Although experimental studies have shown consistently that tea preparations and tea polyphenols may inhibit the induction of a variety of cancers, including lung cancer, 1 epidemiologic studies of tea consumption and cancer are limited and the results are inconclusive. Different research groups 1–4 have reviewed the available data and they have all concluded that more epidemiologic data are needed to determine whether drinking green and black teas reduces the risk of cancer. We developed the present study to identify the causes for the unusually high incidence rates of lung cancer among nonsmoking women living in Shanghai, People’s Republic of China, with a focus on exposures from Chinese cooking and from environmental tobacco smoke. 5,6 As we also obtained data on green tea consumption, we report herein associations in both smoking and nonsmoking women.
Subjects and Methods
The methods for this population-based case-control study have been reported in detail elsewhere. 5,6 Briefly, cases were women who were permanent residents of Shanghai, 35–69 years of age, diagnosed between February 1, 1992 and January 31, 1994 with an incident, primary lung carcinoma (International Classification of Diseases, 9th revision, 162). We identified case subjects from the population-based Shanghai Cancer Registry. For each eligible case reported to the registry, we examined the hospital chart to confirm the diagnosis. The histologic type of lung cancer was obtained from pathology reports or hospital discharge summaries. For cases whose pathologic or cytologic reports were not available, we used clinical and radiologic evidence as the basis of the diagnosis.
We selected control subjects from the general population by frequency-matching to the age distribution of new female lung cancer cases occurring in 1987–1989. Control subjects were selected randomly from the personal identification cards retained by the population-based Shanghai Residential Registry. Each card contained the name, address, date of birth, gender, and occupation of residents of Shanghai. Using the expected age distribution of the cases, we randomly selected a set of starting positions in the card file and then selected the first two cards of women whose ages fell within predetermined 5-year intervals. When a primary control subject (first card) could not be interviewed, we then contacted the alternate control subject (second card).
Subjects were interviewed either in hospital, at home, or at work by trained interviewers who used a structured questionnaire to obtain information on level of education and income, all occupations held for at least 2 years, residential history, family history of lung cancer, cooking activities, dietary habits, active smoking, environmental tobacco smoke at home and at work, alcohol use, and tea drinking.
We obtained information on the usual type of tea consumed and for regular tea drinkers the average annual amount of tea consumed over the last 5 years. The questions used to elicit tea consumption were identical to those used by researchers at the Shanghai Cancer Institute of China and the U.S. National Cancer Institute in their studies of green tea consumption. 7,8 Subjects were classified as nonregular green tea drinkers or regular green tea drinkers if they drank at least three cups of green tea per week for 1 year or more. The few subjects (16 subjects) who drank only black tea or both green and black tea were grouped together in a separate category.
Diet was assessed by inquiring about the usual frequency and amount of consumption of the primary foods consumed by residents of Shanghai (91 items). We assessed the average monthly amount (in grams) and monthly frequency of consumption 5 years before the interview, ignoring recent changes. Daily intake of nutrients was calculated by multiplying the amount of each food item consumed by the per-gram values in the National Chinese Food Composition Tables. 9
For each job held for at least 2 years, subjects were asked to describe their occupation, industry, and duties carried out, and to provide the number of years of employment. Job and industry titles were coded according to the system used in the 1982 population census of the People’s Republic of China. 10 The following occupations were considered as being at potentially high risk 11 : hat and shoe makers; timber processing; wood, bamboo, hemp, rattan, palm and straw products makers; paper manufacturers and paper product makers; stone sharpeners; welding and flame cutters; and painters. We created an indicator variable for subjects whose occupation of longest duration was in any of these high-risk occupations. We also developed variables representing cumulative duration of employment in each of these occupations. It has been found that employment in the textile industry may be protective for lung cancer, 12 but there were too few subjects to include in the statistical model.
We computed odds ratios (ORs) between the risk of lung cancer and drinking of tea and associated 95% confidence intervals (95% CIs) using unconditional logistic regression. 13 To gain insight into patterns of exposure response and to provide a finer level of control of confounding, we used the generalized additive models. 14 These models allow the fit of continuous independent variables using smoothing techniques. We used locally weighted regression smoothers (LOESS) because of their flexibility in specifying the amount of smoothing. We entered age, income, daily intake of vitamin C, total pack-years of cigarettes smoked, and number of years of exposure to environmental tobacco smoke at work as loess terms in the models [smoothing parameter (span) set at 25% of the data]. Cooking at very high temperatures, family history of lung cancer, high-risk occupations, and respondent status were entered as categorical variables. For the amount of green tea consumed, appropriate cutpoints were selected from the loess plots.
