The Impact of Green Tea and Coffee Consumption on the Reduced Risk of Stroke Incidence in Japanese Population
The Japan Public Health Center-Based Study Cohort
Background and Purpose—Few prospective studies have examined the impact of both green tea and coffee consumption on strokes. We investigated the association of the combination of those consumption with stroke incidence in a general population.
Methods—We studied 82 369 Japanese (aged 45–74 years; without cardiovascular disease [CVD] or cancer in 1995 and 1998 for Cohort I and II, respectively) who received 13 years of mean follow-up through the end of 2007. Green tea and coffee consumption was assessed by self-administered food frequency questionnaire at baseline.
Results—In the 1 066 718 person-years of follow-up, we documented the incidence of strokes (n=3425) and coronary heart disease (n=910). Compared with seldom drinking green tea, the multivariable-adjusted hazard ratios (95% confidence intervals) of all strokes were 0.86 (0.78–0.95) and 0.80 (0.73–0.89) in green tea 2 to 3 and ≥4 cups/d, respectively. Higher green tea consumption was associated with inverse risks of CVD and strokes subtypes. Compared with seldom drinking coffee, the multivariable-adjusted hazard ratios (95% confidence intervals) of all strokes were 0.89 (0.80–0.99), 0.80 (0.72–0.90), and 0.81 (0.72–0.91) for coffee 3 to 6 times/week and 1 and ≥2 times/day, respectively. Coffee consumption was associated with an inverse risk of CVD and cerebral infarction. Higher green tea or coffee consumption reduced the risks of CVD and stroke subtypes (especially in intracerebral hemorrhage, P for interaction between green tea and coffee=0.04). None of the significant association was observed in coronary heart disease.
Conclusions—Higher green tea and coffee consumption were inversely associated with risk of CVD and stroke in general population.
Tea is one of the world’s most popular beverages. Among the various kinds of teas, green tea has been extensively studied for its antioxidant activities and potential for reducing lifestyle-related diseases.1,2 However, the evidence of cohort study is limited. Prospective cohort studies have shown that green tea consumption is associated with reductions in all-cause and cardiovascular disease (CVD) mortalities.3,4 Only 1 has examined the inverse association between green tea consumption and incident of stroke.5
Furthermore, coffee, which is also the most popular beverage in the world, has recently come into the limelight because of its association with reduced diabetes mellitus6 and cancer.7 However, prospective studies on the association between coffee and stroke8,9 have been inconsistent. These inconsistencies may be partly because of the different end points, types of participants, and small sample sizes. However, no large prospective study has examined the associations of green tea and coffee consumption with the incidence of stroke simultaneously. We examined the hypothesis that both green tea and coffee consumption would be independently preventative for stroke in the Japanese general population.
Subjects and Methods
Study Design and Samples
This is an ongoing cohort study focusing on CVD and cancer.10 The study design was described in detail elsewhere.11 Briefly, the age distributions at the time of entry were 40 to 59 for Cohort I (started in 1990) and 40 to 69 for Cohort II (started in 1993). Participants were identified using the population registry in each city or town without overlap (65 803 men and 67 520 women).
Baseline Data Collection
We assessed the dietary habits using a food-frequency questionnaire in 1995 and 1998 for Cohorts I and II, respectively. The questionnaires were returned by 47 400 men (72%) and 53 538 women (79%). We estimated the dietary intake of each individual based on a 138-food-item questionnaire, which was previously validated for estimating various nutrients and food groups.12 The questionnaire also collected data on demographics, lifestyle factors, height, weight, smoking, alcohol consumption, and physical activity. Cohort members were excluded from the analyses, if they reported CVD or cancer in the questionnaires (n=5061), lost to follow-up and moved out of the area before the baseline survey (n=1327), or incompletely answered the food-frequency questionnaire (n=12 572). After applying these exclusions, 38 029 men and 43 949 women in total were included in the analyses. This study was approved by the institutional review board of the National Cancer Center, Tokyo. Each participant provided informed consent on completion of the baseline questionnaire, which described the study purposes and follow-up.
