Green tea consumption and colorectal cancer risk: a report from the Shanghai Men’s Health Study

Abstract

Tea and its constituents have demonstrated anticarcinogenic activity in both in vitro and in vivo animal studies. Results from epidemiologic studies, however, have been inconsistent. Some factors that coexist with tea consumption, such as cigarette smoking, may confound or modify the association between tea consumption and cancer risk. The objective of this study was to comprehensively evaluate the association between green tea consumption and colorectal cancer risk in a population-based prospective cohort study, the Shanghai Men’s Health Study. The analysis included 60 567 Chinese men aged 40–74 years at baseline. During ∼5 years of follow-up, 243 incident cases of colorectal cancer were identified. Multivariable Cox regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of developing colorectal cancer. Regular green tea consumption (ever drank green tea at least three times per week for more than six consecutive months) was associated with reduced risk of colorectal cancer in non-smokers (multivariable-adjusted HR = 0.54, 95% CI: 0.34–0.86). The risk decreased as the amount of green tea consumption increased (P_trend = 0.01). Each 2 g increment of intake of dry green tea leaves per day (approximately equivalent to the amount of tea in a tea bag) was associated with a 12% reduction in risk (HR = 0.88, 95% CI: 0.78–0.99). No significant association was found among smokers (HR = 0.94, 95% CI: 0.66–1.34). This study suggests that regular consumption of green tea may reduce colorectal cancer risk among non-smokers.

Introduction

Colorectal cancer is the third most common cancer worldwide (1) and the third leading cause of cancer mortality in many countries (1,2). Although the incidence rate of colon cancer over the past three decades has doubled in both men and women in Shanghai, China (3,4), the rate is still considerably lower than in Western countries (1), suggesting that potential protective factors, such as diet and other lifestyle factors, may play a role in risk for colorectal cancer in this population (5,6).

Tea is one of the most commonly consumed beverages in the world. Non-herbal teas (green, black and oolong) are all made from the leaf of Camellia sinensis. Both green and black tea and their constituents, such as tea polyphenols, have been shown to have many cancer-inhibitory properties in numerous in vitro and in vivo animal studies (7,8). Green tea, however, contains a much higher level of antioxidant tea polyphenols than black or oolong tea since green tea is made from unfermented unoxidized tea leaves. Organ sites that are accessible directly to orally administered tea, such as the digestive tract, are thought to represent good targets for potential chemoprevention by tea because of the high bioavailability of bioactive tea compounds (7). Administration of tea infusion or tea polyphenols to animals has been shown to reduce the occurrence and regress the growth of experimental tumors (8–10).

The association between tea consumption and colorectal cancer risk has been previously evaluated in several populations, mostly in studies with a case–control study design, and results have been inconsistent (11,12). Differences in tea consumption habits, types and amounts of tea consumed may have contributed to some of the inconsistencies (11–13). Failure to account for effect modifiers, improper control of confounding factors and inadequate assessment of tea consumption may also be important contributors (11,14). In particular, tea drinkers are also likely to smoke cigarettes and drink alcoholic beverages in many populations; failure to take these factors into consideration could have led to biased assessment of associations with tea consumption. In previous studies, we found that cigarette smoking conferred substantial effect modification on the association between tea consumption and cancer risk (15,16). Decreased risk associated with tea consumption for cancers of the esophagus (15) and lung (16) was more pronounced among never-smokers.

In this report, we describe a comprehensive evaluation of the association of green tea consumption with colorectal cancer risk in a large prospective cohort study conducted in urban Shanghai, the Shanghai Men’s Health Study (SMHS). In particular, we evaluated whether the association between green tea consumption and colorectal cancer was confounded or modified by cigarette smoking, a factor that often coexists among Chinese men who drink tea.

Materials and methods

Study participants

The SMHS is a population-based prospective cohort study conducted in urban Shanghai. The study was approved by the relevant Institutional Review Boards for human research in China and the USA, and written informed consent was obtained from all study participants. The details of the study design and methods have been published elsewhere (17). Briefly, between 2002 and 2006, the SMHS recruited 61 500 men who were aged 40–74 years, were free of cancer, and lived in one of the eight selected urban communities in Shanghai (participation rate: 74.1%). In-person interviews were conducted to obtain information on demographics, lifestyle and dietary habits, medical history and other characteristics through structured questionnaires (18,19). Anthropometric measurements, including current weight, height and circumferences of the waist and hips, were also taken.

