Esophageal cancer is the eighth most commonly occurring cancer and the sixth leading cause of cancer deaths in the world, with the majority of cases occurring in developing countries [1]. China, with ~250,000 cases diagnosed yearly, lies in the "esophageal cancer belt" [2], and contributes to about half of the world's cases [3]. The prevalence of esophageal cancer varies greatly across China, but there are two major endemic areas: the southern Taihang Mountain area (Linxian, Henan Province) and the northern Jiangsu area. In the northern Jiangsu area, the esophageal cancer incidence is over 80/100,000 which is six times greater than the national average rate (13/100,000). Many studies have been conducted in the southern Taihang Mountain area; however, despite very distinct geographic environments, fewer studies have been done in the northern Jiangsu area. Since esophageal squamous cell carcinoma (ESCC) represents more than 99% of esophageal cancer cases in China [4], this study focused only on ESCC.

Previous epidemiological studies have demonstrated that exposure to environmental carcinogens plays a significant role in the etiology of ESCC [5–7]. The recognized risk factors for ESCC include excessive use of tobacco, alcohol consumption, and the consumption of salt-pickled, salt-cured or moldy foods [8–10]. Other factors associated with increased risk of ESCC include vitamin and trace mineral deficiencies [11]. On the other hand, genetic polymorphisms may also play an important role in the development of ESCC and in determining individual susceptibility to carcinogens. Many studies have suggested that polymorphisms in xenobiotic-metabolizing enzymes (XME) including phase I enzymes such as CYP1A1, CYP1B1, CYP2A6 and CYP2E1, and phase II enzymes such as GSTM1, GSTT1, GSTP1, and microsomal epoxide hydrolase (EPHX) may confer different risks of susceptibility to cancer [12–20], including ESCC [21–23]. For example, genetic variations, including CYP2E1 c1/c1 allele, GSTM1 and GSTT1 non-null genotype and EPHX slow allele, were associated with increased risk of ESCC, while CYP1A1 3' polymorphism (variant allele) may be one of the protective factors for ESCC [22, 23]. It is clear that many cancers including ESCC, are the result of complex interactions between inherited and environmental factors [24].

Although extensive studies have been carried out during the last two to three decades in the southern Taihang Mountain area [8, 11, 22, 23], data are scarce from the northern Jiangsu area [21]. In the present study, we conducted a population based case-control study of ESCC in the Huaian area. In addition to regular epidemiological and food frequency questionnaire analyses, we investigated the polymorphisms of several phase I metabolic enzymes, including CYP1A1, CYP1B1, CYP2A6, and CYP2E1, and phase II metabolic enzymes, including GSTM1, GSTT1, GSTP1 and EPHX, which are believed to be involved in the metabolism of the environmental carcinogens for many types of cancers. The overall goal of this study is to evaluate the etiological roles of dietary carcinogen exposure and XEM polymorphisms, and to investigate possible gene-gene and gene-environment interactions associated with ESCC in this high risk area.

In the present study, introverted personality, passive smoking, a family history of cancer, esophageal lesion, HP infection, eating fast, fatty meat consumption, pickled vegetables intake, moldy food intake, favor to salty and acrid food were all associated with increased risk of ESCC. However, regular clean up of food storage utensils, alcohol abstinence lasted for more than half a year, and tea drinking were protective for ESCC. Among the factors associated with ESCC in this study, an introverted personality type, clean up of food storage utensils, and alcohol abstinence are first reported. Other major findings are consistent with previous studies conducted in China and elsewhere [31–33] including studies conducted in the Huaian area [34, 35]. For instance, previous epidemiological surveys have shown that people in this area used to consume a lot of salted food and pickled vegetables. These foods are a rich source of nitrosamine compounds, and have been reported to be risk factors for ESCC [36]. Moldy foods are frequently contaminated by fungal toxins, and it has long been reported that mycotoxins, especially Fumonisin B₁ (FB₁), are associated with ESCC [37]. This may also be an explanation for our finding that maintaining clean storage utensils reduces the risk of ESCC. Damp and dirty environments favor the growth of mold and mildew, and regular clean up of the utensils may be a good habit to prevent food from developing mildew and molds. Eating fast, as well as frequently consuming acrid foods, which are common physical stimulations to the esophagus, most often will lead to the chronic esophagitis. This has been considered the earliest tissue perturbation in the progressive process leading to malignant transformation of the squamous epithelium of the esophagus [38].

We found that passive smoking is significantly associated with increased risk of ESCC. Passive smoking is a major risk factor for lung cancer, and emerging studies have shown that it may also be related to increased risk of breast cancer in women [39]. It has been suggested that passive smoking or environmental tobacco smoke (ETS), 85% of which consisted by sidestream smoke due to burning cigarette, was more risky than active smoking, and concentrations of some carcinogens are higher in sidestream than in mainstream smoke, which is exhaled by the smoker [40]. ETS has been classified as a Group A carcinogen which means it is known to cause cancer in humans [41]. The number of known carcinogens in ETS has been reported to be 69 in 2000 [42]. The role of active smoking has been examined by age started, smoking intensity, years smoked, cessation, type of tobacco, and type of cigarette. No association between active smoking and ESCC was found.

