Consumption of green and roasted teas and the risk of stroke incidence: results from the Tokamachi–Nakasato cohort study in Japan
1 Division of Health Promotion, Niigata University Graduate School of Medical and Dental Sciences, Japan.
2 Division of Public Health, Niigata University Graduate School of Medical and Dental Sciences, Japan.
3 Division of Cardiology, Niigata University Graduate School of Medical and Dental Sciences, Japan.
4 School of Health Sciences, Faculty of Medicine, Niigata University, Japan.
* Corresponding author. Naohito Tanabe, MD, PhD, Division of Health Promotion, Niigata University Graduate School of Medical and Dental Sciences, 1-757 Asahimachi-dori, Chuo Ward, Niigata 951-8510, Japan. E-mail: tanabe{at}med.niigata-u.ac.jp
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Background and purpose Green tea consumption is inversely associated with death from stroke. The purpose of the present study was to assess whether it is inversely associated with subsequent stroke incidence and whether this association is preserved even with roasted tea leaves.
Methods In 1998, 6358 Japanese adults (2087 men and 4271 women) aged 40–89 years without a history of stroke or heart disease completed a lifestyle questionnaire, including consumption of green tea or roasted tea. By the end of 2003, 110 stroke events (59 cerebral infarction events, 34 cerebral haemorrhage events, 15 subarachnoidal haemorrhage events and two stroke events of unspecified subtype) had been documented. Cox proportional hazards regression analysis was used to calculate the multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (95% CIs) for total stroke events, cerebral infarction events and cerebral haemorrhage events according to consumption categories of green tea and roasted tea.
Results A considerably lower risk was observed for total stroke incidence in both the middle (multivariable HR, 0.43; 95% CI, 0.25–0.74; P = 0.002) and the high (multivariable HR, 0.41; 95% CI, 0.24–0.70; P = 0.001) categories of green tea consumption. This inverse association was consistent even when cerebral infarction and cerebral haemorrhage were analysed separately. The consumption of roasted tea was not associated with stroke risk.
Conclusions Green tea consumption is associated with a reduced risk of total stroke incidence, cerebral infarction and cerebral haemorrhage.
Keywords Tea, stroke, incidence, cerebral infarction, cerebral haemorrhage, cohort studies, epidemiology
Accepted 12 May 2008
Interest about the preventive effects of green tea consumption on stroke is growing. Experimental studies have investigated the preventive effects of green tea extract on cardiovascular disease through its antioxidant activity.1–3 In humans, Kuriyama et al.4 recently reported an inverse association between the consumption of green tea and death from cardiovascular disease and cerebral infarction. Historically, a high consumption of green tea has been associated with a lower mortality from total stroke,5 cerebral haemorrhage5 and total cardiovascular disease.6 However, whether green tea prevents the incidence of these events or inhibits their progression in humans is unclear because only fatal cases were evaluated in these epidemiological reports. Furthermore, the effects of other types of tea have not been fully investigated.
The Tokamachi–Nakasato cohort study was started in 1998 to prospectively analyse the association between lifestyle factors and cardiovascular disease incidence. At the baseline assessment, we evaluated tea consumption for green tea and roasted tea. Green tea and roasted tea are the most commonly consumed teas in Japan; therefore, these are sometimes called Japanese tea as in the present study. We designed a prospective analysis to evaluate the risk of stroke incidence according to tea consumption. The purpose of the present study was to assess whether green tea consumption is inversely associated with subsequent stroke incidence and whether this association is preserved even with roasted tea leaves.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Study subjects
The study areas were Tokamachi city and its neighbouring village, Nakasato, Niigata prefecture, Japan. The total population of the entire areas was 49 419 and those aged 40–89 years were 29 323 in the 2000 census. For the present study, we distributed a self-administered questionnaire to the 8161 inhabitants, aged 40–89 years, who participated in the official population-based health check-up programme from June to September of 1998. This programme was managed by local governments of the city and the village in accordance with the Law of the Health and Medical Services for the Aged, mainly for inhabitants aged 40 years or over who wished to check their general health status. Thus, the study subjects were non-random samples of the general population, and consisted of relatively motivated subjects for maintaining their own health. Of those participants, 7766 responded to the questionnaire and 7753 agreed to participate in the study. They were the baseline cohort of the present study and followed for the incidence of stroke until the end of 2003 (Figure 1). Of the entire cohort, 199 subjects already had a history of cerebrovascular disease at baseline and were excluded from the study. Of the remaining subjects, 353 were lost for follow-up before the evaluation of their first stroke event. We also excluded 394 subjects with missing data for green tea and roasted tea consumption, as well as 425 subjects with a history of heart disease at baseline. Those with missing data on body mass index, blood pressure, total cholesterol or grain product consumption were also excluded (n = 20) because these were treated as continuous variables in the statistical analysis for the adjustment of the confounding effects. Furthermore, four subjects with missing data on vegetable consumption were excluded because statistical analysis could not be completed if they were included by creation of a dummy variable. Finally, 6358 subjects were selected for the present study. However, the number of subjects evaluated varies by type of tea because not all subjects answered for both types of tea. The number of subjects and total follow-up periods for the first stroke event were 6207 and 31 070 person-years for green tea and 5337 and 27 861 person-years for roasted tea.
