Effect of Coffee and Green Tea Consumption on the Risk of Liver Cancer: Cohort Analysis by Hepatitis Virus Infection Status

Coffee and green tea, the most popular beverages in the world, contain polyphenolic antioxidants, which are thought to contribute to cancer prevention (1). Although an association between these beverages and liver cancer has been speculated (2, 3), epidemiologic evidence is insufficient (4) and varies by beverage. Several cohort and case-control studies (5-15) and meta-analyses (16, 17) have reported a significant inverse association between coffee consumption and liver cancer, but none of these considered hepatitis virus infection status. Meanwhile, epidemiologic evidence for the association between green tea consumption and liver cancer is sparse (10, 14, 18), and little is known about its effects on liver cancer.

For both beverages, few studies have taken account of hepatitis virus infection status in evaluating risk (8, 10, 11, 13-15), although hepatitis B (HBV) and C (HCV) virus infections are the most important risk factors of liver cancer (19). This problem is particularly prominent in cohort studies having no baseline data on hepatitis virus infection. Indeed, because we had insufficient data, our previous analysis of coffee consumption and liver cancer (5) also failed to take account of hepatitis virus infection status at baseline.

To investigate the association between these putatively anticarcinogenic beverages and liver cancer in consideration of HCV and HBV infection status, we conducted a cohort analysis of the association between the consumption of coffee and green tea and the risk of liver cancer using a large-scale population-based study in Japan, which has a relatively high incidence of liver cancer (20). Our main purpose was to clarify whether coffee and green tea consumption reduce the risk of liver cancer in those with HCV or HBV infection (or both), who are at high risk of liver cancer, using a prospective study design. Infection status was determined at baseline using stored blood samples.

The Japan Public Health Center-Based Prospective Study Cohort II was initiated in 1993 to 1994. Subjects were drawn from six public health center areas across Japan. The study design has been described in detail previously (21). The study population was defined as all residents ages 40 to 69 years at the start of the baseline survey. A part of one public health center area was excluded because its study population was defined differently to the others. Initially, we defined a population-based cohort of 68,975 subjects after exclusion of ineligible subjects (n = 103). The study was approved by the Institutional Ethical Review Board of the National Cancer Center.

A self-administered questionnaire survey on various lifestyle factors was conducted at baseline (response rate, 82%). Among respondents, 39% voluntarily provided 10 mL blood during health checkups provided by their local government. The plasma and buffy layer were divided into four tubes holding 1.0 mL each (three tubes for plasma and one for the buffy layer) and stored at −80°C until analysis.

For the present analysis, we restricted subjects to those who responded to the questionnaire and for whom blood samples were available. We further excluded those with a history of liver cancer and those with missing data on variables included in the statistical model, including coffee and green tea consumption, smoking status, weekly ethanol intake, body mass index, history of diabetes mellitus, serum alanine aminotransferase (ALT) level, and HCV and HBV infection status. Finally, a total of 18,815 individuals were included in the present analysis.

All exposures were collected at baseline. Information on coffee and green tea consumption was obtained in terms of the frequency and amount of each beverage consumed using the pre-coded answers of almost never, 1-2 days/wk, 3-4 days/wk, and almost daily (further divided into 1-2, 3-4, ≥5 cups/d). For the present analysis, we further grouped these categories based on their distribution among the subjects (coffee consumption: almost never, <1, 1-2, ≥3 cups/d and green tea consumption: <3, 3-4, ≥5 cups/d). The validity of coffee and green tea consumption reported by the cohort was assessed using dietary records for 28 days (7-day dietary records in four seasons) or 14 days. The rank correlation coefficients for coffee consumption between the questionnaire and dietary record data were 0.59 for men and 0.51 for women, whereas that for green tea consumption was 0.37 for men and 0.43 for women.

Plasma samples were screened for anti-HCV antibody using a third-generation immunoassay (Lumipulse II Ortho HCV; Ortho-Clinical Diagnosis; ref. 22) and for HBV antigen (HBsAg) by reversed passive hemagglutination with a commercial kit (Institute of Immunology). This HBsAg detection kit is one of the most commonly used in Japan. The sensitivity of this kit (5 IU/mL) is similar to that of the radioimmunoassay method and lower than that of the chemiluminescent immunoassay method (23) and is considered suitable for the screening of general Japanese populations, particularly given the rarity of HBsAg-positive cases with low titers. In the present study, positivity for anti-HCV was regarded as indicating HCV infection and positivity for HBsAg as indicating HBV infection.