The cohort was followed for occurrence of cancer and other chronic diseases by a combination of active surveys conducted every 3 years and annual record linkage of the cohort to databases of the population-based Shanghai Cancer Registry and the Shanghai Municipal Vital Statistics Unit. Nearly, all cohort members were successfully followed; the response rate for the first in-person follow-up survey (2004–2008) was 97.6%. All possible incident cancer cases were verified through home visits. Medical charts were reviewed to verify cancer diagnoses and to collect detailed diagnostic information.

In this study, we excluded participants who drank black or oolong tea regularly and exclusively or other types of tea such as herbal tea (n = 899, 1.46%; number of events = 7), reported extreme total energy intake (<500 or >4500 kcal/day) (n = 34; number of events = 0), or were lost to follow-up immediately after study enrollment (n = 82; number of events = 0). After these exclusions (not mutually exclusive), a total of 60 567 men remained for the present study.

Assessment of tea consumption

Tea consumption was assessed at the baseline survey for all participants. Each participant was asked whether he drank tea regularly (ever drank tea at least three times per week for more than six consecutive months) and at what age he started this habit, followed by questions on the type and amount (dry weight) of tea (tea leaves) consumed during the year preceding the interview as well as the current status of tea consumption. Those who were former tea drinkers were further asked about the age at which they last drank tea regularly.

Statistical analysis

Hazard ratios (HRs) of developing colorectal cancer and their 95% confidence intervals (CI) were estimated by using the Cox proportional hazards regression model, with age as the time scale. Entry time was defined as age at study enrollment and exit time was defined as age at colorectal cancer diagnosis or censoring (either 31 December 2008, the date of the latest record linkage or the date of death), whichever came first. A total of 60 567 participants included in this analysis contributed 273 776 person-years in a mean follow-up of 4.6 years between 2002 and 2008. Among green tea drinkers, the amount and years of green tea consumption were dichotomized according to the median amount or years of consumption; non-drinkers of tea were used as the reference. Tests for trend were performed by entering categorical variables as continuous variables in the model.

Potential confounding variables were chosen based on a priori considerations, including age at baseline (continuous), education (two categories), cigarette smoking (ever smoked at least one cigarette per day for more than six consecutive months; yes or no), pack-years (average number of packages of cigarettes smoked per day multiplied by the number of years smoked), alcohol consumption (ever drank beer, wine or spirits at least three times per week for more than six consecutive months; yes or no), regular participation in leisure-time exercise (during the past 5 years, exercised at least one time per week for more than three consecutive months; yes or no), body mass index [BMI (kg/m²), calculated as weight in kilograms divided by the square of height in meters; continuous], history of diabetes (yes or no), family history of colorectal cancer (yes or no) and intakes (continuous) of total fruits, vegetables and red meat. Additional adjustments for use of non-steroidal anti-inflammatory drugs and other dietary factors, such as intakes of total calories, dietary fiber, calcium, and folic acid, did not appreciably alter the results; these variables were, therefore, not included in the final model. We also conducted sensitivity analyses by excluding early years of observation or stratifying by tumor stage to address the potential influence of prediagnosed disease on the risk estimates.

Stratified analyses were conducted to assess possible effect modification by age, cigarette smoking, alcohol consumption and other known risk factors for colorectal cancer, such as body mass index, physical activity and intakes of red meat, fruits and vegetables. Multiplicative interactions were determined by use of the likelihood ratio test between the main effect model and models with both the main effect and cross-product terms. The ‘proportionality assumption’ that underlies the Cox model was checked by graphical methods and was found not to have been violated. Statistical analyses were carried out using SAS version 9.1 (SAS Institute, Cary, NC). All statistical tests were based on two-sided probability and P <0.05 was considered statistically significant.