The frequency response choices for each food item were as follows: never; 1 to 3 times/month; 1 to 2, 3 to 4, and 5 to 6 times/week; 1, 2 to 3, 4 to 6, and ≥7 times/day. A standard portion was shown for each food item. The relative portion sizes were as follows: small (≈50% smaller than the standard), standard, and large (≈50% larger than the standard) size.13 Daily intakes of individual nutrients were calculated using the food composition table developed for each questionnaire based on the fifth revised edition of the Standard Tables of Food Composition in Japan.14
Green tea and coffee consumption were obtained from the frequencies and amount of each beverage consumed using the choice of 0, 1 to 2, 3 to 4, and 5 to 6 times/week, and almost daily (further divided into 1, 2–3, 4–6, 7–9, and ≥10 cups/day). For the present analysis, we further grouped these categories based on their distribution among the subjects: green tea consumption for 0, 1 to 2, 3 to 6 times/week, 1, 2 to 3, and ≥4 cups/day; and coffee consumption for 0, 1 to 2, 3 to 6 times/week, 1 and ≥2 cups/day. The rank correlation coefficients for green tea and coffee consumption between the questionnaire and dietary record data were 0.37 and 0.59 for men and 0.43 and 0.51 for women, respectively.15 We did not collect the type of coffee (decaffeinated or caffeinated), because decaffeinated coffee is not commonly consumed in Japan.
Confirmation of Stroke and Coronary Heart Disease
In the 9 Public Health Center areas, totally 54 major hospitals were capable of computed tomopgraphy (CT) scanning and/or magnetic resonance imaging.10 Medical records were reviewed by hospital workers, physicians, or researchers who were blinded to the baseline data. Incidences of CVD were registered during the follow-up period. To complete surveillance for fatal CVD, we also conducted a systematic search for death certificates. We obtained information on the underlying cause of death by checking against death certificate files with permission to confirm mortality from CVD according to the International Classification of Death, 10th Revision: I00-I99. All cases of CVD based on death certificates only had been registered as death certificates only cases.
Strokes were confirmed according to the National Survey of Stroke criteria. These criteria require the rapid onset of a constellation of neurological deficits lasting at least 24 hours (or until death). For each stroke subtype (cerebral infarction [CI; thrombotic or embolic stroke], intracerebral hemorrhage [ICH], and subarachnoid hemorrhage), a definite diagnosis was established based on the examination of CT scans, magnetic resonance imaging, or autopsy findings.16 Coronary heart disease (CHD) indicated in the medical records was confirmed according to the criteria of the MONICA project, which requires chest pain, electrocardiographic evidence, cardiac enzyme abnormalities, and autopsy findings.17 In the absence of diagnosis of myocardial infarction, deaths that occurred within 1hour from onset of event were regarded as sudden cardiac deaths. In this study, CVD was defined as stroke or CHD.
For each subject, we calculated person-years of follow-up from the baseline to whichever of the following came first: the first end point, death, emigration, or December 31, 2007. Changes in residential status were identified through the residential registry in each area. Subjects who moved from their original residence (2% of the total participants) were censored at that time.
The Cox proportional hazards ratios and 95% confidence intervals were fitted to the categorized consumption (the reference group is the never consumption of green tea or coffee), after adjusting for age, sex, and other potential confounding factors: smoking status (never, ex-smoker, or current smoker of 1–19 or ≥20 cigarettes/day); alcohol intake (nondrinkers, occasional drinkers [1–3 days/month], weekly ethanol intake of 1–149, 150–299, 300–449, or ≥450 g/week); body mass index (in quintiles); history of diabetes mellitus (yes/no); medication use for hypertension or hypercholesterolemia (yes/no); quintiles of energy-adjusted dietary intakes of fruits, vegetables, and fish; leisure time spent engaged in exercise (<1, 1 to 3 days/month, or ≥1 day/week); public health center; and coffee or green tea consumption. Trend tests were conducted by assignment for green tea and coffee consumption to test the significance of these variables. Green tea and coffee consumption interactions were also analyzed (see the Statistical Analysis in the online-only Data Supplement). All statistical analyses were conducted using the SAS statistical package (version 8.2, SAS Institute Inc, Cary, NC).