We also found that alcohol abstinence is a protective factor for ESCC. It is possible that genetic damage such as chromosome aberrations and micronuclei caused by alcohol are reversible by abstinence [43]. Although several epidemiological studies have showed that exposure to tobacco smoking and alcohol drinking are major risk factors for ESCC [44–46], we did not find any association between smoking and alcohol drinking and the risk of ESCC. It may be resulted from that men usually smoke and drink while women do not smoke or drink in this area.

Data from the present study revealed that HP infection was a risk factor for ESCC in the high risk population. However, results on HP infection and risk of ESCC have been inconsistent to date, an inverse relationship between HP infection and esophageal adenocarcinoma risk has been reported [47, 48]. Furthermore, HP infection was protective against developing ESCC in a Taiwanese population [49]. Possible reasons for this discrepancy are not clear, and further investigation is needed.

Our study indicated that green tea is an important protective factor for ESCC, which is consistent with previous studies conducted in the same area [21] and in Shanghai, China [50]. Recent evidence suggests that the anti-oxidative and anti-inflammatory properties of green tea make it a promising agent for human cancer preventions [51]. In rats, green tea consumption prevents the esophageal carcinogenesis induced by N-nitrosomethylbenzylamine (NMBA), which is a potent carcinogen that is frequently found in moldy food and pickled vegetables [52]. A cohort study also showed an inverse relationship between urinary green tea polyphenols and risk of ESCC of men [53]. One possible mechanism may be that the green tea polyphenols induce epidermal growth factor receptor inhibition in esophageal cancer cells [54]. Taken together, these findings suggest that green tea might be used as a chemoprevention agent for human esophageal cancer.

Numerous studies have investigated the association between different XME polymorphisms and ESCC, but with conflicting results [18]. In the present study, a difference was found in the distribution of GSTT1 null genotypes between cases and controls; however, the difference only approached statistical significance. This difference was more striking in the male subjects, resulting in a 2.8-fold increase in risk of ESCC compared to male subjects who carried the GSTT1 non-null genotype. This suggests that the GSTT1 null genotype may contribute to the development of ESCC in this high risk population. Although this was different from previous studies on ESCC [22, 23, 55], it is biologically plausible because the presence of the GSTT1 null genotype leads to lower GST enzyme activity levels and, consequently, impaired detoxification of environmental or dietary carcinogens. In addition, GSTT1 null has been shown to play a role in esophageal carcinogenesis through a pathway of abnormalities in the p53 tumor suppressor gene, which is highly reported in ESCC [18]. Previous data showed that CYP1A1 Val/Val and CYP2E1 c1/c1 genotypes were related with increased risk of ESCC [22, 56]. EPHX was also reported to be associated with high susceptibility of ESCC [23]. Nevertheless, we did not find significant association between CYP1A1, CYP2E1, CYP2A6, and EPHX and ESCC. The exact reasons for these inconsistent findings remain to be elucidated, but it is demonstrated that the variation in enzyme activity with ethnicity and gender may contribute to the differences in effects on the risk of cancers [57].

In order to find an overall risk for ESCC, variables with statistical significance in univariate conditional logistic regression were analyzed in a multivariable model. Esophageal lesion, eating fast, a family history of cancer, and HP infection were risk factors for ESCC (P < 0.05), while clean up of storage utensils and green tea consumption were protective factors for ESCC (P < 0.05). Personal character, fatty meat intake, pickled vegetables intake, and alcohol abstinence were not statistically significant in the multivariable model which might be due to their colinearity with other variables in the multivariable analysis. For instance, eating fast and pickled vegetables intake was highly correlated (P < 0.05, data not shown). Although, gene-gene and gene-environment interactions have been reported for ESCC [22, 58], we did not find such interactions in our study, which might be related to the smaller sample size in our study.

The present study indicated that fatty meat consumption, eating fast, consuming pickled vegetables, moldy food intake, preference for salty and acrid food along with esophageal lesion, HP infection, passive smoking, and a family history of cancer are risk factors for ESCC in Huaian area, China. Regular cleaning of food storage utensils, green tea consumption, and alcohol abstinence are important protective factors for ESCC in this high risk population. In addition, there was a significant association between the GSTT1 null genotype and risk of ESCC in male subjects. One potential limitation of the current study is the relatively smaller sample size, which may reduce statistical power for detecting gene-gene and gene-environmental differences. Hence, further studies with larger sample size and prospective cohort are mandated to confirm the present findings.