Results

Approximately 66.6% of cohort members were regular drinkers of green tea. Many people in this study started drinking green tea regularly in early adulthood (median age = 25 years), and the majority (95.5%) had maintained the habit since then. Among green tea drinkers, the median monthly consumption amount was 250 g (dry weight) of green tea leaves (interquartile range: 150–425 g). The median duration of lifetime green tea consumption at baseline was 25 years (interquartile range: 19–32 years).

Baseline characteristics of the study population according to status of green tea consumption are presented in Table I. Compared with non-drinkers of green tea, regular drinkers of green tea were slightly younger and were more likely to smoke cigarettes and drink alcoholic beverages. They were also likely to be less physically active and to consume more red meat.

Table I.

Baseline characteristics by green tea consumption habits, the SMHS^a

		Regular green tea consumption
Characteristics	All participants	No	Yes
No. of participants	60 567	20 212	40 355
Age (years)	54.9 (9.7)	56.3	54.2
Education (high school and above), %	58.8	57.6	59.4
Ever smoked regularly, %	69.5	52.6	78.0
Ever drank alcohol regularly, %	33.6	22.1	39.4
Exercised regularly, %	35.5	38.6	34.0
Family history of colorectal cancer, %	2.1	2.1	2.1
History of diabetes, %	6.3	6.2	6.3
Body Mass Index (kg/m2)	23.7 (3.1)	23.6	23.8
Vegetable intake (g/day)	343.7 (192.1)	336.2	347.0
Fruit intake (g/day)	151.5 (125.7)	156.1	149.2
Red meat intake (g/day)	63.1 (44.1)	58.6	65.4

^aUnless otherwise indicated, data are expressed as the mean (standard deviation). All tests for differences between green tea drinkers and non-drinkers were statistically significant (P < 0.05), except for history of diabetes and family history of colorectal cancer.

During a mean follow-up of 4.6 years between 2002 and 2008, 243 incident cases of colorectal cancer, including 133 colon and 110 rectal cancers, were identified; 123 patients had early-stage tumors and 93 had late-stage tumors. The mean age at diagnosis was 66.4 years (standard deviation, 8.7) and ranged from 45 to 79 years. Table II presents HR for colorectal cancer associated with green tea consumption at baseline. Overall, green tea consumption was inversely associated with colorectal cancer risk, but the result was only marginally significant (P = 0.06); multivariable-adjusted HR was 0.77 (95% CI: 0.59–1.01). This inverse association was statistically significant among non-smokers (HR = 0.54, 95% CI: 0.34–0.86) but was not observed among smokers (HR = 0.94, 95% CI: 0.66–1.34), although the test for multiplicative interaction between green tea consumption and ever cigarette smoking was not statistically significant (the test for interaction between green tea consumption and pack-years of cigarette smoking was borderline significant, P = 0.06). Demographic, lifestyle and dietary factors did not significantly confound the association (Table II); further adjustment for total caloric intake did not alter the results. Among non-smokers, the direction and magnitude of the association were similar for both colon cancer (HR = 0.51, 95% CI: 0.28–0.93) and rectal cancer (HR = 0.59, 95% CI: 0.29–1.22; P_{heterogeneity} = 0.42), although the latter was not statistically significant (probably due to a limited number of events in the subgroup).

Table II.

HR and 95% CI of colorectal cancer associated with green tea consumption, the SMHS (2002–2008)

		Colorectal cancer		Colon cancer		Rectal cancer
Regular green tea consumption	Person-years	No. of cases	HR (95% CI)a	No. of cases	HR (95% CI)a	No. of cases	HR (95% CI)a
All participants
No	91 040	100	1.0	60	1.0	40	1.0
Yes	182 736	143	0.77 (0.59, 1.01)	73	0.69 (0.48, 0.98)	70	0.89 (0.59, 1.34)
Non-smokers
No	43 744	54	1.0c	33	1.0	21	1.0
Yes	41 347	29	0.54 (0.34, 0.86)	17	0.51 (0.28, 0.93)	12	0.59 (0.29, 1.22)
Smokersb
No	47 290	46	1.0c	27	1.0	19	1.0
Yes	141 389	114	0.94 (0.66, 1.34)	56	0.82 (0.51, 1.30)	58	1.10 (0.65, 1.86)

^aHRs were adjusted for age, education, cigarette smoking, pack-years of cigarette smoking, alcohol consumption, regular exercise, body mass index, history of diabetes, family history of colorectal cancer and intakes of vegetables, fruits and red meat in a Cox model with age as the time scale.

