Green tea consumption, inflammation and the risk of primary hepatocellular carcinoma in a Chinese population

Abstract

Objective

Green tea has been found to possess anti-inflammatory, anti-oxidative and anti-carcinogenic properties. The present study examines the association between green tea drinking and hepatocellular carcinoma (HCC) and its interactions with other risk or protective factors and single nucleotide polymorphisms (SNP) of inflammation and oxidative stress related genes.

Methods

A population-based case-control study with 204 primary HCC cases and 415 healthy controls was conducted in Taixing, China. Epidemiological data were collected using a standard questionnaire. SNPs of genes of the inflammation and metabolic pathways were genotyped at the UCLA Molecular Epidemiology Laboratory. Logistic regression was performed to estimate adjusted odds ratios and 95% confidence intervals.

Results

Longer duration and larger quantities of green tea consumption were inversely associated with primary HCC. Individuals who drank green tea longer than 30 years were at lowest risk (adjusted OR = 0.44, 95% CI: 0.19-0.96) compared with non-drinkers. A strong interaction was observed between green tea drinking and alcohol consumption (adjusted OR for interaction = 3.40, 95% CI: 1.26-9.16). Green tea drinking was also observed to have a potential effect modification on HBV/HCV infection, smoking and polymorphisms of inflammation related cytokines, especially for IL-10.

Conclusion

Green tea consumption may protect against development of primary HCC. Potential effect modifications of green tea on associations between primary HCC and alcohol drinking, HBV/HCV infection, and inflammation-related SNPs were suggested.

1. Introduction

Primary liver cancer (PLC) is the sixth most common cancer and the third most common cancer-related death worldwide. PLC has a wide geographic variation, with approximately 82% of PLC cases and deaths occurring in the developing countries [1]. According to Globocan 2002, an estimated 345,844 new PLC cases and 321,851 deaths occurred in China each year, accounting for more than half of PLC cases and deaths in the world. Hepatocellular carcinoma (HCC) comprises 85-90% of PLC, and the two terms are often used interchangeably [2].

Chronic infections by hepatitis B (HBV) and hepatitis C (HCV), aflatoxin B1 (AFB1) and excessive alcohol consumption have been associated with increased risk for HCC [3-7]. Tobacco smoking may increase HCC risk, but the results are conflicting [8,9]. Among these known or suspected risk factors, inflammation may play an important role in infectious and non-infectious pathways leading to HCC. Chronic infections with HBV/HCV can induce chronic inflammation in liver tissue, which creates a procarcinogenic environment for HCC development. Chronic alcohol consumption and tobacco smoking are known non-infectious causes of chronic inflammation and may play a role in liver cell injury [10,11]. Chronic inflammation is characterized, in part, by altered cytokine levels, which may contribute to the pathogenesis of HCC [12]. In addition, liver tissue inflammation can increase oxidative stress, which is thought to be important in the initiation and promotion of HCC [13,14]. Single nucleotide polymorphisms (SNPs) of cytokines may alter gene expression level and are functionally associated with liver disease and HCC [12,15].

Green tea contains many polyphenols. In vitro and in vivo studies have suggested that polyphenols, particularly epigallo-cathechin-3 gallate (EGCG), have anti-inflammatory and antioxidant properties [16-19]. Epidemiological studies suggest that, though not conclusive, green tea consumption is associated with reduced risk of gastrointestinal cancers, including stomach cancer [20-22], esophageal cancer [23-25] and colorectal cancer[26,27]. Increased green tea consumption is inversely associated with breast cancer in case-control studies [28,29], but not in cohort studies [30,31]. In addition, green tea drinking is associated with breast cancer recurrence [32,33]. Very few studies have been conducted to examine the effect of green tea drinking on the development of HCC, and the published results are inconclusive [34-38]. There have been no published epidemiological studies that investigate interactions between polymorphisms of inflammation-related cytokines and green tea consumption on HCC risk. We hypothesize that the antiinflammatory effects of green tea may protect against HCC and SNPs of inflammation pathways may modify the effect of green tea drinking on HCC.