Table 1 shows the baseline characteristics according to green tea and coffee consumption categories. Higher frequencies of green tea consumption tended to have a higher prevalence of exercise. Higher frequencies of coffee consumption tended to be younger, had a higher prevalence of smoking and exercise, and had a lower prevalence of antihypertensive drug users and history of diabetes mellitus.
During a follow-up period (13-year of average), we documented 3425 strokes (1964 CI, 1001 ICH, and 460 subarachnoid hemorrhages) and 910 CHD events (489 definite myocardial infarctions and 28 sudden cardiac deaths). In total, 4335 CVD events were documented.
We observed inverse associations between green tea consumption and the incidences of CVD, strokes, and its subtypes (Table 2). Age- and sex-adjusted hazard ratios of CI for green tea consumption of 2 to 3 and ≥4 cups/day were inversely significant. After further adjustment, the significance was inversely associated with CI for green tea consumption of ≥4 cups/day.
We observed an inverse association between coffee consumption and the incidences of CVD, all strokes, and CI (Table 3). We observed inverse associations between coffee consumption (≥3 times/week) and the incidences of ICH by age- and sex-adjusted. However, after further adjustment, the association was attenuated.
The Figure shows the multivariable hazard ratios of CVD and stroke subtypes, according to combination of green tea and coffee consumption frequencies. Compared with seldom drinking green tea or coffee, higher green tea (≥2 times/day) or coffee (≥1 time/day) consumption reduced the risks of CVD, all strokes, CI, and ICH (In particular, P for interaction between green tea and coffee consumption was 0.04 in ICH.) The results were similar among men and women (data not shown). No significant association was observed in CHD (data not shown).
In this study, higher green tea and coffee consumption were found to be inversely associated with the incidences of CVD and stroke subtypes. In addition, higher green tea or coffee consumption reduced the risks of CVD, strokes, and its subtypes. To the best of our knowledge, this is the first study on the association between the combination of these 2 popular beverages and the incidences of CVD and stroke subtypes, especially in ICH independently.
Green tea consumption has been shown to reduce the risk of mortality because of all-cause and CVD.3 Compared with green tea <1 cup/day, ≥5 cups/day had 15% and 26% reduced risk of all-cause and CVD mortality, respectively. Green tea consumption has also been associated with a reduced risk of incidence of all strokes, CI, and ICH.5 There has been no prospective study showing evidence that green tea consumption is preventive of CHD, which is consistent with our study.
Green tea content catechins, especially (–)-epigallocatechin-3-gallate, exerts vascular-protective effects through multiple mechanisms, including antioxidative,18 anti-inflammatory,19 antiproliferative,20 increasing the plasma antioxidant capacity,21 and antithrombogenic effects.22 Moderate consumers of green tea are less likely to develop hypertension.23 However, a meta-analysis of randomized controlled trials has shown that green tea is not associated with blood pressure, partly because of small studies number (n=2).24 Further studies are waited for.
The association between coffee consumption and the risk of stroke has been controversial. No association between coffee consumption and fatal and nonfatal stroke has been found in healthy populations.25,26 However, coffee consumption has been associated with an increased risk of CI in hypertensive men.27 In a meta-analysis of prospective studies,28 moderate consumption of coffee showed a weak inverse, but heavy coffee consumption (≥7 cups/day) was not significantly associated with stroke risk. Recently, in a cohort of Swedish women (coffee >1 cups/day), risk of stroke was reduced in CI, but not in ICH, compared with <1 cup/day.8
In recent meta-analyses, the association between coffee consumption and the risk of CHD still remains controversial.29,30 Our study showed no significant association between coffee consumption and CHD. Coffee intake (≥2 cups/day) was positively associated with CHD in age-adjusted analysis. However, after further adjustment, including smoking, the positive association was disappeared, which was similar to a cohort study in United States.31 The finding of the present study might suggest that the unfavorable effects of smoking overcome the beneficial effects of coffee consumption.
