor A but not to factor B; A0B means exposure to factor B but not to factor A.

100

Y. Su et al. / Annals of Epidemiology 24 (2014) 98e103

Results

Table 2 Variable coding of the unconditional logistic regression

Subject demographics Table 1 lists the demographic data of all the HCC patients and controls. The patients were older (mean age: 60.2  13.3 years) than the control group (58.3  10.7 years; t test, P < .001). The education level and annual income were lower for the HCC patients compared with the controls. The rate of divorced, widowed, and single in HCC patients were significantly higher than that of the controls. Multivariate analysis To evaluate the comprehensive effect of different factors on HCC and control interference from potential confounding factors, we performed multiple factor unconditional logistic regression analysis (level of inclusion and exclusion ¼ 0.05). Many variables, such as age, occupation, education, marital status, smoking, drinking, hepatitis B virus surface antigen (HBsAg), and family history of cancer (Table 2), were introduced into the multivariate logistic regression because of their possible correlation with HCC. The results demonstrated that HBV infection, smoking, BPDE-DNA adducts and alcohol drinking were the risk factors for HCC (P & .05; Table 3). BPDE-DNA adduct levels in peripheral blood and HCC risk

Variables

Assignment instructions

X1 Age X2 Gender X3 Educational level

0 1 1 2 3 1 2 1 0 0 0 0 0 0 0 0 0

X4 Marital status X5 Occupation X6 Cigarette smoking X7 Alcohol drinking X8 Coffee intake X9 Long-term tea drinking X10 Contaminated water drinking X11 Pickle intake X12 Hair-dyeing X13 HBsAg X14 Family history of cancer

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

40 y; 1 ¼ >40 y Woman; 2 ¼ Man Illiteracy or primary school; Middle school; High school graduate or more Married; Divorced, widowed, single Farmer; 2 ¼ Worker; 3 ¼ Others No; 1 ¼ Yes No; 1 ¼ Yes No; 1 ¼ Yes No; 1 ¼ Yes No; 1 ¼ Yes No; 1 ¼ Yes No; 1 ¼ Yes Negative; 1 ¼ Positive No; 1 ¼ Yes

the second, third, and fourth quartile, respectively, compared with adduct level in the first quartile (x2 ¼ 203.61, Ptrend < .01; Table 4). Interactions of BPDE-DNA adducts and other risk factors in HCC

The association of BPDE-DNA adducts with HCC risk is shown in Table 4. The analyses of BPDE-DNA adduct levels were based on the log-transformed data. The mean level of BPDE-DNA adducts in HCC patients was 0.65  0.22 fmol/mg, which was significantly higher than that in the controls (0.35  0.25 fmol/mg) (t ¼ 8.151, P < .05). The result of trend test showed that the risk of HCC increased with adjusted OR of 1.78 (95% CI, 1.18e2.68), 5.56 (95% CI, 3.92e7.87), and 7.44 (95% CI, 5.29e10.45) for subjects with BPDE-DNA adducts in

Table 5 lists the results of analyzing the interaction between paired risk factors among the groups divided according to the mean level of BPDE-DNA adducts in controls (0.35  0.25 fmol/mg). Multivariate logistic regression analysis was performed to estimate the interaction between BPDE-DNA adducts levels and HBsAg and alcohol drinking on HCC. The results showed that the interactions of these factors were a positive additive model, and RERI was 34.71 and 54.92, respectively. When BPDE-DNA adducts were combined with HBsAg and alcohol drinking, the attributable proportion of interaction for HCC development was 41.53% and 75.59%, respectively.

Table 1 Characteristics of the HCC patients and controls

Reproducibility of our experiment

Variable Gender Male Female Age (y) 40 41e50 51e59 60e69 70 Mean age (y) Education level Illiterate Primary school Middle school High school College degree Marital status Married Divorced/widowed/single Income annual (RMB) 5000 5001e10,000 10,001e20,000 20,001 Smoking Yes No Alcohol drinking Yes No RMB ¼ Ren Min Bi.