2. Materials and methods

2.1. Study population

The current study includes 204 newly diagnosed HCC cases and 415 controls from a population-based case-control study conducted in Taixing city, China. The original study population has been described in detail previously [20,39]. Briefly, the original study included three cancer sites (esophagus, stomach and liver) and one common control group. The ratio of the combined three cancer cases to the common control is 3:2. The current study only includes newly diagnosed HCC patients and all population controls. Therefore, the HCC case-to-control ratio is 1:2 in this study. Eligible cases were identified using the Taixing Tumor Registry between January 1 and June 30, 2000. Healthy controls were randomly selected from the residential areas where cancer cases originated and were frequency-matched to the combined case group (esophagus, stomach and liver) on residential area, gender and age. Cases and controls were at least 20 years old and living in Taixing at least 10 years. If a selected control did not fit the criteria, or refused to be interviewed, we recorded their basic demographic data and used the same selection process to choose another control. Recruitment rates were 57% for eligible cases and 89% for controls. Since controls were not individual-matched to cases, but were instead frequency-matched to the combined case group, the age and gender distribution of controls corresponds to all three cancer sites and does not exactly match the HCC cases.

2.2. Data collection

All cases and controls completed a standardized interviewer-based questionnaire to collect information on demographic factors, smoking history, alcohol drinking habits, tea drinking habits, drinking water, personal, family cancer history, etc. An 8-milliliter blood sample was collected from 194 cases (95%) and 397 controls (96%). Genomic DNA was isolated from the specimens using a modified phenol-chloroform protocol.

2.3. Laboratory assays

2.3.1. PCR-analysis of gene polymorphisms

Genotyping was performed in the Molecular Epidemiology Laboratory at Department of Epidemiology, School of Public Health at UCLA. Genotypes for GSTM1, GSTT1 and ALDH2 rs671 polymorphisms were determined using the PCR-RFLP method as previously described [20,40]. The other SNPs (ALDH2 rs886205, IL-1α rs17561, IL-10 rs1800871, IL-10 rs 1800872, IL-10 rs 1800896, IL-13 rs20541, TNF-α rs1800629, and TNF-β rs909253) were genotyped using the SNPlex assay (Applied Biosystems [ABI], Foster City, CA) with call rates of >85% and reproducibility of 0.978 (randomly regenotyping 3% of samples). Two SNPs (ALDH2 rs886205 and IL-10 rs1800871) were also genotyped using the ABI’s Taqman assay with call rates of >97% and reproducibility of 0.989 (randomly regenotyping 5% of samples). The concordance between SNPlex and Taqman is 94.3% for IL-10 rs1800871 and 98.8% for ALDH2 rs886205. Detailed description of SNPlex and Taqman methods were published elsewhere [41-43].

2.3.2. HBsAg, anti-HCV and plasma aflatoxin B1-albumin adduct detection

The presence of HBsAg in serum was measured by enzyme-linked immunosorbant assay (ELISA) using kits from the Reagent Company of the Shanghai Hospital for Infectious Diseases. Anti-HCV IgG antibody was measured by ELISA using kits from Shanghai Huamei Biological Company. Aflatoxin B1 (AFB1)-albumin adduct levels were measured from plasma as previously described [44].

2.4. Definition of green tea and alcohol drinking

Green tea ever-drinkers were defined as individuals who drank at least one cup of green tea per day for more than half a year. Tea concentration was categorized into three levels: low (tea leaves were <25% of the volume of the cup), moderate (tea leaves were between 25% and 50% of the volume of the cup) and high (volume of tea leaves was >50% of cup volume). One standard drink of alcohol was defined as any alcoholic beverage containing 14 g of pure alcohol (National Institute on Alcohol Abuse and Alcoholism, the United States).

2.5. Statistical analysis

All analyses were performed using SAS 9.2 software. We used unconditional logistic regression models to obtain odds ratios (ORs) and 95% confidence intervals (CIs). We adjusted each independent variable for potential confounders, including age (continuous), gender (male or female), education (ordinal, five categories), income (continuous), body mass index (BMI, continuous), family history of primary liver cancer (yes or no), pack-years of smoking (continuous), alcohol drinking (never/occasionally vs. often/everyday) and HBSAg status (positive vs. negative). Associations were considered statistically significant if the P-value <0.05 at two-sided test. Unconditional logistic regression was used to evaluate multiplicative interactions between green tea and other risk factors. Departures from multiplicative effects were assessed by including main effect variables and their product terms in the logistic regression model.