|
Written informed consent was obtained from all participants and the study protocol was reviewed and approved by the ethics committee of Niigata University School of Medicine.
Exposure data assessment
Japanese tea consumption was evaluated by one self-administrated questionnaire each for green tea and roasted tea in the following six categories: (i) >10 cups every day; (i) 5–10 cups every day; (iii) one to four cups every day; (iv) several cups every 2–3 days; (v) several cups a week; and (vi) never or occasionally. Although consumption was evaluated in these six categories, groups of two categories were combined for statistical analysis because the number of respondents was insufficient in some categories. Accordingly, the categories for the green and roasted tea analyses were consolidated to: (a) five cups every day or more (High); (b) more than several cups every 2–3 days but <5 cups every day (Middle); and (c) several cups a week or less (Low). The usual amount of green tea is about 70–90 ml per cup and that of roasted tea is about 100–130 ml per cup. Although the usual amount of tea extract is different between these types of tea, the amount of tea leaves for brewing is 2–3 g per person for both types. Tea can be brewed a few times from the same tea leaves.
Possible confounders, including a history of hypertension, diabetes, hyperlipidaemia, the presence of a medically treated disease, a parental history of stroke, smoking, alcohol drinking, frequency of fruit intake, balance of meat and fish intake, the average amount of time spent walking each day, consumption of black or oolong tea and coffee consumption were also evaluated through a self-administered questionnaire. Black tea and oolong tea were simultaneously assessed as an index of non-Japanese tea consumption. Personal and parental histories were categorized as positive when a participant declared them in the questionnaire. Otherwise, he/she was not considered to have them. The presence of a medically treated disease was categorized as positive when a subject had received chronic outpatient treatment for any disease. Alcohol drinking was assigned to one of three categories: never, sometimes and almost every day. Subjects who drink alcoholic beverages ‘almost every day’ were considered regular alcohol drinkers. Consumption of black or oolong tea and coffee were evaluated in the same six categories as Japanese tea consumption. For black or oolong tea, the top five categories were combined because the number of respondents barely reached 28.5% even when combined, to give two categories: (i) several cups a week or more and (ii) never or occasionally. Coffee consumption was also summarized in the following two categories: (i) 1–4 cups per day or more and (ii) several cups every 2–3 days or less.
Daily intake of grain products, vegetables, salted vegetables and soybean paste soup (misoshiru in Japanese), and weekly intake of milk were assessed by nutritionists using food models as references. Daily intake of grain products were assessed as total intake of cooked rice, bread and noodles and summarized in units; 1 unit of grain products is equivalent to 80 kcal. Soybean paste soup consumption is assessed in bowls (one bowl generally serves about 150 ml of soup) and milk consumption is translated into bottles (one bottle means 200 ml of milk). Vegetable and salted vegetable consumptions were assessed in three categories as described in Table 2. Body mass index was calculated from height and weight measured at the baseline assessment. Blood pressure was measured using a manual sphygmomanometer. Total cholesterol was measured from serum samples obtained at baseline assessment.
|
End-point determination
Diagnosis of stroke was determined by medical chart review. Candidates for the review were identified through death certificate reviewing or a mailed questionnaire survey. A single physician reviewed all medical charts. Stroke subtype was determined by computed tomography or magnetic resonance imaging. Three fatal cases that died before imaging were diagnosed as subarachnoid haemorrhage by lumbar puncture at necropsy. Subtypes of two stroke events could not be confirmed because imaging testing was not performed; they were included only in the risk evaluation for total stroke. Thus, 106 stroke cases were identified. Because four cases experienced two different types of stroke, 110 stroke events (59 cerebral infarction events, 34 cerebral haemorrhage events, 15 subarachnoidal haemorrhage events and two stroke events of unspecified subtype) were registered as end-points.