^bMissing data on status of cigarette smoking for one participant.

^cP = 0.13 for interaction between ever smoking and green tea consumption; P = 0.06 for interaction between pack-years of cigarette smoking and green tea consumption.

Among non-smokers, risk of colorectal cancer decreased as the amount of green tea consumption increased (P_trend = 0.01) as indicated by a trend test in which the categorical variable of green tea consumption was treated as a continuous variable in the model (Table III). Compared with non-drinkers of green tea, men consuming >250 g/month of dry green tea leaves had a multivariable HR of 0.45 (95 CI: 0.23–0.90). When intake of green tea was analyzed on a continuous basis, each 2 g increment of intake of dry green tea leaves per day (approximately equivalent to the amount of tea in a tea bag) was associated with a 12% reduction in risk (HR = 0.88, 95% CI: 0.78–0.99, P = 0.03) after fully adjusting for potential confounding factors. No association with the amount of green tea consumption was found for colorectal cancer in cigarette smokers (P_interaction = 0.04). However, we found that longer duration of green tea consumption (i.e. drinking tea for more years) was unrelated to lower risk, either overall or by smoking status.

Table III.

HR and 95% CI of colorectal cancer associated with amount and years of green tea consumption, the SMHS (2002–2008)

	No. of cases	HR (95% CI)a
Amount of green tea consumedbc (g/month)
All participants
Never	100	1.0
<250 (median)	48	0.66 (0.46, 0.93)
≥250	91	0.85 (0.62, 1.15)
Ptrend		0.27
Non-smokers
Never	54	1.0d
<250	19	0.62 (0.37, 1.05)
≥250	10	0.44 (0.22, 0.88)
Ptrend		0.009
Smokers
Never	46	1.0d
<250	29	0.70 (0.44, 1.11)
≥250	81	1.04 (0.72, 1.50)
Ptrend		0.66
Years of green tea consumptionc
All participants
Never	100	1.0
<25 (median)	45	0.75 (0.52, 1.07)
≥25	97	0.78 (0.58, 1.06)
Ptrend		0.12
Non-smokers
Never	54	1.0e
<25	10	0.48 (0.24, 0.94)
≥25	18	0.56 (0.33, 0.97)
Ptrend		0.02
Smokers
Never	46	1.0e
<25	35	0.95 (0.60, 1.49)
≥25	79	0.96 (0.66, 1.39)
Ptrend		0.83

^aHRs were adjusted for age, education, cigarette smoking, pack-years of cigarette smoking, alcohol consumption, regular exercise, Body Mass Index, history of diabetes, family history of colorectal cancer and intakes of vegetables, fruits and red meat in a Cox model with age as the time scale.

^bMeasured based on dry weight of green tea leaves.

^cMissing data on the amount of green tea consumption for 234 participants and the years of green tea consumption for 318 participants.

^dP = 0.04 for interaction between ever smoking and amount of green tea consumed; P = 0.03 for interaction between pack-years of cigarette smoking and amount of green tea consumed.

^eP = 0.16 for interaction between ever smoking and years of green tea consumption; P = 0.06 for interaction between pack-years of cigarette smoking and years of green tea consumption.

Sensitivity analyses were conducted by excluding the early years of observation and colorectal cancer cases occurring during the same time period in order to minimize the possible effect of lifestyle changes related to subclinical disease. Among non-smokers of cigarettes, the HRs (95% CI) associated with green tea consumption were not materially altered when excluding the first year of observation, 0.59 (0.35–0.99) or when excluding the first 2 years of observation, 0.54 (0.29–1.00). Nor did the association vary substantially by tumor stage, with HRs (95% CI) of 0.51 (0.28–0.95) for early-stage tumors (TNM stage I or II) and 0.58 (0.29–1.17) for late-stage tumors (TNM stage III or IV), although the latter HR was not statistically significant (probably due to a limited number of events (n = 35) in the subgroup).

