ABSTRACT
Background
Many cohort studies have explored the relation between tea consumption and stroke risk; however, the conclusions have been inconsistent. In addition, evidence is lacking in China, where the patterns of tea consumption and main types of tea consumed differ substantially from those in high-income countries.
Objective
We aimed to systematically assess the association of tea consumption with the risk of stroke based on a Chinese large-scale cohort study.
Methods
A total of 487,377 participants from the China Kadoorie Biobank were included in the present study. Detailed information about tea consumption (including frequency, duration, amount, and tea type) was self-reported at baseline. After ∼4.3 million person-years of follow-up, 38,727 incident cases of stroke were recorded, mainly through linkage with mortality and morbidity registries and based on the national health insurance system.
Results
Overall, 128,280 adults (26.3%) reported drinking tea almost daily (41.4% men, 15.9% women), predominantly green tea (86.7%). Tea consumption had an inverse and dose–response relation with the risk of stroke (Ptrend < 0.001). Compared with nonconsumers, those who consumed tea occasionally, weekly, and daily had adjusted HRs and 95% CIs of 0.96 (0.94, 0.99), 0.94 (0.90, 0.98), and 0.92 (0.89, 0.95) respectively, with little difference by stroke type. Among those who consumed tea daily, the HRs for stroke decreased with the increasing duration and amount of tea consumed (all P < 0.001). These inverse associations were significant for green tea but not for other types of tea. Among men, but not women, the inverse relations could be detected, and similar inverse associations could be found for male noncurrent alcohol-consumers and noncurrent smokers as well.
Conclusions
Among Chinese adults, higher consumption of tea, especially green tea, was associated with a lower risk of ischemic and hemorrhagic stroke.
Keywords: tea consumption, stroke, risk, association, cohort study, China, CKB
Introduction
Tea is one of the most frequently consumed beverages worldwide (1), with the global consumption amount increasing annually. In China, a key tea-producing and exporting country worldwide, there are many tea consumers (2). Given the widespread consumption of tea, it is important to understand its potential health effects in diverse populations (1, 3). Over the past few decades, exploratory research has reported that tea may influence the risks of numerous diseases, including cardiovascular disease, diabetes mellitus, and cancer (2-4).
In 1989, it was originally proposed that drinking tea might contribute to the prevention of stroke (4). Thereafter, various studies, including experimental and population-based studies, were conducted to explore the relation between drinking tea and stroke risk (5-7). Experimental evidence has shown that polyphenols, which are important constituents of tea, can inhibit the development of atherosclerosis, have anti-inflammatory properties, improve endothelial function, and contribute to protection against stroke (8, 9). Many previous population-based cohort studies have explored the relation between tea consumption and stroke risk (1, 10-14). Nevertheless, these prospective studies have failed to reach a consistent conclusion: some studies confirmed the negative correlation between tea intake and stroke risk (13), but the others did not obtain similar results (14). Relatively small sample sizes, lack of reproducibility of tea consumption data, nonrepresentative study populations, and the absence of important modifying factors might have led to the aforementioned different results.
The existing cohort studies mainly originated from Western and Japanese populations. The commonly consumed tea types (green tea or black tea), habits of tea drinking (drinking tea with or without milk), production regions, and manufacturing technologies of tea (i.e., different technologies for green removal) are very different between populations from various countries (15-17), which might also result in inconsistent associations of tea consumption with stroke risk. As mentioned, most of the related cohort studies focusing on tea consumption and stroke risk were conducted in Western and Japanese populations (1, 10-12). However, in China, a country with one of the highest amounts of tea consumption, evidence related to this association from a cohort study is still lacking. Moreover, the incidence of hemorrhagic stroke is higher in China than in Western countries (18, 19), where most strokes are of the ischemic type.
Therefore, in this study, based on the large prospective cohort of the China Kadoorie Biobank (CKB) with >0.5 million participants, we integrally evaluated the relation between tea consumption and the risk of stroke, including the different types (including ischemic and hemorrhagic stroke), in the Chinese population.