Statistical analysis
The age- and sex-adjusted statistical difference of baseline characteristics according to green tea and roasted tea consumption levels were tested by multinomial logistic regression analysis using a log-likelihood chi-square statistic. Items related to tea consumptions with a P-value of <0.1 were selected for further adjustment. Multinomial logistic regression analysis was also applied in the calculation of propensity scores7 using selected confounders as independent variables and tea consumption as an dependent variable, with the ‘Low’ category as a reference.
Cox proportional hazards regression analysis was used to calculate the hazard ratios (HRs) and 95% confidence intervals (CIs) for total stroke events, cerebral infarction events and cerebral haemorrhage events according to consumption categories of different types of Japanese tea. Age- and sex-adjusted HRs were calculated in analyses for both sexes; age-adjusted HRs were calculated separately for men and women. The multivariable adjusted HRs were also calculated using propensity scores.
Statistical analyses were performed using the SPSS 13.0 software package (SPSS Inc, Chicago, IL).
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
The study subjects more frequently consumed green tea than roasted tea in both sexes (Table 1). According to age- and sex-adjusted analysis, values for age, sex, personal history of hypertension, personal history of diabetes, presence of a medically treated disease, body mass index, systolic blood pressure, serum total cholesterol, smoking, daily intake of grain products, vegetable intake, salted vegetable intake, fruit intake, balance of meat and fish intake, daily intake of soybean paste soup, weekly intake of milk, amount of time spent walking each day, consumption of another type of Japanese tea, consumption of black or oolong tea and coffee consumption associated with green or roasted tea consumption with a P-value of <0.1 (Table 2) and were selected to calculate propensity scores.
|
Multivariable analysis (Table 3) showed that green tea consumption was positively associated with age (P < 0.001), systolic blood pressure (P = 0.077), daily intake of grain products (P < 0.001), vegetable intake (P < 0.001), salted vegetable intake (P < 0.001), fruit intake (P < 0.001), balanced or fish-dominant intake rather than meat-dominant intake (P = 0.006), daily intake of soybean paste soup (P = 0.002), amount of time spent walking each day (P = 0.047), consumption of black or oolong tea (P = 0.005); but negatively associated with smoking habit (P = 0.001), consumption of another type of Japanese tea (P < 0.001) and coffee consumption (P < 0.001). Roasted tea was positively associated with age (P < 0.001), fruit intake (P < 0.001), balanced or fish-dominant intake rather than meat-dominant intake (P = 0.006), amount of time spent walking each day (P = 0.093), consumption of black or oolong tea (P < 0.001); but negatively associated with smoking (P = 0.006), consumption of another type of Japanese tea (P < 0.001) and coffee consumption (P < 0.001).
For green tea consumption, the highest incidence of total stroke was observed in the ‘Low’ consumption category for both sexes (Table 4). Compared to this category, both the ‘Middle’ and the ‘High’ categories showed considerably lower HRs every in age- and sex-adjusted analysis and in multivariable analysis. The multivariable-adjusted HR was 0.43 (95% CI, 0.25–0.74; P = 0.002) for the ‘Middle’ category and 0.41 (95% CI, 0.24–0.70; P = 0.001) for the ‘High’ category. Decreased HRs for these two categories were also observed both for men and women, although these were slightly higher for women.
|
|
These decreased HRs in the ‘Middle’ and the ‘High’ categories were also observed, even when analysed for cerebral infarction and cerebral haemorrhage separately.
No considerable association was found between roasted tea and the risk of stroke. Although multivariable-adjusted HR for total stroke seemed to be low in the ‘Middle’ and ‘High’ categories of roasted tea, these were inconsistent for men and women, as well as for cerebral infarction and cerebral haemorrhage.