With the exception of cigarette smoking, the association between green tea consumption and colorectal cancer did not vary by other known risk factors for colorectal cancer, including age, body mass index, waist-to-hip ratio, physical activity or intakes of red meat, fruits and vegetables. In stratified analyses, continuous covariates were dichotomized based on the median of these factors in the entire cohort. Because diabetic patients tend to drink more water and also have a higher risk for colorectal cancer, we conducted analyses excluding individuals who reported a history of diabetes at baseline and censored the observation at diagnosis of diabetes during follow-up. No material changes in the risk estimates for colorectal cancer were found.

Discussion

In this large population-based prospective cohort study, green tea consumption was inversely associated with overall risk of colorectal cancer (P = 0.06, marginally significant). This inverse association was statistically significant among non-smokers after Bonferroni correction for multiple comparisons. No significant association was found among smokers.

Green tea contains many polyphenolic compounds, mainly catechins, comprising 30–40% of the extractable solids of dried green tea leaves (8). These compounds, especially epigallocatechin-3-gallate, the major catechin of green tea, have been shown to have many cancer-inhibitory properties in both in vitro and in vivo animal studies. These include antioxidative and anti-inflammatory activities (20,21), inhibition of the formation of N-nitroso compounds and heterocyclic amines (potential carcinogens for colorectal cancer) (22), modulation of xenobiotic metabolizing enzymes (23), trapping of activated forms of carcinogens (24) and modulation of signal transduction pathways, which leads to inhibition of cell proliferation and transformation, induction of apoptosis and cell cycle arrest and inhibition of tumor invasion and angiogenesis (8,25,26). Because green tea catechins are not completely absorbed by the gut, catechins can be present in native form at high concentrations in the intestinal lumen (27). Thus, it has been postulated that the digestive tract may represent a good target for potential chemoprevention with tea because of the high bioavailability of tea catechins in digestive tract organs (7).

Green tea is the most commonly consumed tea in Shanghai, accounting for 98.6% of tea consumed by male tea drinkers. The finding of an inverse association between green tea consumption and colorectal cancer risk in our prospective cohort study is consistent with our previous findings from two large case–control studies in Shanghai (28,29) as well as another prospective cohort study, the Shanghai Women’s Health Study (SWHS) (30). In both case–control studies, which involved 1328 and 1805 incident cases of colorectal cancer, respectively, green tea consumption was found to be associated with a reduced risk of colorectal cancer in a dose–response manner (28,29). The inverse association was further confirmed in an analysis conducted among participants of the SWHS, a cohort of 74 942 women, 97.3% of whom are non-smokers (30). Results from other previous case–control studies, mainly conducted in Japan, have been mixed but generally pointed toward an inverse association, with a summary odds ratio of 0.74 (95% CI: 0.63–0.86) (12). Five cohort study reports on this topic, including our report from the SWHS, have been published to date. Three cohort studies reported an inverse association of green tea consumption with both colon and rectal cancers (30), colorectal cancer combined (31) or rectal cancer only (32); two studies reported a null association (32,33); and one study suggested no effect of green tea on colorectal cancer in women, but increased risk for advanced stage cancer in men (34). A potential effect modification by cigarette smoking on the association was not evaluated in the study (34). In contrast, we found that green tea consumption was associated with reduced risk for both early-and late-stage colorectal cancers in our earlier report on female non-smokers (30) and the current study of male non-smokers.