Of the 706 women with lung cancer, 57 (8.1%) had died before they could be asked to participate in the study. One hundred eight of 675 control subjects (16.0%) were alternate subjects, enrolled mainly because the primary control subject had moved out of Shanghai and could not be traced. Response rates were 91.9% (649 of 706) and 86.2% [675 of (108 + 675)] in the case and control groups, respectively. Interviews were conducted with surrogate respondents for 17 (2.5%) and 141 (21.7%) of control and case subjects, respectively. Fifty-three per cent of these surrogate interviews were conducted with the husbands of subjects, and the remaining surrogate respondents were the subjects’ parents or children.
Four hundred seventy-three (72.9%) lung cancer cases were diagnosed by tissue biopsy or by cytology, and 176 (27.1%) by repeated radiographs combined with other clinical evidence. Among the 473 pathologically or cytologically confirmed lung cancer cases, adenocarcinoma was the predominant histologic type (70.0%). The remaining cell types were distributed as follows: 83 squamous cell carcinomas (17.5%), 13 small-cell carcinomas (2.7%), one large-cell carcinoma (0.2%), and 45 mixed-cell carcinomas (9.5%).
Table 1 shows the distributions of potential confounding variables. For the purposes of presentation in this table, quartiles of the distributions in the control group were used as cutpoints for all continuous variables except age. As expected, a greater percentage of case subjects than control subjects were smokers; case subjects also tended to have lower daily intake of vitamin C and beta-carotene, more often reported cooking food at high temperatures, and had positive family histories of lung cancer.
Among control subjects, 25.7% (19 of 74) of subjects who smoked and 20.0% (120 of 601) of subjects who did not smoke were regular green tea drinkers. Among tea drinkers, 94.1% (253 of 269) of them drank green tea exclusively, and the remainder (16) drank either both green tea and black tea or black tea alone. Table 2 shows the associations between lung cancer and consumption of green tea by smoking status. These estimates were adjusted for all of the variables listed in Table 1 as well as respondent status. Because of a high correlation between the intake of vitamin C and beta-carotene (Pearson correlation coefficient = 0.9), we found it sufficient to include only the intake of vitamin C in the regression models. We found no important association with other dietary items, such as fat and protein (data not shown). Among nonsmokers, there was an overall reduction in the risk of lung cancer for subjects who were regular green tea drinkers (OR = 0.65; 95% CI = 0.45–0.93). Moreover, risks decreased with increasing consumption of green tea, consistent with a linear trend.
Among smokers, however, we found little association between the incidence of lung cancer and green tea consumption (OR = 0.94; 95% CI = 0.40–2.22). We found that cumulative cigarette smoking confounded the association between lung cancer and green tea consumption. For example, not including it in the model led to an increased relative risk (OR = 1.69;Table 3). Assuming that green tea consumption has no effect among smokers, we found that the model that most accurately portrayed this relation was the one in which cumulative smoking (pack-years) was included as a loess term.
An analysis by histologic type among nonsmoking women revealed a similar pattern of reduced risk of lung cancer with green tea consumption, although the trends were stronger (but based on small numbers of cases) for nonadenocarcinomas and for unknown cell types than for the adenocarcinomas (Table 4).
As a sensitivity analysis, we evaluated the potential confounding effect of occupation by using duration of employment in each of the high-risk occupations in the models; we found similar results to those analyses that included simple binary variables representing ever-employment in these high-risk occupations. In addition, we did not find any other factor that modified the effect of green tea consumption on the risk of lung cancer.
As tea is the most widely consumed beverage in the world, an association with the risk of lung cancer, however modest, has important public health implications. Despite the large number of epidemiologic studies for which risk factors for lung cancer have been investigated, in only nine 15–23 studies was the role of tea drinking examined. In five 15–19 of these nine studies no meaningful association was found, in two studies increased risks were reported, 20,21 and reduced risks were reported in the remaining two investigations. 22,23 In one 20 of the two studies showing positive associations, the ORs increased monotonically with increasing consumption. In the second study, 21 almost a 200% increase in the risk of developing lung cancer for green tea consumption was reported.