Subjects were followed from the date of the baseline survey until December 31, 2006. Residence status, including survival, was confirmed through the residential registry. Inspection of the registry is available to anyone under the resident registration law. Information on the cause of death was obtained from the death certificate, provided by the Ministry of Health, Labour, and Welfare with the permission of the Ministry of Internal Affairs and Communications, in which cause of death is defined according to the International Classification of Disease, 10th Edition (24). Resident and death registration are required by law in Japan and the registries are believed to be complete. Among study subjects, 1,894 died, 1,088 moved out of the study area, and 49 (0.3%) were lost to follow-up during the follow-up period.

Incidence data on liver cancer were obtained by active patient notification from major hospitals in the study area and data linkage with population-based cancer registries, with permission from the local governments responsible for the registries. Death certificates were used as a supplementary information source. In our cancer registry system, the proportion of cases for which information was available from death certificates only was 4.7%. The site and histology of each liver cancer case were coded using the International Classification of Diseases for Oncology, Third Edition (C22.0; ref. 25). For the present analysis, the earliest date of diagnosis was used in subjects with multiple primary cancers at different times. A total of 110 newly diagnosed cancer cases (73 in men and 37 in women) were identified.

The number of person-years in the follow-up period was counted from the date of completion of the baseline questionnaire until the date of liver cancer diagnosis, date of emigration from the study area, date of death, or end of the study period, whichever came first. For subjects who withdrew from or were lost to follow-up, the date of withdrawal and the last confirmed date of presence, respectively, were used as the date of censor.

The relative risk of liver cancer associated with coffee and green tea consumption was described using hazard ratios (HR) and 95% confidence intervals (95% CI). Analyses were conducted among total subjects (n = 18,815), those who were either or both anti-HCV and HBsAg positive (n = 1,499), and those who were anti-HCV positive without regard to HBV status (n = 1,058). The Cox proportional hazards model was employed to control for potential confounding factors: sex (stratified; men and women combined only), age at baseline (stratified; 5-year age categories), area (stratified; six public health center areas), smoking status (never, past, current), weekly ethanol intake (past, less than weekly, <150 g/wk, ≥150 g/wk), body mass index (<25, 25 to <27, ≥27), history of diabetes mellitus (yes, no), serum ALT level (IU/L; continuous), and HCV (anti-HCV negative, positive; not included in analyses restricted to HCV-positive subjects) and HBV (HBsAg negative, positive) infection status in addition to coffee consumption (almost never, <1, 1-2, ≥3 cups/d) and green tea consumption (<3, 3-4, ≥5 cups/d). These variables, obtained from the questionnaire, are either known or suspected from previous studies as risk factors for liver cancer. Trend was assessed by assigning ordinal values for original coffee and green tea consumption categories. Sex, age, and area were treated as strata to allow for a different baseline hazard for each stratum. Testing of the proportional hazards assumption by Schoenfeld and scaled Schoenfeld residuals found no violation of proportionality. All statistical analyses were done using Stata 10 (Stata; ref. 26).

During 238,517 person-years of follow-up (average follow-up period, 12.7 years) for 18,815 subjects (6,414 men and 12,401 women), a total of 110 cases of newly diagnosed liver cancer (73 men and 37 women) were identified and included in the analyses.

Baseline characteristics of subjects according to coffee and green tea consumption are shown in Table 1 (coffee) and Table 2 (green tea). Among total subjects, 33% almost never drank coffee, whereas 9% consumed ≥3 cups/d. In contrast, 63% of subjects consumed at least 3 cups/d of green tea. The distribution of coffee and green tea consumption was similar between those with and without HCV and HBV infection (data not shown).

Subjects with higher coffee consumption tended to smoke more, consume less green tea, be less obese, and have less history of diabetes mellitus. These tendencies were similar between all subjects and those who were either or both HBV and HCV infection positive, except that fewer infection-positive subjects with higher coffee consumption tended to be past drinkers and more had a lower ALT level. Body mass index did not markedly differ by coffee consumption in those who were either or both HCV and HBV positive (Table 1).