Methods
Study population
The details of the prospective cohort design, procedure, population characteristics, and other information related to the CKB have been reported elsewhere (20, 21). Overall, 515,681 persons participated in the baseline survey between 2004 and 2008. After quality control of the data, a total of 512,891 individuals, from 30 to 79 y old, were recruited from 10 geographically diverse areas in China (Supplemental Figure 1) (20). For this study, we excluded 8,884 individuals with stroke or transient ischemic attack, 15,472 persons with heart disease, and 2,577 persons with cancer at baseline. After the aforementioned exclusions, our final analyses included 487,377 individuals, consisting of 199,294 men and 288,083 women (Supplemental Figure 1).
Before the baseline survey, each participant provided written informed consent. The study was approved by the Ethics Review Committee of the Chinese Center for Disease Control and Prevention (CDC), the Oxford Tropical Research Ethics Committee, the University of Oxford, and the institutional research boards at the local CDCs in the 10 geographically diverse areas.
Data collection
At the local assessment clinics, detailed information about demographic and socioeconomic status (age, sex, residential area, education, marital status, and household income), lifestyle behaviors (smoking status, alcohol consumption, physical activity, and intakes of red meat, fresh vegetables, and fruits), personal medical history (hypertension and diabetes), family medical history (stroke or transient ischemic attack), and other indexes (height and weight) were collected by trained health workers.
The daily physical activity was calculated for each participant by summing the metabolic equivalent (MET) hours per day. BMI was calculated as kg/m2. Habitual dietary intake (including red meat, fresh fruits, and fresh vegetables) over the past 12 mo was assessed using a qualitative food frequency questionnaire. The diagnostic criteria for prevalent hypertension in our study included measured systolic blood pressure ≥ 140 mm Hg, measured diastolic blood pressure ≥ 90 mm Hg, self-reported hypertension, or use of medicine for hypertension at baseline. Diagnostic criteria for prevalent diabetes included measured fasting blood glucose ≥ 7.0 mmol/L, measured random blood glucose ≥ 11.1 mmol/L, or self-reported diabetes.
Assessment of tea consumption
All participants were required to answer the question of how often they had consumed tea during the past 12 mo, and this question had 5 response choices (never or almost never, only occasionally, only in certain seasons, every month but less than weekly, and usually at least once a week). Individuals who usually consumed tea at least once a week were further required to answer other questions, including how many days they drank tea in a typical week (1–2 d/wk, 3–5 d/wk, or daily or almost every day) and which type of tea they consumed most commonly (green tea, oolong tea, black tea, or other tea). Among the aforementioned participants, additional data were collected, including the number of cups of tea consumed (based on a 300-mL-sized cup) in a drinking day, the amount of tea leaves (in grams) added each time, the number of times that the tea leaves were changed during 1 consuming day, and the age at which they began to drink tea continuously. Each participant was offered a pictorial guide that showed the size of a standard cup and different gram amounts of tea leaves. The tea leaves added during 1 drinking day were calculated by multiplying the amount added each time by the number of times the tea leaves were changed.
In the present analyses, according to the frequency of tea consumption, all individuals were classified into 4 categories: never, occasionally (including only occasionally, only in certain seasons, and every month but less than weekly), weekly (i.e., weekly but less than daily), and daily. In addition to some common indexes (duration, daily cups, and daily grams of tea consumed), we introduced another 2 indicators to measure the amount of tea consumption, which were generated by multiplying the daily cups of tea consumed by the years of tea drinking, and multiplying the daily grams of tea consumed by years of tea drinking. Furthermore, daily tea drinkers were divided into classifications on the basis of tertiles of the different tea consumption indexes, including the duration of tea consumption (<19, 19–30, or >30 y), daily grams of tea consumption (<2, 2–4, or >4 g/d), daily cups of tea consumption (<2, 2–4, or >4 cups/d), number of gram-years of tea consumption (<46, 46–105, or >105 gram-years), and number of cup–years of tea consumption (<35, 35–105, or >105 cup–years).
To test the reproducibility of the tea consumption data, Li et al. (22) analyzed results for the frequency of tea consumption obtained from a subsample of the participants at 2 different times, and reported that the weighted κ coefficient for drinking tea was 0.77. In addition, the accuracy of the baseline information (including tea consumption) could be guaranteed by several quality control methods, including quality control surveys, monitoring, and resurveys; the details have been described in previously published articles (20).