3. Results

The distributions of demographic characters among cases and controls are shown in Table 1. We observed a higher proportion of males and younger individuals among cases, which is consistent with characteristics of HCC in high-risk areas. Education levels and BMI were higher among controls (P < 0.05). There is no obvious difference in the income class between cases and controls (P = 0.063).

Table 1.

Distribution of demographic characters among cases and controls.

	Cases		Controls		p-Value
	N	%	N	%
Gender
Male	159	77.9	287	69.2	0.0221
Female	45	22.1	128	30.8
Age
<40	31	15.2	31	7.5	0.0003
40-50	54	26.5	69	16.6
50-60	54	26.5	136	32.8
60-70	42	20.6	116	28.0
≥70	23	11.3	63	15.2
Education
Illiteracy	44	21.6	73	17.6	0.0018
Primary school	77	37.8	142	34.2
Middle school	70	34.3	124	29.9
High school	13	6.4	66	15.9
College and above	0	0.0	10	2.4
Monthly income per capita (Yuan)
<60	63	30.9	88	21.2	0.0643
60-100	35	17.2	74	17.8
100-160	58	28.4	135	32.5
≥160	48	23.5	118	28.4
BMI
≤22	106	52.0	180	43.4	0.0440
>22	98	48.0	235	56.6

Compared with green tea non-drinkers, subjects who drank more than 250 g of green tea per month (about ≥2 cups had per day), a crude OR of 0.58 (95% CI: 0.34–1.00) and an adjusted OR of 0.55 (95% CI: 0.28–1.09), while subjects who drank green tea longer than 30 years had a crude OR of 0.38 (95% CI: 0.20–0.74) and an adjusted OR of 0.44 (95% CI: 0.19-0.96). No obvious association was observed between green tea concentration, green tea drinking age onset, green tea temperature and HCC. After adjusting for potential confounding factors, no obvious associations were observed between alcohol drinking, tobacco smoking and HCC risk. Chronic HBV and HCV infection markers, HBsAg and anti-HCV, were much more prevalent among cases than controls, with adjusted ORs of 5.07 (95% CI: 3.38–7.60) and 4.44 (95% CI: 1.82–10.85), respectively. Ingestion of moldy foods was moderately associated with HCC, with an adjusted OR of 2.39 (95% CI: 1.44–3.98). Using a refrigerator was protective, with an adjusted OR of 0.36 (95% CI: 0.19-0.69). Raw water drinking was associated with HCC, while tap water drinking was inversely related to HCC (Ptrend < 0.05). A family history of HCC was associated with HCC, with an adjusted OR of 3.06 (95% CI: 1.80-5.19) (Table 2).

Table 2.

Tea drinking and potential risk factors among cases and controls.