Subjects with higher green tea consumption tended to consume less alcohol and less coffee, have less history of diabetes mellitus, and have a lower ALT level. These subjects had a greater tendency to be smokers, particularly those with either or both HCV and HBV infection (Table 2).

HR (95% CI) of liver cancer according to coffee consumption are shown in Table 3. Compared with those who almost never drank coffee, increased coffee consumption was associated with a lower risk of liver cancer (<1 cup/d: HR, 0.67; 95% CI, 0.42-1.07; 1-2 cups/d: HR, 0.49; 95% CI, 0.27-0.91; ≥3 cups/d: HR, 0.54; 95% CI, 0.21-1.39; P_trend = 0.025). A similar tendency was observed when analysis was restricted to those with either or both HCV and HBV infection (<1 cup/d: HR, 0.55; 95% CI, 0.33-0.93; 1-2 cups/d: HR, 0.47; 95% CI, 0.24-0.93; ≥3 cups/d: HR, 0.61; 95% CI, 0.23-1.62; P_trend = 0.036) and when further restricted to those with HCV infection (<1 cup/d: HR, 0.56; 95% CI, 0.32-0.99; 1-2 cups/d: HR, 0.40; 95% CI, 0.18-0.88; ≥3 cups/d: HR, 0.78; 95% CI, 0.28-2.15; P_trend = 0.065). When analyzed by sex, a more significant reduction in risk was observed in men than in women.

The association between liver cancer risk and green tea consumption differed to that with coffee (Table 4). No significant association with risk was observed for total subjects, those with either or both HCV and HBV infection, or those with HCV infection. No clear difference was seen by sex.

Among covariates in the multivariate analysis, we found significant positive associations with a history of diabetes mellitus, smoking, high body mass index, high ALT level, and positivity for HCV and HBV infection and a significant inverse association with alcohol drinking.

In this prospective cohort study among a large Japanese population, we found that coffee and green tea consumption differed in their association with the risk of liver cancer. Specifically, coffee consumption reduced the risk of liver cancer in all subjects as well as in those who were either or both HCV and HBV infection positive. In contrast, green tea consumption showed no significant association with the risk of liver cancer overall or by HCV or HBV status.

Few studies investigating the association of coffee and green tea with liver cancer risk have taken account of hepatitis virus infection status, although HBV and HCV infection are the most important risk factors of liver cancer. Because it is difficult to obtain baseline data on virus infection status for the large populations, this problem is particularly prominent in cohort studies. A previous report from the present cohort reported an inverse association between coffee and liver cancer (5), but this finding was weakened by the lack of data on hepatitis virus infection, which could therefore not be controlled. For the present analysis, we newly analyzed the association of coffee and green tea with liver cancer by determining infection status at baseline using stored blood samples and extending the follow-up period. To our knowledge, this study is the first prospective cohort analysis of this association to consider hepatitis virus infection status.

A growing number of epidemiologic studies have observed a reduced risk of liver cancer with increased consumption of coffee (5-15), and several meta-analyses have provided confirmation (16, 17). Although several of the case-control (10, 11, 13-15) and nested case-control (8, 9) studies took HCV or HBV infection status into consideration, most adjusted for HCV or HBV infection status in the model, and only a few observed risk among those who were HCV or HBV infection positive (8, 10, 13, 14). Most observed a decreased risk trend for liver cancer with increased coffee consumption following adjustment for HCV and HBV infection status (8, 9, 11-13) or among HCV- or HBV-positive subjects (8, 10, 13, 14). Consistent with these previous reports, our present results show a clear and significant inverse association between coffee consumption and liver cancer both among total subjects after controlling for HCV and HBV infection status and among those who were either or both HCV and HBV positive, although statistical significance was observed for only one of the consumption categories. Biologically, three major components of coffee have been considered to contribute to the beneficial effects of coffee on liver carcinogenesis: chlorogenic acids, caffeine, and the coffee diterpenes cafestol and kahweol (27, 28). Because diterpene content is negligible in instant, drip-filtered, and percolated brews (29), the preparation methods most commonly used by Japanese, the main putative candidates in our population are chlorogenic acids and caffeine. Further, coffee and caffeine have been suggested to improve liver enzyme activity (30-34), and coffee has also been associated with a reduced risk of liver disease and cirrhosis (35-37), a major risk factor or pathogenic step in the process of liver carcinogenesis. Thus, coffee may act to mitigate the inflammation of liver cells, suppress the aggravation of liver disease, and ultimately prevent the development of liver cancer (10).