Follow-up for mortality and morbidity
Incident stroke, including mortality and morbidity, was the endpoint of the present follow-up study. Detailed information on the long-term follow-up has been described in previously published articles (22, 23). Briefly, cause-specific mortality and morbidity were ascertained through electronic linkage via unique personal identification numbers to electronic hospital records from the nationwide health insurance system (which has >98% coverage across the 10 study regions), to established local disease registries, and to local death registries. The accuracy of reported stroke types was verified by a review of the original medical records by a panel of certified neurologists and stroke physicians in China. Among the stroke cases selected, >90% were confirmed by using brain imaging (23).
The 10th revision of the International Classification of Diseases was used to code incident strokes, such as ischemic stroke (I63), hemorrhagic stroke [subarachnoid (I60) or intracerebral (I61)], and other or unknown stroke types (I64), by trained staff who were “blinded” to the baseline information.
Statistical analysis
In this study, means ± SDs or percentages were calculated to describe the baseline characteristics of the participants, adjusting for age, sex, and residential area as appropriate using either multiple linear regression (for continuous outcomes) or logistic regression (for binary outcomes). Follow-up time (person–years) was calculated for each participant from the baseline date to the date of the incidence of stroke, loss to follow-up, or 31 December, 2015 (the end of follow-up in this study), whichever occurred first. The Cox proportional hazards regression model, stratified jointly by age at baseline in 5-y intervals, sex, and study area (10 regions), was used to estimate the HRs and 95% CIs. To determine the proportional hazards assumption for the Cox model, graphs and tests based on Schoenfeld residuals were applied, and no violation was identified. The multivariate-adjusted HRs of stroke were estimated with the following adjustments: age (continuous); sex (man or woman); level of education (no formal school, primary school, middle school, high school, technical school/college, or university); marital status (married, widowed, divorced/separated, or never married); household income (<2,500, 2,500–4,999, 5,000–9,999, 10,000–19,999, 20,000–34,999, or ≥35,000 yuan/y); alcohol consumption (nonconsumer, occasional consumer, ex-consumer, or current regular consumer); smoking status (never smoker, occasional smoker, ex-smoker, or current regular smoker); physical activity (continuous); BMI (continuous); prevalent hypertension or prevalent diabetes at baseline (presence or absence); intake frequencies of red meat, fresh fruits, and vegetables (daily, 4–6 d/wk, 1–3 d/wk, monthly, or rarely or never); and family history of stroke (presence or absence). Linear trends of stroke risk related to different tea consumption patterns were calculated by modeling the levels of frequency of tea consumption and the years or the amount of tea consumption as continuous variables in separate models.
For the present study, we explored the relations between tea consumption and incident stroke in accordance with the frequency, years, and amount of tea consumption using individuals who never consumed tea during the past year as a reference. Stratified analyses were carried out to examine whether the associations of daily tea consumption with incident stroke differed according to the median age (<50 or ≥50 y), sex (male or female), residential area (rural or urban), marital status (married or widowed/separated/divorced/never married), education (illiterate/primary school or middle school, and above), household income (<35,000 or ≥35,000 yuan/y), smoking status (never or ever), alcohol consumption (never or ever), median physical activity (<18.10 or ≥18.10 MET-h/d), BMI (<24.00, 24.00–28.00, or ≥28.00), hypertension (yes or no), or diabetes (yes or no). The test for the interaction was performed using a likelihood ratio test comparing models with and without a cross-product term. We intended to conduct sensitivity analyses by excluding the first 2 y of follow-up, excluding the participants younger than 40 y of age, and further adjusting for waist-to-hip ratio. In addition, to eliminate residual confounding related to the important effect factors (i.e., smoking), we conducted an association analysis among certain participants (i.e., noncurrently smoking men).
All CIs were estimated at the 95% level. All statistical tests were 2-sided and significance was defined as P < 0.05. The statistical analyses were performed with R software version 3.4.2 (28 September, 2017; R Foundation for Statistical Computing, http://www.cran.r-project.org/).
Results
Of the individuals included in this study, 40.9% of the participants (199,294) were men. In both men and women, compared with never tea consumers during the past 12 mo, many characteristics of participants who consumed tea more frequently were very different, including residence area, age, household income, smoking, alcohol consumption, etc. (Table 1). In total, 128,280 individuals (26.3%) reported drinking tea almost daily among all participants (Supplemental Table 1). Approximately 41.4% and 15.9% of individuals consumed tea daily among men and women, respectively. Green tea was the most commonly consumed tea type among daily tea consumers (86.7%). Among all daily tea consumers, the men showed a longer duration of tea consumption (25.11 compared with 24.15 y) and a greater amount of tea consumed (4.78 compared with 3.33 g/d; 117.00 compared with 81.00 gram-years) than the women, as shown in Supplemental Table 1.