	Cases N	Controls N	cOR and 95% CI	aOR and 95% Cla
Years of green tea drinking
Never	111	216	Ref	Ref
<20	32	60	1.04 (0.64–1.69)	0.69 (0.36–1.32)
20–30	27	60	0.88 (0.53–1.46)	1.05 (0.55–1.99)
≥30	12	61	0.38 (0.20–0.74)	0.44 (0.19–0.96)
P trend			0.0131	0.1297
Monthly consumption of green tea (g/month)
Never	111	216	Ref	Ref
<125	23	42	1.07 (0.61–1.86)	1.21 (0.62–2.36)
125–250	24	50	0.93 (0.55–1.60)	0.76 (0.38–1.51)
≥250	21	70	0.58 (0.34–1.00)	0.55 (0.28–1.09)
P trend			0.0849	0.0806
Green tea concentration
Never	111	216	Ref	Ref
Low	10	20	0.97 (0.44–2.15)	0.84 (0.33–2.14)
Moderate	42	125	0.65 (0.43–0.99)	0.60 (0.34–1.05)
High	21	36	1.14 (0.63–2.04)	1.12 (0.54–2.31)
P trend			0.3190	0.4685
Green tea temperature
Never	111	216	Ref	Ref
Normal temperature	50	125	0.78 (0.52–1.16)	0.76 (0.45–1.31)
High temperature	14	47	0.58 (0.31–1.10)	0.50 (0.23–1.07)
P trend			0.0571	0.0688
Age (years) of green tea drinking onset
Never	111	216	Ref	Ref
>28	29	89	0.63 (0.39–1.02)	0.56 (0.30–1.02)
≤28	41	92	0.87 (0.56–1.34)	0.89 (0.49–1.60)
P trend			0.3159	0.5411
Alcohol drinking
Never/occasionally	116	279	Ref	Ref
Often/everyday	76	133	1.37 (0.96–1.96)	1.29 (0.79–2.09)
Drinks per day (1960–1999)
Never	123	248	Ref	Ref
0–2	29	66	0.89 (0.54–1.44)	0.90 (0.49–1.65)
>2	52	101	1.04 (0.70–1.55)	1.30 (0.75–2.25)
P trend			0.9322	0.3719
Pack–year of smoking
Never	85	217	Ref	Ref
<20	53	85	1.59 (1.04–2.44)	1.12 (0.62–2.01)
20–40	42	86	1.25 (0.80–1.95)	0.88 (0.48–1.64)
≥40	12	26	1.18 (0.57–2.44)	0.92 (0.35–2.37)
P trend			0.2768	0.6904
HBV infection
HBsAg−	72	312	Ref	Ref
HBsAg+	132	102	5.61 (3.90–8.07)	5.07 (3.38–7.60)
HCV infection
Anti-HCV−	183	403	Ref	Ref
Anti-HCV+	18	12	3.03 (1.56–7.00)	4.44 (1.82–10.85)
Moldy food intake
No	143	339	Ref	Ref
Yes	48	66	1.72 (1.13–2.62)	2.39 (1.44–3.98)
Using refrigerator
No	173	305	Ref	Ref
Yes	18	86	0.37 (0.22–0.63)	0.36 (0.19–0.69)
Raw water drinking
Never	95	248	Ref	Ref
Few/ever	27	44	1.60 (0.94–2.73)	1.84 (0.96–3.53)
Sometimes	52	62	2.19 (1.41–3.39)	2.18 (1.30–3.67)
Often	11	5	5.74 (1.94–16.97)	3.78 (0.96–14.96)
P trend			<0001	0.0006
Tap water drinking
Never use tap water	151	263	Ref	Ref
Use tap water <3yrs	23	67	0.60 (0.36–1.00)	0.56 (0.30–1.02)
Use tap water >3yrs	30	85	0.61 (0.39–0.98)	0.63 (0.36–1.10)
P trend			0.0154	0.0416
Family history of liver cancer
No	150	375	Ref	Ref
Yes	53	39	3.40 (2.16–5.35)	3.06 (1.80–5.19)

^aAdjusted for age, gender, education, income, BMI, family history, pack-year, alcohol drinking and HBSAg.

Table 3 shows the possible interactions between environmental risk factors and green tea drinking. A more than multiplicative interaction between green tea and alcohol drinking was observed (adjusted OR for interaction 3.40, 95% CI: 1.26–9.16). The adjusted ORs for interactions between green tea drinking and smoking, and green tea drinking and HBV or HCV infection were 1.98 (95% CI: 0.78-5.03) and 1.87 (95% CI: 0.81-4.31), respectively. Green tea drinking does not modify the effect of AFB1 on HCC.

Table 3.

Interactions between green tea and other risk factors.