Although several studies have reported the beneficial effects of coffee on abnormal liver biochemistry, cirrhosis, and liver cancer (38), the mechanism of the effect of coffee on HCV and HBV is not well understood. Our present results are consistent with previous findings (13, 35), which suggest that coffee may be protective against chronic liver disease no matter whether it is caused by HCV or HBV infection. In other words, coffee may act on the clinical condition underlying the chronic liver disease rather than on the HCV or HBV itself (13) by the mechanisms mentioned above (10). Further basic research to clarify the mechanism of this effect is warranted.

The effect of green tea on liver cancer is not well understood. Several experimental studies have suggested a hepatoprotective effect and anticarcinogenic potential in the liver (39, 40). Only a few epidemiologic studies have evaluated risk (10, 18); however, one of these observed a null association, although without considering HCV or HBV infection (18), whereas the second observed an elevated risk with increased green tea consumption among those positive for HCV (10). Our present results and a previous report in HCV-positive subjects suggest that green tea consumption does not appear to be inversely associated with liver cancer. Green tea contains vitamin C, which has antioxidant potential, but is also known to stimulate iron uptake from food (41). In the Framingham Heart Study, dietary vitamin C was positively associated with ferritin, which was used as a measure of body iron stores, whereas coffee intake was inversely associated with it (42, 43). Given the integral role of excess iron in the progression of hepatic fibrosis (44), higher consumption of green tea might harm hepatic cells by increasing iron uptake from food via dietary vitamin C. Although the contribution of green tea intake to total vitamin C intake in Japanese is ∼13% (45), it is possible that the positive effect of green tea, if any, might be due to vitamin C, at least in part, and that the lack of association for green tea might result from the opposing effects of its constituents. A better understanding of this issue will require further research.

The major strength of the present study is its prospective design, in which information was collected before the subsequent diagnosis of liver cancer, thereby avoiding the exposure recall bias inherent to case-control studies. A second strength was that virus infection status was determined at baseline for the entire population, allowing us to clarify the association in a high-risk population using a prospective design. Other strengths include study subjects were selected from the general population, the proportion of loss to follow-up (0.3%) was negligible, the quality of our cancer registry system was satisfactory over the study period, and potential confounding factors could be adjusted to minimize their influence on risk values in spite of the possible influence of residual confounding.

Against this, several obvious limitations can be identified. First, the correlation coefficients were moderate for coffee and even lower for green tea. Inaccurate measurement of coffee and green tea consumption necessarily results in random misclassification, which in turn attenuates the true association. Second, misclassification might also have resulted from our evaluation of coffee and green tea consumption by a single, self-reported measurement at baseline. If present, however, this would likely have been nondifferential and led to an underestimation of results. Third, although the study was based on a cohort of ∼20,000 Japanese with 13 years of follow-up, the number of cases was nevertheless relatively small, and the possibility that the results were due to chance cannot be ruled out. Fourth, we had no information on the clinical severity of hepatitis or the treatment of subjects with HCV or HBV infection before and during the study period. If infected subjects had received treatment, the occurrence of liver cancer may have been decreased. Fifth, our study subjects were restricted to those who responded to the questionnaire and provided blood samples when they participated in the baseline health checkup, suggesting that these individuals may not be representative of the entire cohort. The incidence of liver cancer in this study population during the follow-up period was 46.1 per 100,000 person-years versus 67.5 in the whole Japan Public Health Center-Based Prospective Study, suggesting that subjects who were already under care for HCV or HBV infection may have been less willing to attend a health checkup. For these reasons, interpreting or generalizing these results should be done with care.