Coffee contains several biologically active substances, such as caffeine and diterpene, which were inconsistently associated with serum cholesterol and blood pressure levels, and decrease insulin sensitivity.32–34 Coffee also contains chlorogenic acid and quinides, which may reduce body weight and blood glucose tolerance.35 In a prospective study, higher coffee consumption may reduce the incidence of diabetes mellitus.36 In our study, higher coffee consumption had a lower prevalence of diabetes mellitus. Elevated blood glucose is a risk of incident CI.37 Therefore, coffee consumption may reduce the risk of CI by improving glucose levels.
Our study showed that combination of higher green tea and coffee consumptions contributed to the reduced risk of stroke as an interaction effect for each other. Protective effects may be related to weak but strength the different antioxidant and other biological contents in these 2 beverages, but the underlying combinational mechanism is still not clear.
The present study has certain methodological strengths compared with previous investigations. First, we evaluated a large cohort enrolled from the Japanese general population, and the higher response rate to the questionnaire (>80%) and the smaller loss to follow-up (0.4%) were acceptable for study settings. Second, incidence is a more direct measure of CVD risk than death because treatment influences CVD death. Third, we estimated green tea and coffee consumption using a validated questionnaire.
Our study has several limitations. First, data regarding present illnesses, green tea, and coffee consumption were self-reported, raising the problem of potential misclassification. However, our self-reported data may be reasonably accurate because nationwide annual health screenings, conducted since 1992 in Japan, produced similar results. Our validation study showed relatively high validity.38 Second, measurement errors concerning nutrient intake are inevitable when using a food-frequency questionnaire. Changes in green tea and coffee consumption during the follow-up period may also have caused misclassification. In the current study, the reproducibility of green tea and coffee consumption estimates suggests that any overestimates or underestimates are likely to cancel each other out and result in an accurate overall estimate. We have no decaffeinated coffee in our cohort study. The Japanese do not drink the decaffeinated coffee. In the previous article, similar associations were observed whether participants drank predominantly caffeinated or decaffeinated coffee.39
Higher green tea or coffee consumption was beneficial for reducing risks of CVD, all strokes, CI, and ICH, as a preventive medical point of view.
The authors thank all of the staff members in each study area and in the central office for their painstaking efforts in conducting the baseline survey and follow-up research.
Sources of Funding
This study was supported by Grants-in-Aid for Cancer Research and the Third-Term Comprehensive Ten-Year Strategy for Cancer Control from the Ministry of Health, Labor and Welfare of Japan.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.677500/-/DC1.
- Received September 20, 2012.
- Accepted January 8, 2013.
- © 2013 American Heart Association, Inc.
- Kuriyama S
- Tanabe N,
- Suzuki H,
- Aizawa Y,
- Seki N
- Larsson SC,
- Virtamo J,
- Wolk A
- Kokubo Y,
- Iso H,
- Ishihara J,
- Okada K,
- Inoue M,
- Tsugane S
- 14.↵Standard Tables of Food Composition in Japan, the fifth revised and enlarged edition. Tokyo: Printing Bureau, Ministry of Finance, Tokyo;2005.
- Inoue M,
- Kurahashi N,
- Iwasaki M,
- Shimazu T,
- Tanaka Y,
- Mizokami M,
- et al
- Tunstall-Pedoe H,
- Kuulasmaa K,
- Amouyel P,
- Arveiler D,
- Rajakangas AM,
- Pajak A
- Hofmann CS,
- Sonenshein GE
- Hooper L,
- Kroon PA,
- Rimm EB,
- Cohn JS,
- Harvey I,
- Le Cornu KA,
- et al
- Lopez-Garcia E,
- Rodriguez-Artalejo F,
- Rexrode KM,
- Logroscino G,
- Hu FB,
- van Dam RM
- Larsson SC,
- Orsini N
- Kawachi I,
- Colditz GA,
- Stone CB
- Lopez-Garcia E,
- van Dam RM,
- Willett WC,
- Rimm EB,
- Manson JE,
- Stampfer MJ,
- et al
- Greenberg JA,
- Boozer CN,
- Geliebter A