		Cases N	Controls N	cOR and 95% CI	aOR and 95% CIa
Green tea drinking	Alcohol drks/day
Ever	<2	46	105	Ref	Ref
Ever	≥2	27	76	0.81 (0.46–1.42)	0.85 (0.44–1.64)
Never	≤2	86	191	1.03 (0.67–1.58)	0.93 (0.52–1.66)
Never	>2	25	25	2.28 (1.19–4.39)	2.68 (1.24–5.82)
OR for interaction				2.74 (1.20–6.27)	3.40 (1.26–9.16)
Green tea drinking	Smoking
Ever	Never	22	60	Ref	Ref
Ever	Ever	51	121	1.15 (0.64–2.07)	0.76 (0.38–1.55)
Never	Never	61	153	1.09 (0.61–1.93)	0.89 (0.41–1.90)
Never	Ever	50	63	2.17 (1.17–4.00)	1.34 (0.63–2.85)
OR for interaction				1.73 (0.81–3.69)	1.98 (0.78–5.03)
Green tea drinking	HBV/HCV infection
Ever	No	26	126	Ref	Ref
Ever	Yes	45	55	3.97 (2.23–7.06)	3.88 (2.10–7.19)
Never	No	31	163	0.92 (0.52–1.63)	1.01 (0.52–1.94)
Never	Yes	79	52	7.36 (4.26–12.74)	7.29 (3.84–13.82)
OR for interaction				2.02 (0.93–4.38)	1.87 (0.81–4.31)
Green tea drinking	AFB1
Ever	Low	35	117	Ref	Ref
Ever	High	32	44	2.43 (1.35–4.39)	2.17 (1.14–4.15)
Never	Low	57	127	1.50 (0.92–2.45)	1.55 (0.86–2.81)
Never	High	40	73	1.83 (1.07–3.14)	1.90 (1.00–3.61)
OR for interaction				0.50 (0.23–1.09)	0.56 (0.24–1.32)

^aAdjusted for age, gender, education, income, BMI, family history, pack-year, alcohol drinking and HBSAg.

Table 4 shows the possible interactions between gene polymorphisms and green tea drinking. More than multiplicative interactions were observed between green tea drinking and polymorphisms of IL-10/-819 (rs 1800871) and IL-10/-592 (rs 1800872). Among green tea non-drinkers, an increased risk for HCC was associated with the presence of C/C genotype in IL-10/-819 and IL-10/-592, while for green tea drinkers, a decreased risk for HCC was found with the presence of C/C genotype in IL-10/-819 and IL-10/-592. When evaluating the interactions for green tea and gene polymorphisms, we assumed green tea drinkers with major alleles to be at lowest risk and usually used this category as reference. For IL-10/-819 and IL-10/-592, we used the green tea drinkers with any variant alleles as reference instead. IL-10 is an anti-inflammatory cytokine and its major alleles may be associated with HCC risk [45-47]. Therefore, we assume green tea drinkers with any variant alleles of IL-10 might be at lower risk. For IL-10/-1082 (rs 1800896), green tea drinkers with homozygous variant alleles (A/A) were used as reference, due to the very few subjects with homozygous major alleles (G/G). No obvious interactions were found between polymorphisms of GSTM1, GSTT1, IL-1α (rs17561), IL-13 (rs20541), TNF-α (rs1800629), TNF-β (rs909253) and green tea consumption on HCC risk (data are not shown).

Table 4.

Interactions between green tea and gene polymorphism.

		Cases N	Controls N	cOR &95%CI	aOR &95%CIa
Green tea drinking	ALDH2 rs886205
Ever	C/C	49	135	Ref	Ref
Ever	C/T or T/T	17	35	1.34 (0.69–2.60)	1.10 (0.50–2.40)
Never	C/C	84	169	1.37 (0.90–2.08)	1.39 (0.79–2.43)
Never	C/T or T/T	22	36	1.68 (0.90–3.14)	2.28 (1.03–5.03)
OR for interaction				0.92 (0.38–2.24)	1.50 (0.52–4.37)
Green tea drinking	IL-10 rs1800871
Ever	T/T or C/T	62	140	Ref	Ref
Ever	C/C	3	24	0.28 (0.08–0.97)	0.21 (0.05–0.83)
Never	T/T or C/T	82	183	1.01 (0.68–1.50)	1.17 (0.68–2.00)
Never	C/C	19	15	2.86 (1.37–5.99)	3.17 (1.29–7.79)
OR for interaction				10.01 (2.39–41.97)	13.14 (2.60–66.52)
Green tea drinking	IL-10 rs1800872
Ever	A/A or A/C	57	136	Ref	Ref
Ever	C/C	2	21	0.23 (0.05–1.00)	0.18 (0.04–0.95)
Never	A/A or A/C	77	177	1.04 (0.69–1.56)	1.18 (0.68–2.043)
Never	C/C	14	13	2.57 (1.14–5.81)	2.81 (1.08–7.28)
OR for interaction				10.89 (2.02–58.77)	13.13 (1.98–86.94)
Green tea drinking	IL-10 rs1800896
Ever	A/A	53	125	Ref	Ref
Ever	A/G or G/G	7	40	0.41 (0.17–0.98)	0.35 (0.13–0.92)
Never	A/A	79	153	1.22 (0.80–1.86)	1.40 (0.80–2.48)
Never	A/G or G/G	19	37	1.21 (0.64–2.30)	1.20 (0.54–2.67)
OR for interaction				2.41 (0.83–6.97)	2.46 (0.72–8.43)