Case (%)

Control (%)

P

243 (70.44) 102 (29.56)

609 (63.37) 352 (36.63)

.020

12 (3.48) 43 (12.46) 97 (28.12) 112 (32.46) 81 (23.48) 60.22  13.31

46 (4.78) 163 (16.96) 287 (29.86) 306 (31.85) 159 (16.55) 58.34  10.73

.030

38 131 139 27 10

86 360 382 114 19

(8.94) (37.46) (39.76) (11.87) (1.97)

.008

289 (83.76) 56 (16.24)

865 (90.01) 96 (9.99)

.002

113 59 98 75

199 185 304 273

(20.71) (19.25) (31.63) (28.41)

.000

256 (74.21) 89 (25.79)

264 (27.47) 697 (72.53)

.000

196 (56.81) 149 (43.19)

168 (17.48) 793 (82.52)

.000

The reproducibility of detecting BPDE-DNA adducts by an immunoperoxidase method was carried out as follows: The BPDEDNA adducts of peripheral blood from randomly selected 15 cases and 30 controls were detected repeatedly by two laboratory operator in two separate days. Results showed that inter- and intracorrelation coefficients of the experiments were 0.95 and 0.94, respectively, and indicating a high level of consistency and repeatability of the laboratory method.

.001

Discussion (11.01) (37.97) (40.28) (7.84) (2.90)

(32.75) (17.10) (28.41) (21.74)

Studies have suggested that the development of HCC is a complex and multistep process with a multifactorial etiology [3e15].

Table 3 Multivariate logistic regression analysis of the relationship between risk factors and HCC Risk factor HBsAg (þ) Alcohol drinking Smoking BPDE-DNA adducts Long-term tea drinking Drinking contaminated water

b

SE (b)

4.41 2.00 2.62 1.74 1.50 0.49

0.97 0.78 0.91 0.92 1.15 0.86

Wald

OR (95% CI)

P

81.55 7.42 13.71 5.67 0.22 1.62

.000 .010 .004 .049 .191 .582

c2 20.54 6.57 8.35 3.56 1.72 0.32

b ¼ regression coefficient; SE (b) ¼ standard error of b.

(12.16e546.84) (1.60e34.35) (2.32e80.94) (1.14e34.44) (0.02e2.14) (0.30e8.82)

Y. Su et al. / Annals of Epidemiology 24 (2014) 98e103

101

Table 4 Quartile grouping of dose-dependent risks from BPDE-DNA adducts for HCC BPDE-DNA adducts (fmol/mg)

Cases 345 (%)

Controls 961 (%)

ROR

AOR*

95% CI

c2

1

d

d

Q1 (<0.31) Q2 (0.31e0.48)

34 (9.9) 51 (14.8)

240 (25.0) 240 (25.0)

1 1.50

Q3 (0.48e0.71)

123 (35.7)

241 (25.0)

3.62

P 7.75

d .005

127.61

.000

204.23

.000

203.61

.000

1.78 (1.18e2.68) 5.56 (3.92e7.87) Q4 (>0.71)

137 (39.7)

240 (25.0)

4.03 7.44 (5.29e10.45)

Trend analysis AOR ¼ adjusted odds ratio; Q ¼ quartile; ROR ¼ raw odds ratio. * AOR adjusted for age, sex, education level, HBsAg, alcohol drinking, cigarette smoking, and contaminated water drinking.

However, no large case-control studies have been performed to investigate the association between BPDE-DNA adducts in peripheral blood DNA and HCC in humans up to date. We, therefore, designed a case-control study to investigate the association between the incidence of HCC and exposure to B[a]P by examining the BPDE-DNA adduct levels in peripheral blood DNA. Our study indicated that HBV infection, smoking, alcohol drinking, and BPDE-DNA adducts were the major risk factors for HCC in Xiamen City (Table 3), which is located at the center of the high-risk area for HCC in China. There is strong evidence of PAHs carcinogenicity in animals [25e28] and an emphasis has been placed on the genotoxic properties of B[a]P. B[a]P is notable for being the first chemical procarcinogen to be discovered by the International Agency of Research on Cancer. Epidemiologic studies have shown increased mortality due to lung cancer in humans exposed to coke oven emissions, roofing-tar emissions, and cigarette smoke. Each of these mixtures contains B[a]P [31,32]. Although there is increasing recognition of the role of B[a]P in the etiology of many types of carcinoma, the epidemiologic data about the relationship between B[a]P and HCC occurrence are still limited. The main reason for this is that few accurate methods are available to measure exposure to B [a]P in epidemiologic studies. Many chemical carcinogens, including B[a]P, can be converted to reactive electrophilic metabolites by oxidative enzymes, mainly CYP450, and can bind to DNA and protein to form adducts [33e35] in humans. DNA adducts represent premutagenic lesions that play an essential role in the initiation phase of carcinogenesis [22]. Thus, high levels of BPDE-DNA adducts are indicative of susceptibility to HCC and may be a better indicator of the body’s response to carcinogenic exposure. The PAHTable 5 Multivariate logistic regression analysis of interactions of BPDE-DNA Adducts with known risk factors of HCC Variables