TABLE 1.
Baseline characteristics of the study participants according to the frequency of tea consumption1
| Men (n = 199,294) | Women (n = 288,083) | |||||||
|---|---|---|---|---|---|---|---|---|
| Characteristics | Never | Occasionally | Weekly | Daily | Never | Occasionally | Weekly | Daily |
| Participants, n | 38,360 | 58,527 | 19,940 | 82,467 | 131,974 | 94,595 | 15,701 | 45,813 |
| Age, y | 51.95 ± 11.23 | 51.88 ± 10.59 | 51.82 ± 10.46 | 51.75 ± 10.52 | 50.92 ± 10.61 | 50.45 ± 9.88 | 49.99 ± 9.81 | 49.52 ± 10.18 |
| Rural area | 58.95 | 58.05 | 57.15 | 56.25 | 55.05 | 56.72 | 58.39 | 60.03 |
| Married | 92.95 | 93.39 | 93.80 | 94.19 | 92.31 | 92.63 | 92.93 | 93.23 |
| Middle school and higher | 64.00 | 61.84 | 59.63 | 57.38 | 35.54 | 41.51 | 47.75 | 54.05 |
| Household income ≥35,000 yuan/y | 13.81 | 17.00 | 20.76 | 25.09 | 15.04 | 16.45 | 17.98 | 19.61 |
| Ever smoker2 | 79.09 | 83.96 | 87.87 | 90.93 | 3.49 | 4.11 | 4.82 | 5.65 |
| Ever alcohol consumer3 | 80.68 | 81.09 | 81.50 | 81.89 | 34.80 | 36.23 | 37.69 | 39.17 |
| Physical activity, MET-h/d | 22.64 ± 16.79 | 22.62 ± 15.68 | 22.61 ± 14.08 | 22.59 ± 14.47 | 22.02 ± 13.53 | 20.71 ± 12.53 | 19.40 ± 12.01 | 18.09 ± 11.11 |
| BMI,4 kg/m2 | 23.42 ± 3.17 | 23.40 ± 3.20 | 23.38 ± 3.26 | 23.36 ± 3.27 | 23.65 ± 3.44 | 23.76 ± 3.36 | 23.86 ± 3.38 | 23.97 ± 3.56 |
| Average weekly consumption5 | ||||||||
| Red meat, d/wk | 3.39 ± 2.61 | 3.73 ± 2.53 | 4.07 ± 2.39 | 4.41 ± 2.43 | 3.18 ± 2.53 | 3.57 ± 2.49 | 3.95 ± 2.39 | 4.34 ± 2.41 |
| Fresh vegetables, d/wk | 6.84 ± 0.78 | 6.84 ± 0.74 | 6.84 ± 1.04 | 6.84 ± 0.72 | 6.79 ± 0.89 | 6.84 ± 0.70 | 6.89 ± 0.75 | 6.93 ± 0.53 |
| Fresh fruits, d/wk | 2.06 ± 2.31 | 2.17 ± 2.31 | 2.28 ± 2.32 | 2.40 ± 2.31 | 2.49 ± 2.47 | 2.79 ± 2.59 | 3.09 ± 2.64 | 3.39 ± 2.61 |
| Diabetes | 4.39 | 4.34 | 4.30 | 4.26 | 4.46 | 4.54 | 4.62 | 4.70 |
| Hypertension | 34.40 | 34.72 | 35.04 | 35.36 | 30.36 | 29.80 | 29.25 | 28.70 |
| Family history of stroke | 21.41 | 19.14 | 17.06 | 15.16 | 18.33 | 17.31 | 16.33 | 15.40 |
1Values are means ± SDs or percentages unless otherwise indicated, which were adjusted for age and residential region, as appropriate, using either multiple linear regression (for continuous outcomes) or logistic regression (for binary outcomes). All baseline characteristics were associated with the frequency of tea consumption, with P < 0.001 across categories, except for physical activity (P = 0.468), BMI (P = 0.002), fresh vegetable intake (P = 0.250), and diabetes (P = 0.213) among men and diabetes (P = 0.019) among women. MET, metabolic equivalent.