^aAdjusted for age, gender, education, income, BMI, family history, pack-year, alcohol drinking and HBsAg.

Fig. 1 shows the study participants stratified into three groups based on their green tea drinking status: non-drinkers, low-level drinkers (≥250 g/month) and high-level drinkers (<250 g/month). The stratified analyses showed the greatest protection from green tea drinking (>250 mg) on individuals who drank alcohol and with HBV or HCV infection.

Fig. 1.

Alcohol drinking, smoking and HBV&HCV infections among different green tea drinkers. (a) Alcohol drinking, green tea drinking and HCC. Study subjects were stratified into three groups based on their green tea drinking amounts: never drinking, less than 250 g/month and more than 250 g/month. In each green tea drinking categories, the adjusted ORs of alcohol drinking were presented, using subjects who never or occasionally drank alcohol as reference. (b) HBV/HCV infection, green tea drinking and HCC. Study subjects were stratified into three groups based on their green tea drinking amounts: never drinking, less than 250 g/ month and more than 250 g/month. In each green tea drinking categories, the adjusted ORs of HBV/HCV infection were presented, using subjects who have no HBV and HCV infection as reference.

4. Discussion

Of the three major types of tea, green tea (non-fermented), oolong tea (half-fermented) and black tea (fermented), green tea has the highest quantity of tea polyphenols [48]. Green tea has been used as herbal medicine and healthy beverage in China since ancient times. Green tea drinking is thought to provide protection against tumor initiation and development in multiple organs [48-50]. The present study found a protective role of green tea drinking against HCC. In the crude analysis, dose–response relationships were suggested, and longer years and increased consumption of green tea were associated with decreased odds of developing HCC. However, the associations were no longer apparent after adjusting for potential confounding factors.

The association between green tea drinking and HCC are consistent with published studies in China [51]. A prospective cohort study found an inverse association between tea drinking and HCC mortality rates in males [34]. One phase II chemoprevention trial suggested that green tea polyphenols are effective in diminishing oxidative DNA damage in individuals at high-risk of HCC [52]. Conflicting results were observed in populations other than the Chinese population. In Japan, one prospective cohort study found green tea was inversely associated with liver cancer incidence, while no association was observed in another study[37,38]. A case-control study in Italy reported no association between tea and HCC [35].

The conflicting results may arise for the following reasons. First, most previous analyses were simply based on frequency of tea consumption, with no detailed information on type of tea, consumption amount and duration of tea drinking. Other limitations include the study heterogeneity due to inclusion of a variety of liver diseases as study endpoints, and not adjusting for the effects of viral hepatitis infections in some studies. In the present study, green tea drinking habits were assessed in more detail. Our analyses were also limited to green tea drinking and HCC cases, which may preclude the possible influence from different concentrations of polyphenols in various types of tea as well as the heterogeneous causes of different liver diseases.

This study found a possible interaction between green tea drinking and infections of HBV and/or HCV (HBV/HCV). Among individuals infected with HBV/HCV, the risk of HCC in green tea non-drinkers is almost twice that of drinkers. The results implicated that green tea consumption might have a potential HCC prevention benefits among individuals with chronic hepatitis infection. Intervention trials are needed in the future to provide more convincing evidence. Persistent infections with HBV/HCV induce chronic inflammation which may also be associated with oxidative stress. Both chronic inflammation and oxidative stress may contribute to the development of HCC [53]. The anti-inflammatory and anti-oxidative properties of tea polyphenols may explain the positive interactions between green tea and HBV/ HCV infections.