BPDE-DNA adduct (fmol/mg)

Cases/ controls (345/961)

ROR

HBsAg Negative <0.35 21/649 1 Positive <0.35 64/60 32.96 Negative 0.35 65/180 11.16 Positive 0.35 195/72 83.70 Interaction: API (AB) ¼ 41.53%; RERI ¼ 34.71; u ¼ 3.10; Alcohol drinking No <0.35 30/576 1 Yes <0.35 56/132 8.14 No 0.35 119/216 10.58 Yes 0.35 140/37 72.65 Interaction: API (AB) ¼ 75.59%; RERI ¼ 54.92; u ¼ 4.05;

AOR (95% CI)

1 38.72 (18.00e84.71)* 11.14 (6.46e19.38)* 83.57 (48.79e144.43)* P ¼ .001 1 8.15 (4.90e13.58)y 10.58 (6.75e16.65)y 72.65 (42.11e126.20)y P ¼ .000

AOR ¼ adjusted odds ratio; ROR ¼ raw odds ratio. * AOR adjusted for age, sex, education level, alcohol drinking, cigarette smoking, and contaminated water drinking. y AOR adjusted for age, sex, education level, HBsAg, cigarette smoking, and contaminated water drinking.

DNA adducts have been used as biomarkers to monitor PAH exposure and provide relatively precise data on exposure levels from all routes [36,37]. Cigarette smoking and consumption of smoked and grilled foods have been reported as two important contributors to body levels of PAH-DNA adducts [38]. The work of measurement of the association between BPDE-DNA adduct levels and major sources of PAH, such as smoking, was complicated and was not carried out in our study. It is, therefore, possible that the BPDE-DNA adduct levels observed in our study were associated with different exposures of PAHs, such as smoking, air pollution, and grilled foods, and so forth. Sensitive methods are available to detect BPDE-DNA adducts in the blood, serum, and other tissues of humans following exposure to B[a]P. Immunoassays are currently being developed to detect the presence of carcinogenic PAH adducts bound covalently to macromolecules, which are accurate, precise, reproducible, and sensitive enough to measure the PAH adducts in the body. Research have used immunoassay techniques for detecting PAH-DNA adducts in the blood and tissues of humans occupationally exposed to PAHs and clearly demonstrate that lowering exposure to PAHs results in decreased levels of PAH-DNA adducts [21,30,39]. In our study, we found that the mean level of BPDE-DNA adducts in HCC was significantly higher than that of the controls, which indicated that HCC were probably exposed to more B[a]P from different environmental sources than controls in Xiamen. Our study showed that high levels of BPDE-DNA adducts were related to HCC development regardless of whether the PAH levels were grouped into two (cases and controls) or four (quartiles 1e4). There was also a trend toward increased risk for HCC with elevated level of BPDEDNA adducts, suggesting that adducts are independent risk factors for HCC risk (Table 4). Our study also evaluated the joint effect of PAH and other major risk factors on HCC risk. Our result showed that there was a significant positive synergistic interaction between BPDE-DNA adducts and HBsAg and alcohol drinking (Table 5). This supports the hypothesis that people with a combined effect of PAH and these factors are more vulnerable to HCC. PAHs are widely distributed in the environment because of coal-dependent energy consumption, serious air pollution with fine particulate material, and open cooking with wood or coal combustion in China, and exposure is virtually unavoidable. The precise role for the enhanced effect of PAH on the risk of HCC in people with viral infection (HbsAg positive) might be due to their double effect on the liver, which is the most important metabolic organ for most xenobiotics. Our results showed an underlying biological interaction between B[a]P exposure and HBV infection for HCC risk, suggesting that B[a]P plays a role in human hepatocarcinogenesis in conjunction with HbsAg carrier status. The positive interaction of BPDE-DNA adducts and alcohol drinking may be attributed to their coincident action with HCC. It has been shown that alcohol abuse can cause an increase in PAH-DNA adducts in the body and there is a significant dose response between alcohol and HCC [40,41]. Alcohol drinking