2Ever smokers included occasional smokers, ex-smokers, and current regular smokers.
3Ever alcohol consumers included occasional alcohol consumers, ex-alcohol consumers, and current regular alcohol consumers.
4BMI was defined as the body weight divided by the square of the height.
5Average weekly intakes of red meat, fresh vegetables, and fresh fruits were calculated by assigning participants to the midpoint of their consumption category (daily, 4–6 d/wk, 1–3 d/wk, monthly, or rarely or never).
During the 4,289,584 person-years of follow-up, a total of 38,727 incident stroke cases (∼9.03 cases/1000 person-years) were identified (including 30,312 ischemic stroke, 6,945 hemorrhagic stroke, and 1,470 other stroke cases). The frequency of tea consumption was inversely associated with the risk of incident stroke in either the age- and sex- or multivariable-adjusted models (Ptrend < 0.001). Compared with individuals who never consumed tea, the multivariable-adjusted HRs (95% CIs) of incident stroke were 0.96 (0.94, 0.99), 0.94 (0.90, 0.98), and 0.92 (0.89, 0.95) for those who consumed tea occasionally, weekly, and daily, respectively. The inverse relation persisted for both ischemic and hemorrhagic stroke (Table 2). Moreover, the risk of incident stroke among daily tea consumers decreased with increasing duration and amount of tea consumed compared with the risk of nonconsumers, which was also the case for both ischemic and hemorrhagic stroke (Table 3).
TABLE 2.
Association of frequency of tea consumption with the risk of stroke among 487,377 participants1
| Frequency of tea consumption | ||||||
|---|---|---|---|---|---|---|
| Stroke type | Participants, n | Never | Occasionally | Weekly | Daily | P trend |
| Participants, n | 487,377 | 170,334 | 153,122 | 35,641 | 128,280 | |
| Person years, n | 4,289,584 | 1,492,052 | 1,353,461 | 316,773 | 1,127,298 | |
| Total stroke2 | ||||||
| Cases, n | 38,727 | 16,283 | 10,628 | 2341 | 9475 | |
| Model 13 | 1.00 | 0.95 (0.93, 0.98) | 0.94 (0.90, 0.99) | 0.94 (0.91, 0.97) | <0.001 | |
| Model 24 | 1.00 | 0.96 (0.94, 0.99) | 0.94 (0.90, 0.98) | 0.92 (0.89, 0.95) | <0.001 | |
| Ischemic stroke | ||||||
| Cases, n | 30,312 | 13,291 | 8394 | 1798 | 6829 | |
| Model 13 | 1.00 | 0.96 (0.93, 0.98) | 0.96 (0.91, 1.01) | 0.96 (0.93, 0.999) | 0.040 | |
| Model 24 | 1.00 | 0.96 (0.93, 0.98) | 0.94 (0.90, 0.995) | 0.92 (0.89, 0.96) | <0.001 | |
| Hemorrhagic stroke | ||||||
| Cases, n | 6945 | 2421 | 1836 | 449 | 2239 | |
| Model 13 | 1.00 | 0.92 (0.86, 0.98) | 0.86 (0.77, 0.95) | 0.85 (0.79, 0.92) | <0.001 | |
| Model 24 | 1.00 | 0.96 (0.90, 1.03) | 0.89 (0.80, 0.99) | 0.86 (0.80, 0.93) | <0.001 | |
1Values were obtained from a Cox proportional hazards analysis. Ptrend, P for the linear trend test.
2Including hemorrhagic stroke, ischemic stroke, and stroke of unknown type.
3Model 1: HRs (95% CIs) were achieved after adjusting for age and sex.
4Model 2: HRs (95% CIs) were achieved after adjusting for age; sex; marital status; education; annual household income; smoking status; alcohol consumption; physical activity; BMI; history of hypertension; history of diabetes; intake frequencies of red meat, fresh fruits, and fresh vegetables; and family history of stroke.
TABLE 3.