Results from the present study also suggest possible interactions between green tea drinking and SNPs of inflammation related cytokines. Cytokines are generally grouped into two categories, pro-inflammatory cytokines (e.g., IL-1α, IL-1β, IL-2, TNF-α and IFN-γ) and anti-inflammatory cytokines (e.g., IL-4, IL-5, IL-8, IL-10 and IL-13) [12]. Some studies have shown that high IL-10-producing alleles (major allele of IL-10/-819 and IL-10/-592) were associated with increased HCC risk [45-47]. In the current study, IL-10 major alleles were associated with increased risk among green tea non-drinkers but decreased risk among green tea drinkers. Green tea might reverse the harmful effect of the high IL-10-producing alleles. The relatively small number of subjects in certain categories may limit our ability to detect the true association. Large-scaled studies are needed to further explore the interactions between green tea and SNPs of inflammation related cytokines.

Alcohol consumption is known to cause HCC in developed countries [4]. In the present study, no obvious association was observed between alcohol drinking and HCC without considering tea drinking status. However we observed a strong interaction between green tea and alcohol drinking, and the highest odds of HCC was found among individuals who drank alcohol without concurrent green tea drinking. Possible explanations for this interaction may be due to the anti-oxidative effects of tea polyphenols. Ethanol is eliminated from the body by its oxidation. Acetaldehyde, the first oxidation product of ethanol, is thought to be mainly responsible for the carcinogenic effect of alcohol [54]. Previous evidence showed that green tea may protect against alcohol-related oxidative modification [55-57]. Aldehyde dehydrogenase (ALDH) is the key enzyme in eliminating acetaldehyde, and the null-allele variant of ALDH2 may lead to accumulation of acetaldehyde [58]. Our study found that green tea non-drinkers with variant ALDH2 (rs886205) alleles were at higher risk than other groups. However, no significant interaction between ALDH2 and green tea drinking was found. This result is consistent with one animal experiment [59].

Several methodological issues need to be discussed. Selection bias may occur due to the relatively low participation rate (57%) of HCC cases. High fatality rate of HCC led to a great proportion of newly diagnosed cases dying before our interviewers could reach them. Thus it is possible the enrolled cases only represent less severe patients among all the qualified cases. In retrospective case-control studies, information bias, especially recall bias and corresponding differential misclassification, are important issues. Since the association between green tea and liver cancer had not been established yet, participants in this study were not aware of the possible benefits from green tea drinking. Besides that, several green tea drinking related variables were collected. Thus, the possibility of differential recall bias would be minimal. There is a possibility of reverse causality if cases of liver cancer started to drink green tea after diagnosis. However, this will lead to an underestimation of the observed protective association. Standard protocols and blinding were utilized in the assays so as to reduce other information bias and possible misclassification. Since we used common controls for all three cancer sites, potential residual confounding effects of age and gender might exist, even though these two factors were adjusted for in the multivariate analysis. The relatively small sample size also limited our ability to detect moderate interactions and compromised the precision of measurements. However, the current report has provided preliminary evidence on the association of green tea drinking with HCC and its potential modifying effect on other factors. Large-scale epidemiological studies are needed in the future to further study the complex role of green tea drinking in the development of HCC.

In conclusion, our results support the hypothesis that long years and high amount of green tea drinking may be a protective factor against primary HCC. Green tea drinking modified the effect of alcohol drinking on the development of primary HCC. Potential effect modifications of green tea drinking on associations between primary HCC and hepatitis virus infections and inflammation related SNPs, especially for IL-10, were suggested.

Abbreviations

PLC

primary liver cancer

HCC

hepatocellular carcinoma

HBV

hepatitis B

HCV

hepatitis C

HBsAg

hepatitis B virus surface antigen

SNPs

single nucleotide polymorphisms

AFB1

aflatoxin B1

OR

odds ratio

95% CI

95% confidence interval

ALDH

aldehyde dehydrogenase