102

Y. Su et al. / Annals of Epidemiology 24 (2014) 98e103

may promote HCC development combined with BPDE-DNA adducts through various plausible potential mechanisms including acetaldehyde production, induced activity of CYP450, glutathione depletion, S-adenosylmethionine depletion, dysregulated retinoid metabolism, and induction of liver fibrosis, and so forth. Acetaldehyde may promote HCC by forming mutagenic/carcinogenic DNA adducts; CYP2E1 induction is likely to promote hepatic carcinogenesis through increased reactive oxygen species generation, lipid peroxidation, inducing mutation of p53 gene, and conversion of procarcinogens to carcinogens. Chronic drinking also can cause nutritional deficiencies of certain nutrients in the body, such as glutathione depletion, hepatic S-adenosylmethionine depletion, and retinoic acid low level and thus may render liver susceptible to many liver diseases, including cancer [5,40,41]. In our study, selection bias and information bias were stringently controlled. Potential confounding factors were ruled out using stratified analysis or multiple factor analysis. There were several limitations for the present study. First, as discussed above, the levels of BPDE-DNA adducts were compound measurements for all the major sources of PAH, thus it was not clear what were the relevant contribution to the BPDE-DNA adduct levels in an individual of the different exposures of PAHs, such as smoking, air pollution, and grilled foods. Second, our study is a case-control study, which is an analytical epidemiologic research method that has its inherent limitations and disadvantages. Case-control studies are retrospective and, therefore, cannot be used to identify the exact chronological order of the rise of BPDE-DNA adducts in peripheral blood and the incidence of HCC case, so it is difficult to establish the temporal relationship between the BPDE-DNA adducts and HCC in our study. More specifically, HCC cases in the study were recruited at the time of diagnosis of HCC, so the level of BPDE-DNA adducts could represent an effect rather than a cause of the disease. Conclusion Our study provides evidence that the presence of high levels of BPDE-adducts is associated with HCC development and suggests that environmental exposure to B[a]P increases the risk of HCC, especially among alcohol drinkers and populations with HBV infection. It provides important clues for the prevention of HCC. In addition to universal HBV vaccination of all newborn infants, reduction of exposure to other hepatocarcinogens (e.g., by controlling environmental pollution), reduction or quitting of cigarette smoking, and reduction of alcohol drinking may all be effective ways to prevent HCC in China. Acknowledgments This work was supported by the Xiamen Municipal Science and Technology Program (3502Z20073015). We thank all who contributed their samples and work in this study. The authors thank Prof. Chengyi Qu (School of Public Health, Shanxi Medical College, People’s Republic of China) for his valuable comments on the manuscript. References [1] Yeh FS, Yu MC, Mo CC, Luo S, Tong MJ, Henderson BE. Hepatitis B virus, aflatoxins, and hepatocellular carcinoma in southern Guangxi, China. Cancer Res 1989;49:2506e9. [2] Chen W, Zheng R, Zhang S, Zhao P, Li G, Wu L, et al. Report of incidence and mortality in China cancer registries, 2009. Chin J Cancer Res 2013;25:10e21. [3] Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;127:2893e917. [4] Chen CJ, Liang KY, Chang AS, Chang YC, Lu SN, Liaw YF, et al. Effects of hepatitis B virus, alcohol drinking, cigarette smoking and familial tendency on hepatocellular carcinoma. Hepatology 1991;13:398e406.