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
This population-based, prospective cohort study clearly demonstrates the inverse association between the consumption of green tea and the risk of stroke. Consuming more than several cups of green tea every 2–3 days was associated with a risk reduction of more than 50% of total stroke, cerebral infarction and cerebral haemorrhage events.
The risk-reduction rates seem to be larger than those observed in a recent large-scale cohort study.4 This difference could partly be explained by the different methods of stroke diagnosis. In that study, diagnosis was based on a review of death certificates, whereas here it was based on a review of medical charts, which should be more accurate. Confusion on the consumption of roasted tea with that of green tea could also have caused underestimation of the effect of green tea in the previous study. Some subjects could have included roasted tea consumption when answering questions about green tea consumption at the baseline assessment because the previous study did not distinguish roasted tea (that made by roasting green tea leaves) from non-roasted green tea.
We did not find a dose-dependent association between green tea consumption and stroke incidence. The lack of the dose-dependency could partly be explained by the fact that subjects in the ‘High’ category and 
90% of the subjects in the ‘Middle’ category consumed green tea everyday. Although the daily amount of green tea was different between these two categories, daily consumption of tea leaves might not have been so different. High consumers might prefer thinner green tea extract or brew green tea more times from the same tea leaves to save them. It is very usual for Japanese people to brew a few or several times from the same tea leaves. Although epidemiological studies have treated the amount of the consumed green tea extract as exposure data,4–6 daily or weekly amounts of tea leaves used for brewing should also be considered in the future studies if possible.
In addition, we do not have any data validating the levels of tea consumption assessed in the present study. We could not distinguish the consumption of tea extract itself within habitual consumers of green tea. However, we think that we can assess stroke risk reduction in habitual consumers of green tea using hazard ratios for the ‘Middle’ and ‘High’ categories because the majority of subjects in the ‘Middle’ and all subjects in the ‘High’ categories stated that they consumed green tea every day, whereas more than half of subjects in the ‘Low’ category stated that they never or occasionally did so.
Another explanation could be the possible presence of the selection bias. Low consumption of green tea might be an index for high-risk stroke subjects. Shimazu et al.8 recently reported that green tea is part of a traditional Japanese dietary pattern that is associated with reduced CVD mortality. In that paper, the ‘Japanese pattern’ was positively associated with consumptions of soybean products, fish, seaweeds, vegetables, fruits and green tea. In the present study, green tea consumption also positively associated with consumptions of soybean paste soup, fish, vegetables and fruits. Although adjustment of these factors did not alter the relationship between green tea consumption and stroke incidence in the present study, adjustment of the ‘Japanese pattern’ might have been insufficient. In order to solve these problems, randomized-controlled trials would be necessary.
Tea polyphenols such as flavonoids can play a major role in preventing cardiovascular disease through antioxidant activities.1–3 Polyphenols also prevent systemic atherosclerosis and thrombosis by lowering platelet aggregation activity, and modulate nitric oxide generation from the vascular endothelium.3
In contrast with cerebral infarction, no sufficient biological mechanisms have been proposed for cerebral haemorrhage. Experimental studies of stroke-prone rats have shown that green tea protects against hypertension,9,10 the most potent risk factor both for cerebral haemorrhage and infarction. A human observational study in Taiwan also demonstrated that the habitual drinking of green or oolong tea was associated with a reduced risk of developing hypertension.11 However, a recent meta-analysis of randomized studies concluded that green tea did not affect blood pressure.12 In addition, systolic blood pressure was positively associated with green tea consumption in the present study subjects. Furthermore, the relationship between green tea consumption and the risk of cerebral haemorrhage was inconsistent in previous studies.4,5 Further studies are required to determine the effect of green tea consumption on cerebral haemorrhage risk.
The present study may be the first report to evaluate the association between roasted tea consumption and the risk of stroke. The consumption, however, did not associate with the stroke risk. According to an in vitro study, roasted tea extract contains pro-oxidants.13 Thus, the antioxidant effects of green tea polyphenols could be counteracted by the pro-oxidant activity generated by the roasting, causing a lack of protection against stroke.
We could not analyse the effect of black or oolong tea consumption on stroke incidence because of the limited number of the habitual consumers. A cohort study from the Netherlands showed a lower hazard ratio for stroke incidence among daily high consumers of black tea,14 but a study from the United States did not.15 There has been no sufficient epidemiological study about oolong tea. The effects of such non-Japanese teas should also be analysed in future studies.