Allowing for these methodologic issues, we found that coffee and green tea consumption differed in their association with the risk of liver cancer. Increased coffee consumption reduced the risk of liver cancer regardless of HCV and HBV infection status, whereas green tea consumption showed no significant association with the risk of liver cancer overall or by HCV and HBV status. Although both coffee and green tea have anticarcinogenic potential, these two beverages might differ in their overall effect on the human liver. The mechanisms underlying this difference, including iron uptake, warrant further clarification.

No potential conflicts of interest were disclosed.

The members of the Japan Public Health Center-Based Prospective Study (principal investigator: S. Tsugane) Group are as follows: S. Tsugane, M. Inoue, T. Sobue, and T. Hanaoka, National Cancer Center, Tokyo; J. Ogata, S. Baba, T. Mannami, A. Okayama, and Y. Kokubo, National Cardiovascular Center, Osaka; K. Miyakawa, F. Saito, A. Koizumi, Y. Sano, I. Hashimoto, T. Ikuta, and Y. Tanaba, Iwate Prefectural Ninohe Public Health Center, Iwate; Y. Miyajima, N. Suzuki, S. Nagasawa, Y. Furusugi, and N. Nagai, Akita Prefectural Yokote Public Health Center, Akita; H. Sanada, Y. Hatayama, F. Kobayashi, H. Uchino, Y. Shirai, T. Kondo, R. Sasaki, Y. Watanabe, Y. Miyagawa, and Y. Kobayashi, Nagano Prefectural Saku Public Health Center, Nagano; Y. Kishimoto, E. Takara, T. Fukuyama, M. Kinjo, M. Irei, and H. Sakiyama, Okinawa Prefectural Chubu Public Health Center, Okinawa; K. Imoto, H. Yazawa, T. Seo, A. Seiko, F. Ito, F. Shoji, and R. Saito, Katsushika Public Health Center, Tokyo; A. Murata, K. Minato, K. Motegi, and T. Fujieda, Ibaraki Prefectural Mito Public Health Center, Ibaraki; K. Matsui, T. Abe, M. Katagiri, M. Suzuki, and K. Matsui, Niigata Prefectural Kashiwazaki and Nagaoka Public Health Center, Niigata; M. Doi, A. Terao, Y. Ishikawa, and T. Tagami, Kochi Prefectural Chuo-higashi Public Health Center, Kochi; H. Sueta, H. Doi, M. Urata, N. Okamoto, and F. Ide, Nagasaki Prefectural Kamigoto Public Health Center, Nagasaki; H. Sakiyama, N. Onga, H. Takaesu, and M. Uehara, Okinawa Prefectural Miyako Public Health Center, Okinawa; F. Horii, I. Asano, H. Yamaguchi, K. Aoki, S. Maruyama, M. Ichii, and M. Takano, Osaka Prefectural Suita Public Health Center, Osaka; Y. Tsubono, Tohoku University, Miyagi; K. Suzuki, Research Institute for Brain and Blood Vessels Akita, Akita; Y. Honda, K. Yamagishi, and S. Sakurai, Tsukuba University, Ibaraki; M. Kabuto, National Institute for Environmental Studies, Ibaraki; M. Yamaguchi, Y. Matsumura, S. Sasaki, and S. Watanabe, National Institute of Health and Nutrition, Tokyo; M. Akabane, Tokyo University of Agriculture, Tokyo; T. Kadowaki, Tokyo University, Tokyo; M. Noda and T. Mizoue, International Medical Center of Japan, Tokyo; Y. Kawaguchi, Tokyo Medical and Dental University, Tokyo; Y. Takashima, u, Sakihae Institute, Gifu; H. Sugimura, Hamamatsu University, Shizuoka; S. Tominaga, Aichi Cancer Center Research Institute, Aichi; H. Iso, Osaka University, Osaka; M. Iida, W. Ajiki, and A. Ioka, Osaka Medical Center for Cancer and Cardiovascular Disease, Osaka; S. Sato, Osaka Medical Center for Health Science and Promotion, Osaka; E. Maruyama, Kobe University, Hyogo; M. Konishi, K. Okada, and I. Saito, Ehime University, Ehime; N. Yasuda, Kochi University, Kochi; S. Kono, Kyushu University, Fukuoka.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.