[5] Donato F, Tagger A, Gelatti U, Parrinello G, Boffetta P, Albertini A, et al. Alcohol and hepatocellular carcinoma: the effect of lifetime intake and hepatitis virus infections in men and women. Amer J Epidemiol 2002;155:323e31. [6] Hassan MM, Spitz MR, Thomas MB, El-Deeb AS, Glover KY, Nguyen NT, et al. Effect of different types of smoking and synergism with Hepatitis C virus on risk of hepatocellular carcinoma in American men and women: case-control study. Int J Cancer 2008;123:1883e91. [7] Liu Y, Wu F. Global burden of aflatoxin-induced hepatocellular carcinoma: a risk assessment. Environ Health Perspect 2010;118:818e24. [8] Yu MW, Chen CJ. Hepatitis B and C viruses in the development of hepatocellular carcinoma. Crit Rev Oncol Hematol 1994;17:71e91. [9] Li FQ, Yoshizawa T, Kawamura O, Luo XY, Li YW. Aflatoxins and fumonisins in corn from the high-incidence area for human hepatocellular carcinoma in Guangxi, China. J Agric Food Chem 2001;49:4122e6. [10] Wang J, Liu XM. Contamination of aflatoxins in different kinds of foods in China. Biomed Environ.Sci 2007;20:483e7. [11] Wang JS, Huang T, Su J, Liang F, Wei Z, Liang Y, et al. Hepatocellular carcinoma and aflatoxin exposure in Zhuqing Village, Fusui County, People’s Republic of China. Cancer Epidemiol Biomarker Prev 2001;10:143e6. [12] Wang LY, Hatch M, Chen CJ, Levin B, You SL, Lu SN, et al. Aflatoxin exposure and risk of hepatocellular carcinoma in Taiwan. Int J Cancer 1996;67:620e5. [13] Qian GS, Ross RK, Yu MC, Yuan JM, Gao YT, Henderson BE, et al. A follow-up study of urinary markers of aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev 1994;3:3e10. [14] Szymanska K, Chen JG, Cui Y, Gong YY, Turner PC, Villar S, et al. TP53 R249S mutations, exposure to aflatoxin, and occurrence of hepatocellular carcinoma in a cohort of chronic hepatitis B virus carriers from Qidong, China. Cancer Epidemiol Biomarkers Prev 2009;18:1638e43. [15] Wu HC, Wang Q, Yang HI, Ahsan H, Tsai WY, Wang LY, et al. Urinary 15-F-2tisoprostane, aflatoxin B-1 exposure and hepatitis B virus infection and hepatocellular carcinoma in Taiwan. Carcinogenesis 2008;29:971e6. [16] Groopman JD, Johnson D, Kensler TW. Aflatoxin and hepatitis B virus biomarkers: a paradigm for complex environmental exposures and cancer risk. Cancer Biomark 2005;1:5e14. [17] Qu CX, Kamangar F, Fan JH, Yu B, Sun XD, Taylor PR, et al. Chemoprevention of primary liver cancer: a randomized, double-blind trial in Linxian, China. J Natl Cancer Inst 2007;99:1240e7. [18] Welzel TM, Katki HA, Sakoda LC, Evans AA, London WT, Chen G, et al. Blood folate levels and risk of liver damage and hepatocellular carcinoma in a prospective high-risk cohort. Cancer. Epidemiol. Biomarkers Prev 2007;16:1279e82. [19] McGlynn KA, London WT. Epidemiology and natural history of hepatocellular carcinoma. Best Pract Res Clin Gastroenterol 2005;19:3e23. [20] Wu HC, Wang Q, Wang LW, Yang HI, Ahsan H, Tsai WY, et al. Polycyclic aromatic hydrocarbon- and aflatoxin-albumin adducts, hepatitis B virus infection and hepatocellular carcinoma in Taiwan. Cancer Lett 2007;252:104e14. [21] ATSDR. Toxicological profile for toxicological profile for polycyclic aromatic hydrocarbons. 1995. http://www.atsdr.cdc.gov/toxprofiles/tp35.pdf. [22] Zhou GD, Richardson M, Fazili IS, Wang J, Donnelly KC, Wang F, Amendt B, et al. Role of retinoic acid in the modulation of benzo(a)pyrene-DNA adducts in human hepatoma cells: implications for cancer prevention. Toxicol Appl Pharmacol 2010;249:224e30. [23] Chen SY, Wang LY, Lunn RM, Tsai WY, Lee PH, Lee CS, et al. Polycyclic aromatic hydrocarbon-DNA adducts in liver tissues of hepatocellular carcinoma patients and controls. Int J Cancer 2002;99:14e21. [24] Zhang YJ, Rossner Jr P, Chen Y, Agrawal M, Wang Q, Wang L, et al. Aflatoxin B1 and polycyclic aromatic hydrocarbon adducts, p53 mutations and p16 methylation in liver tissue and plasma of hepatocellular carcinoma patients. Int J Cancer 2006;119:985e91. [25] Rundle A, Schnabel F, Kelly A, Levine R, et al. Molecular epidemiologic studies of polycyclic aromatic hydrocarbon-DNA adducts and breast cancer. Environ Mol Mutagen 2002;39:201e7. [26] Tang D, Santella RM, Blackwood AM, Young TL, Mayer J, Jaretzki A, et al. A molecular epidemiological case-control study of lung cancer. Cancer Epidemiol Biomarkers Prev 1995;4:341e6. [27] Tang D, Phillips DH, Stampfer M, Mooney LA, Hsu Y, Cho S, et al. Association between carcinogen-DNA adducts in white blood cells and lung cancer risk in the physicians health study. Cancer Res 2001;61:6708e12. [28] Veglia F, Matullo G, Vineis P. Bulky DNA adducts and risk of cancer: a metaanalysis. Cancer Epidemiol Biomarkers Prev 2003;12:157e60. [29] Perera FP, Hemminki K, Young TL, Brenner D, Kelly G, Santella RM. Detection of polycyclic aromatic hydrocarbon-DNA adducts in white blood cells of foundry workers. Cancer Res 1988;48:2288e91. [30] Gammon MD, Santella RM, Neugut AI, Eng SM, Teitelbaum SL, Paykin A, et al. Environmental toxins and breast cancer on Long Island. I. Polycyclic aromatic hydrocarbon DNA adducts. Cancer Epidemio Biomarkers Prev 2002;11:677e85. [31] Johnson NM, Qian G, Xu L, Tietze D, Marroquin-Cardona A, Robinson A, et al. Aflatoxin and PAH exposure biomarkers in a U.S. population with a high incidence of hepatocellular carcinoma. Sci Total Environ 2010;408:6027e31. [32] Thyssen J, Althoff J, Kimmerle G, Mohr U. Inhalation studies with benzo[a] pyrene in Syrian golden hamsters. J Natl Cancer Inst 1981;66:575e7.