The first epidemiologic study suggesting a protective effect of tea consumption came from Japan. 22 It was reported that the risks of developing lung cancer among men who consumed the highest levels of Okinawan tea (similar to green tea) was about one-half of the risk among nondrinkers (OR = 0.57; 95% CI = 0.31–1.06); for women, the risks were further reduced (OR = 0.38; 95% CI = 0.12–1.18). In a study in Uruguay, it was found 23 that heavy drinkers of black tea (two or more cups per day) had a relative risk of 0.34 (95% CI = 0.14–0.84) compared with nondrinkers. It was also reported that the protective effects were more evident among squamous-cell and small-cell cancers and among light smokers.
The inconsistency in the reported associations between tea consumption and the risk of lung cancer may be due to statistical fluctuations within studies or to various biases being present in the studies. Our sensitivity analyses indicate that not adequately adjusting for cumulative smoking may lead to residual confounding effects. This possibility is consistent with results from a simulation study showing that regression splines can remove confounding effects better than simple parameterizations. 24
Information arising from experimental studies have shown consistently that green and black teas may inhibit the induction of a number of physically or chemically induced cancers. 1 In different experimental animal models, it was observed that oolong tea, jasmine tea, and green tea inhibited lung tumorigenesis induced by a variety of carcinogens, including urethan, 25N-nitrosodiethylamine, and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone. 1 The mechanisms of these inhibitory effects are not clear, although some researchers 1,2,26 attributed the effects to polyphenols, which account for 35–52% of the dry weight of tea. 3 Polyphenols appear to be strong antioxidants and are effective in scavenging free radicals. They may also prevent mutagenicity and genotoxicity; inhibit tumor initiation, promotion, and cell proliferation; modulate detoxification enzymes; and trap activated metabolites of carcinogens. 1,26
The present investigation has several major strengths, including the population-based design, a large group of lung cancer cases among nonsmoking women, high response rates (small losses in deceased cases and low refusal rates), and detailed information on most key potential confounding factors, including indoor air pollution from tobacco smoke and Chinese-style cooking.
Although our finding of an inverse association of tea consumption with the risk of lung cancer among nonsmokers is consistent with experimental findings and with two previous epidemiologic studies, 22,23 alternative explanations for the observed association are chance or uncontrolled confounding. It has been reported that drinking tea was associated with healthful lifestyles such as smoking less and consuming more fresh fruits. 27 Although we controlled for a number of key risk factors, it is nevertheless possible that some unidentified lifestyle factors may not have been accounted for. Another possible explanation is recall bias. Given the difference in the health status between lung cancer cases and control subjects from the general population, there may have been differential recall of tea drinking resulting from changes due to their illness. It is unlikely that such differences in recall would have been conscious, as the study subjects were not aware of the possible relation between drinking green tea and lung cancer. It is also possible that there was some misclassification in measuring the amount of green tea consumed over the last 5 years. We classified subjects as regular drinkers and nonregular drinkers, and the latter group may have included subjects who had intermittently drunk green tea but did not meet the definition of regular drinkers. It is likely, however, that misclassification was independent of lung cancer status and, therefore, it would have diluted the effect of green tea consumption. Lastly, although we made use of responses from surrogate subjects (10% of interviews), we did not observe any substantial effects after including an indicator variable to represent respondent status in the logistic regression models nor after carrying out analyses that were restricted to self-respondents (data not shown).
In conclusion, this large population-based case-control study among women living in Shanghai, China suggests that regularly drinking green tea may lower the risk of lung cancer among nonsmoking women. Our results are consistent with experimental findings that tea and tea polyphenols may inhibit tumor formation and growth.
We thank Lisa Hum for her review of the manuscript.
1. Yang CS, Wang ZY. Tea and cancer. J Natl Cancer Inst 1993; 85: 1038–1049.
2. Blot WJ, Chow WH, McLaughlin JK. Tea and cancer: a review of the epidemiological evidence. Eur J Cancer Prev 1996; 5: 425–438.
3. Bushman JL. Green tea and cancer in humans: a review of the literature. Nutr Cancer 1998; 31: 151–159.
4. International Agency for Research on Cancer. Coffee, Tea, Mate Methylxanthines and Methylglyoxal. Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol 51. Lyon, France: International Agency for Research on Cancer, 1991; 207–271.
5. Zhong L, Goldberg MS, Gao YT, Jin F. Lung cancer and indoor air pollution arising from Chinese-style cooking among nonsmoking women living in Shanghai, China. Epidemiology 1999; 10: 488–494.