Association of tea consumption with stroke risk according to duration and amount of tea consumed1
| Total stroke2 | Ischemic stroke | Hemorrhagic stroke | ||||
|---|---|---|---|---|---|---|
| Variables of tea consumption3 | HR (95% CI)4 | P trend | HR (95% CI)4 | P trend | HR (95% CI) 4 | P trend |
| Never | 1.00 | 1.00 | 1.00 | |||
| Duration of tea consumption, y | <0.001 | 0.002 | <0.001 | |||
| <19 | 0.95 (0.91, 0.998) | 0.97 (0.92, 1.02) | 0.88 (0.79, 0.98) | |||
| 19–30 | 0.89 (0.84, 0.94) | 0.89 (0.84, 0.96) | 0.85 (0.75, 0.97) | |||
| >30 | 0.92 (0.87, 0.97) | 0.93 (0.88, 0.99) | 0.81 (0.72, 0.91) | |||
| Amount of tea consumed, g/d | <0.001 | 0.002 | <0.001 | |||
| <2 | 0.96 (0.91, 1.00) | 0.96 (0.91, 1.02) | 0.90 (0.81, 0.99) | |||
| 2–4 | 0.91 (0.87, 0.95) | 0.93 (0.88, 0.98) | 0.82 (0.74, 0.91) | |||
| >4 | 0.91 (0.87, 0.96) | 0.92 (0.87, 0.98) | 0.82 (0.72, 0.92) | |||
| Amount of tea consumed, gram-years | <0.001 | 0.001 | <0.001 | |||
| <46 | 0.95 (0.91, 1.00) | 0.96 (0.91, 1.01) | 0.90 (0.80, 0.996) | |||
| 46–105 | 0.92 (0.88, 0.97) | 0.94 (0.89, 0.997) | 0.83 (0.74, 0.93) | |||
| >105 | 0.90 (0.86, 0.94) | 0.91 (0.86, 0.96) | 0.81 (0.73, 0.91) | |||
| Amount of tea consumed, cups/d | <0.001 | 0.002 | <0.001 | |||
| <2 | 0.93 (0.89, 0.98) | 0.95 (0.90, 1.01) | 0.88 (0.79, 0.98) | |||
| 2–4 | 0.93 (0.89, 0.98) | 0.96 (0.90, 1.01) | 0.88 (0.79, 0.98) | |||
| >4 | 0.92 (0.87, 0.97) | 0.91 (0.86, 0.97) | 0.78 (0.69, 0.88) | |||
| Amount of tea consumed, cup–years | <0.001 | 0.009 | <0.001 | |||
| <34 | 0.95 (0.90, 1.01) | 0.94 (0.89, 1.01) | 0.92 (0.82, 1.03) | |||
| 34–105 | 0.93 (0.88, 0.97) | 0.93 (0.88, 0.99) | 0.80 (0.71, 0.89) | |||
| >105 | 0.91 (0.86, 0.95) | 0.94 (0.89, 0.996) | 0.83 (0.74, 0.93) | |||
1Values were obtained from a Cox proportional hazards analysis. Ptrend, P for the linear trend test.
2Including hemorrhagic stroke, ischemic stroke, and stroke of unknown type.
3Variables of tea consumption (duration and amount of tea consumed) were only calculated among daily tea consumers.
4Adjusted for age; sex; marital status; education; annual household income; smoking status; alcohol consumption; physical activity; BMI; history of hypertension; history of diabetes; intake frequencies of red meat, fresh fruits, and fresh vegetables; and family history of stroke.
As shown in Table 4, the frequency of tea consumption was strongly associated with a reduced risk of incident stroke among men: the multivariable-adjusted HRs (95% CIs) were 0.94 (0.90, 0.98), 0.88 (0.82, 0.94), and 0.89 (0.85, 0.93) for occasional, weekly, and daily tea consumers compared with nonconsumers (Ptrend < 0.001), respectively. For women, although the association of frequency of tea consumption with stroke risk was not obvious for either ischemic or hemorrhagic stroke, the risk of total stroke presented a declining trend as the frequency of tea consumption increased with near statistical significance (Ptrend = 0.053, Table 4). Furthermore, among daily tea consumers, the risk of stroke significantly decreased with increasing duration and amount of tea consumed in men, whereas there were no similar findings for women (Supplemental Tables 2 and 3).
TABLE 4.