Y. Su et al. / Annals of Epidemiology 24 (2014) 98e103 [33] Autrup H, Vestergaard AB, Okkels H. Transplacental transfer of environmental genotoxins: polycyclic aromatic hydrocarbon-albumin in nonsmoking women, and the effect of maternal GSTM1 genotype. Carcinogenesis 1995;16:1305e9. [34] Tang D, Warburton D, Tannenbaum SR, Skipper P, Santella RM, Cereijido GS, et al. Molecular and genetic damage from environmental tobacco smoke in young children. Cancer Epidemiol Biomarkers Prev 1999;8:427e31. [35] Autrup H, Daneshvar B, Dragsted LO, Gamborg M, Hansen M, Loft S, et al. Biomarkers for exposure to ambient air pollutiondcomparison of carcinogenDNA adduct levels with other exposure markers and markers for oxidative stress. Environ Health Perspect 1999;107:233e8. [36] Godschalk RW, Van Schooten FJ, Bartsch H. A critical evaluation of DNA adducts as biological markers for human exposure to polycyclic aromatic compounds. J Biochem Mol Biol 2003;36:1e11.

103

[37] Godschal R, Nair J, van Schooten FJ, Risch A, Drings P, Kayser K, et al. Comparison of multiple DNA adduct types in tumor adjacent human lung tissue: effect of cigarette smoking. Carcinogenesis 2002;23:2081e6. [38] Poirier MC. DNA adducts as exposure biomarkers and indicators of cancer risk. Environ Health Perspect 1997;105:907e12. [39] Rothman N, Correa-Villaseñor A, Ford DP, Poirier MC, Haas R, Hansen JA, et al. Contribution of occupation and diet to white blood cell polycyclic aromatic hydrocarbon-DNA adducts in wildland firefighters. Cancer Epidemiol Biomarkers Prev 1993;2:341e7. [40] Purohit V, Rapaka R, Kwon OS, Song BJ. Roles of alcohol and tobacco exposure in the development of hepatocellular carcinoma. Life Sci 2013;92:3e9. [41] Boffetta P, Hashibe M, La Vecchia C, Zatonski W, Rehm J. The burden of cancer attributable to alcohol drinking. In J Cancer 2006;119:884e7.