6. Zhong L, Goldberg MS, Gao YT, Jin F. A case-control study
of lung cancer and environmental tobacco smoke among nonsmoking women living in Shanghai, China. Cancer Causes Control 1999; 10: 607–616.
7. Ji BT, Chow WH, Hsing AW, McLaughlin JK, Dai Q, Gao YT, Blot WJ, Fraumeni JF Jr. Green tea consumption
and the risk of pancreatic and colorectal cancers. Int J Cancer 1997; 70: 255–258.
8. Gao YT, McLaughlin JK, Blot WJ, Ji BT, Dai Q, Fraumeni JF Jr. Reduced risk of esophageal cancer associated with green tea consumption
. J Natl Cancer Inst 1994; 86: 855–858.
9. Chinese Academy of Medical Sciences. Food Composition Tables (in Chinese). Beijing: People’s Health Publishing Co, 1981.
10. China Report. Ten Percent Sampling Tabulation on 1982 Population Census of the People’s Republic of China. Beijing: Foreign Broadcast Information Services, 1985.
11. Levin LI, Zheng W, Blot WJ, Gao YT, Fraumeni JF Jr. Occupation and lung cancer in Shanghai: a case-control study
. Br J Ind Med 1988; 45: 450–458.
12. Levin LI, Gao Y-T, Blot WJ, Zheng W, Fraumeni JF Jr. Decreased risk of lung cancer in the cotton textile industry of Shanghai. Cancer Res 1987; 47: 5777–5781.
13. Breslow NE, Day NE. Statistical Methods in Cancer Research. The Analysis of Case-Control Studies. IARC Scientific Pub. No. 32. vol. 1. Lyon: International Agency for Research on Cancer, 1980; 192–246.
14. Hastie TJ, Tibshirani RJ. Generalized Additive Models. New York: Chapman and Hall, 1993.
15. Goldbohm RA, Hertog MG, Brants HA, van Poppel G, van den Brandt PA. Consumption of black tea and cancer risk: a prospective cohort study. J Natl Cancer Inst 1996; 88: 93–100.
16. Heibrun LK, Nomura A, Stemmermann GN. Black tea consumption and cancer risk: a prospective study. Br J Cancer 1986; 54: 677–683.
17. Koo LC. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer 1988; 11: 155–172.
18. Zheng W, Doyle TJ, Kushi HL, Sellers TA, Hong CP, Folsom AR. Tea consumption and cancer incidence in a prospective cohort study of postmenopausal women. Am J Epidemiol 1996; 144: 175–182.
19. Mettlin C. Milk drinking, other beverage habits, and lung cancer risk. Int J Cancer 1989; 43: 608–612.
20. Kinlen LJ, Willows ASN, Goldblatt P, Yudkin J. Tea consumption and cancer. Br J Cancer 1988; 58: 397–401.
21. Tewes FJ, Koo LC, Meisgen TJ, Rylander R. Lung cancer risk and mutagenicity of tea. Environ Res 1990; 52: 23–33.
22. Ohno Y, Wakai K, Genka K, Ohmine K, Kawamura T, Tamakoshi A, Aoki R, Senda M, Hayashi Y, Nagao K. Tea consumption and lung cancer risk: a case-control study
in Okinawa, Japan. Jpn J Cancer Res 1995; 86: 1027–1034.
23. Mendilaharsu M, De Stefani E, Deneo-Pellegrini H, Carzoglio JC, Ronco A. Consumption of tea and coffee and the risk of lung cancer in cigarette-smoking
men: a case-control study
in Uruguay. Lung Cancer 1998; 19: 101–107.
24. Brenner H, Blettner M. Controlling for continuous confounders in epidemiologic research. Epidemiology 1997; 8: 429–434.
25. Wu RR, Lin YP, Chen HY. Effect of Fujian oolong tea, jasmine tea, green tea and tea standing overnight on urethan induced lung neoplasia in mice. Presented at the International Tea-Quality–Human Health Symposium (China), 1987;118–119.
26. Ahmad J, Mukhtar H. Green tea polyphenols and cancer: biologic mechanisms and practical implications. Nutr Rev 1999; 57: 78–83.
27. Schwarz B, Bischof HP, Kunze M. Coffee, tea, and lifestyle. Prev Med 1994; 23: 377–384.