Association of frequency of tea consumption with the risk of stroke among male and female participants1
| Frequency of tea consumption | |||||||
|---|---|---|---|---|---|---|---|
| Sex | Stroke type | Participants, n | Never | Occasionally | Weekly | Daily | P trend |
| Men | |||||||
| Participants, n | 199,294 | 38,360 | 58,527 | 19,940 | 82,467 | ||
| Person years, n | 1,722,946 | 322,076 | 507,410 | 175,236 | 718,224 | ||
| Cases, n | 18,038 | 5095 | 5061 | 1417 | 6465 | ||
| HR (95% CI)2 | |||||||
| Total stroke3 | 1.00 | 0.94 (0.90, 0.98) | 0.88 (0.82, 0.94) | 0.89 (0.85, 0.93) | <0.001 | ||
| Ischemic stroke | 1.00 | 0.93 (0.89, 0.97) | 0.87 (0.81, 0.94) | 0.89 (0.85, 0.94) | <0.001 | ||
| Hemorrhagic stroke | 1.00 | 0.92 (0.84, 1.02) | 0.85 (0.73, 0.97) | 0.82 (0.74, 0.90) | <0.001 | ||
| Women | |||||||
| Participants, n | 288,083 | 131,974 | 94,595 | 15,701 | 45,813 | ||
| Person years, n | 2,566,640 | 1,169,977 | 846,051 | 141,537 | 409,075 | ||
| Cases, n | 20,689 | 11,188 | 5567 | 924 | 3010 | ||
| HR (95% CI)2 | |||||||
| Total stroke3 | 1.00 | 0.97 (0.94, 1.00) | 1.01 (0.94, 1.09) | 0.95 (0.90, 0.998) | 0.053 | ||
| Ischemic stroke | 1.00 | 0.97 (0.93, 1.01) | 1.03 (0.95, 1.11) | 0.95 (0.89, 1.01) | 0.120 | ||
| Hemorrhagic stroke | 1.00 | 0.97 (0.89, 1.06) | 0.92 (0.77, 1.10) | 0.92 (0.82, 1.04) | 0.177 | ||
1Values were obtained from a Cox proportional hazards analysis. Ptrend, P for the linear trend test.
2Adjusted for age; marital status; education; annual household income; smoking status; alcohol consumption; physical activity; BMI; history of hypertension; history of diabetes; intake frequencies of red meat, fresh fruits, and fresh vegetables; and family history of stroke.
3Including hemorrhagic stroke, ischemic stroke, and stroke of unknown type.
We further investigated the associations between different types of tea and stroke risk (Figure 1, Supplemental Table 4). After multivariate adjustment, the frequency of green tea consumption was inversely related to total stroke and the 2 main stroke types. In contrast, there was no significant association between nongreen tea consumption and stroke risk. In addition, consistent results were observed in the relations between different types of tea and stroke risk, regarding the duration and amount of tea consumed (Supplemental Tables 5 and 6).
FIGURE 1.
Association of frequency of tea consumption with the risk of stroke according to type of tea. (A) Results for the relations of green tea consumption to risks of total, ischemic, and hemorrhagic stroke; (B) results for the relations of nongreen tea consumption to risks of total, ischemic, and hemorrhagic stroke. Values were obtained from a Cox proportional hazards analysis. Analyses were adjusted for age; sex; marital status; education; annual household income; smoking status; alcohol consumption; physical activity; BMI; history of hypertension; history of diabetes; intake frequencies of red meat, fresh fruits, and fresh vegetables; and family history of stroke. Total stroke included hemorrhagic stroke, ischemic stroke, and stroke of unknown type. Round dots represent the HRs and vertical lines represent the corresponding 95% CIs.
We analyzed the relations of daily tea consumption with stroke risk in specific population subgroups (Figure 2). The strength of the associations of tea consumption with the risk of stroke was largely consistent across subgroups, which were classified by age, smoking status, alcohol consumption, physical activity, BMI, hypertension, and diabetes (Pheterogeneity > 0.05, Figure 2), but the residential region subgroup was not consistent (rural compared with urban) (Pheterogeneity = 0.031): stroke risk was significantly lower in the rural areas, but not in the urban regions.
FIGURE 2.