It should be noted that many of the previous studies on tea and colorectal cancer risk were not designed specifically to evaluate the effect of tea, and thus, data on tea consumption were usually not comprehensive (few studies collected information on amount or years of tea consumption) (11). The lack of heterogeneity of tea consumption in some study populations, e.g. ∼95% of study participants drinking tea every day (35), may have hindered these studies from evaluating the association of cancer risk with green tea consumption. In addition, inadequate controlling for potential confounding or failure to account for effect modifiers may also have contributed to the inconsistency (11,12). In particular, the habit of drinking tea often coexists with cigarette smoking and/or alcohol consumption in many populations, which has been found to confer a substantial effect modification on the association between tea consumption and cancer risk (15,16). In the present study, consumption of green tea was associated with an overall 23% reduction in risk of colorectal cancer. After accounting for effect modification by cigarette smoking, we found that risk was reduced by 46% among non-smokers, similar to findings observed in our previous study of esophageal cancer (15). The strong correlation of tea consumption with cigarette smoking, alcohol consumption, physical inactivity and higher intake of red meat may, in part, explain the null association we observed between tea consumption and cancer risk in smokers. It is also possible that the anticancer effect of green tea consumption is visible only in groups with lower risk of colorectal cancer but may not be strong enough to reverse the adverse effects of other concurrent risk factors such as cigarette smoking, as observed in this study.

In this study, a significant dose–response relationship was observed for the amount of green tea consumed (P_trend = 0.01) but not for the duration in years of lifetime green tea consumption. However, in our earlier report from the SWHS, we found that the reduction in risk was most evident among women with longer duration of green tea consumption (30). The median duration of lifetime green tea consumption in women was 16 years, with a 3-fold interquartile range from 8 to 24 years. In the present study, the variation in years of green tea consumption in men was relatively small (interquartile range: 19–32), which may have limited our ability to evaluate the dose–response relationship for lifetime duration of green tea consumption.

Most tea consumed in Western societies is black tea. Frequent consumption of black tea has also been associated with reduced risk for digestive tract cancers in some (36,37) but not all studies (12,38). In our study, only 743 (1.2%) participants of the SMHS drank black tea regularly and exclusively, which limited our ability to assess the association of black tea with colorectal cancer risk. The association between drinking coffee and colorectal cancer risk has also been evaluated in several previous studies (38). We did not collect information on coffee consumption since coffee is not a commonly consumed beverage in this population.

The specifics of tea consumption habits vary substantially among and between populations with regard to the types and amounts of tea consumed. In studies of tea and health, the number of cups of tea consumed has been commonly used to assess the amount of exposure. The bioactivity of a cup of tea is affected by many factors, especially the amount of dry tea leaves used for tea preparation (39). A recent case–control study found a significant reduction in the risk of rectal cancer with increasing intake of black tea measured in grams of dry weight consumed per month (40). The observed association, however, was substantially diminished when tea consumption was measured in liters of tea consumed. This observation suggests that differences in the methods of assessing tea consumption may have contributed to the conflicting results reported by previous epidemiologic studies. The most common method of green tea preparation in Shanghai is to brew dry green tea leaves with hot water. Higher green tea consumption measured in dry weight has been consistently associated with reduced risk of cancers of the colon and rectum (28–30), esophagus (15), stomach (41), pancreas (29) and lungs (16) in studies conducted in Shanghai.

The current study has several notable strengths. We comprehensively evaluated green tea consumption and colorectal cancer risk by both the amount and duration of green tea consumption. Other strengths of the study include a population-based prospective study design, high participation rates and a virtually complete cohort follow-up, all of which help to minimize many sources of bias inherent in case–control studies. However, as with any observational study, some error in the measurement of tea consumption is likely. Because the exposure assessment was conducted prospectively and prior to cancer diagnosis, it is possible that any measurement error is non-differential by case/control status, which would tend to attenuate the true association between green tea consumption and colorectal cancer risk. In addition, residual confounding may be a potential concern, although we carefully adjusted for a wide range of potential confounding factors, including socioeconomic status and known risk factors for colorectal cancer.

In conclusion, this prospective cohort study among men in Shanghai suggests that regular consumption of green tea may confer protection against colorectal cancer, in particular, for non-smokers. These findings are consistent with data from both in vitro and in vivo experiments, indicating that green tea may serve as an effective chemopreventive agent. With a longer period of follow-up, we should be able to provide a more precise risk estimate according to the duration and amount of tea consumption.

Funding

United States Public Health Service grants (R01 CA082729, in part by R01 CA122364).

Glossary

Abbreviations

CI

confidence interval

HR

hazard ratio

SMHS

Shanghai Men’s Health Study

SWHS

Shanghai Women’s Health Study