In conclusion, ‘Middle’ to ‘High’ consumption of green tea is associated with a reduced risk of total stroke incidence and cerebral infarction, and probably with cerebral haemorrhage, whereas roasted tea consumption is not. Further studies are required to verify the causal relationship between green tea consumption and reduced risk of stroke incidence.
|
Funding |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
Japan Society for the Promotion of Science—the Grants-in-Aid for Scientific Research in fiscal years 2003 to 2006 (No. 15390195); Niigata Prefecture, Japan in fiscal year 1998; Niigata Association for Comprehensive Health Promotion and Research in fiscal year 2004.
KEY MESSAGES
|
|
Acknowledgement |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
We are grateful for the cooperation of Tokamachi city and Nakasato village in the collection of data.
Conflict of interest: None declared.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionFundingAcknowledgementReferences |
|---|
1 Cooper R, Morré DJ, Morré DM. Medicinal benefits of green tea: Part I. Review of noncancer health benefits. J Altern Complement Med (2005) 11:521–28.[CrossRef][Web of Science][Medline]
2 Zaveri NT. Green tea and its polyphenolic catechins: Medicinal uses in cancer and noncancer applications. Life Sci (2006) 78:2073–80.[CrossRef][Web of Science][Medline]
3 Curin Y, Andriantsitohaina R. Polyphenols as potential therapeutical agents against cardiovascular diseases. Pharmacol Rep (2005) 57(Suppl):97–107.[Medline]
4 Kuriyama S, Shimazu T, Ohmori K, et al. Green tea consumption and mortality due to cardiovascular disease, cancer, and all causes in Japan: the Ohsaki study. JAMA (2006) 296:1255–65.
5 Sato Y, Nakatsuka H, Watanabe T, et al. Possible contribution of green tea drinking habits to the prevention of stroke. Tohoku J Exp Med (1989) 157:337–43.[Web of Science][Medline]
6 Nakachi K, Matsuyama S, Miyake S, Suganuma M, Imai K. Preventive effects of drinking green tea on cancer and cardiovascular disease: epidemiological evidence for multiple targeting prevention. Biofactors (2000) 13:49–54.[Web of Science][Medline]
7 D'Agostino RB Jr. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med (1998) 17:2265–81.[CrossRef][Web of Science][Medline]
8 Shimazu T, Kuriyama S, Hozawa A, et al. Dietary patterns and cardiovascular disease mortality in Japan: a prospective cohort study. Int J Epidemiol (2007) 36:600–09.
9 Ikeda M, Suzuki C, Umegaki K, Saito K, Tabuchi M, Tomita T. Preventive effects of green tea catechins on spontaneous stroke in rats. Med Sci Monit (2007) 13:BR40–45.[Web of Science][Medline]
10 Potenza MA, Marasciulo FL, Tarquinio M, et al. EGCG, a green tea polyphenol, improves endothelial function and insulin sensitivity, reduces blood pressure, and protects against myocardial I/R injury in SHR. Am J Physiol Endocrinol Metab (2007) 292:E1378–87.
11 Yang YC, Lu FH, Wu JS, Wu CH, Chang CJ. The protective effect of habitual tea consumption on hypertension. Arch Intern Med (2004) 164:1534–40.
12 Taubert D, Roesen R, Schömig E. Effect of cocoa and tea intake on blood pressure: a meta-analysis. Arch Intern Med (2007) 167:626–34.
13 Yanagimoto K, Ochi H, Lee KG, Shibamoto T. Antioxidative activities of volatile extracts from green tea, oolong tea, and black tea. J Agric Food Chem (2003) 51:7396–401.[CrossRef][Web of Science][Medline]
14 Keli SO, Hertog MG, Feskens EJ, Kromhout D. Dietary flavonoids, antioxidant vitamins, and incidence of stroke: the Zutphen study. Arch Intern Med (1996) 156:637–42.
15 Sesso HD, Paffenbarger RS Jr, Oguma Y, Lee IM. Lack of association between tea and cardiovascular disease in college alumni. Int J Epidemiol (2003) 32:527–33.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  G. D. Smith How do we know, what do we know and what can knowledge do? From John Brownlee to translational medicine Int. J. Epidemiol., October 1, 2008; 37(5): 911 - 913. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||