Subgroup analysis of associations between daily tea consumption and stroke risk according to potential baseline risk factors. HRs and 95% CIs for stroke were obtained by the comparison between daily tea consumers and individuals who never consumed tea during the past year. Values were obtained from a Cox proportional hazards analysis. Analyses were adjusted for age; sex; marital status; education; annual household income; smoking status; alcohol consumption; physical activity; BMI; history of hypertension; history of diabetes; intake frequencies of red meat, fresh fruits, and fresh vegetables; and family history of stroke, except for the stratification factor. Round dots represent the HRs and horizontal lines represent the corresponding 95% CIs. MET, metabolic equivalent.
To eliminate residual confounding factors related to important effect factors, we conducted an association analysis among noncurrent smokers (including nonsmokers) and discovered a negative correlation (Supplemental Table 7). In addition, a similar inverse association could also be found for noncurrent alcohol consumers (including nonconsumers) among men (Supplemental Table 7). Excluding the participants younger than 40 y of age did not substantially modify the observed relation (Supplemental Table 8), and the same was true for further adjustment for waist-to-hip ratio and excluding the first 2 y of follow-up (data not shown).
Discussion
In this study, we observed that the frequency of tea consumption was related to a decreased risk of stroke. Compared with nonconsumers, daily tea consumers had an 8% lower risk of stroke. According to the type of tea, green tea consumption was significantly associated with a reduction in stroke risk; however, this was not the case for nongreen tea. The inverse relations were significant among men but not among women.
Several prior prospective studies have explored the association of tea consumption with stroke (1, 10-14). Nevertheless, previous cohort studies have shown inconsistent results. For instance, Kokubo et al. (13) reported that compared with nonconsumers, the HRs (95% CIs) of stroke were 0.86 (0.78, 0.95) and 0.80 (0.73, 0.95) for tea consumers who drank 2–3 and ≥4 cups of tea per day, respectively (with 82,369 participants). No such significant association, however, was identified in the European Prospective Investigation into Cancer and Nutrition study with 37,514 participants (14). A recent meta-analysis of prospective studies involving 513,804 participants identified that an increase in tea consumption of 3 cups/d was related to a 13% decreased risk of stroke (24). The 14 studies in this meta-analysis were mainly from Japan, Europe, and America, where the commonly consumed tea types (black tea in Europe, whereas green tea is more common in China), tea green-removing technologies (stir-frying green in Japan, steaming green in China), tea-drinking habits (usually drinking tea with milk in Europe, whereas this is uncommon in China), and other factors were different from those in China (15-17).
Currently, as far as we know there is no standard population-based prospective study focusing on the association of drinking tea with stroke in a Chinese population. A Chinese case-control study reported that a significant decline in the risk of ischemic stroke was associated with tea consumption (25). A similar negative relation was observed in our current study. In addition to using a cohort study design with stronger causal verification ability and involving a larger sample size (0.5 million), our study analyzed the main types of stroke (ischemic and hemorrhagic stroke), controlled confounding factors more stringently, and included stricter quality control than the aforementioned case-control study.
Tea is divided into 3 main categories, i.e., green, black, and oolong tea, according to the degree of fermentation during processing: green tea is nonoxidized and black and oolong tea are oxidized and partially oxidized, respectively (26). The different manufacturing processes could affect the components of tea and further influence the functions of tea related to stroke (27). For example, green tea had higher antioxidant properties than nongreen tea, indicating possibly different effects on stroke risk. In accordance with the aforementioned situation, our study identified different effects of green and nongreen tea on stroke risk. Similar to our results, Tanabe et al. (11) found that green tea consumption was associated with reduced stroke risk, but an inverse relation could not be observed between other kinds of tea and stroke risk. In addition, a meta-analysis revealed that green tea consumption was related to stroke risk, but did not find similar effects of other types of tea (3). In our study, the minority of tea drinkers consumed black and oolong tea, which might affect our further exploration of the effects of the 2 kinds of tea on stroke risk.
We found that tea consumption was strongly associated with a reduced stroke risk among men, whereas similar evidence was lacking for women. The proportion of tea consumers was 80.8% and 45.8% among men and women, respectively. In addition, compared with women, the proportion of daily tea consumers and the duration and the amount of tea consumption for male tea consumers were both significantly greater, which might partly explain why the protective effects of tea consumption on stroke were more significant for men. It was reported that differences in risk factors between men and women could strongly affect stroke risk. Alcohol abuse and cigarette smoking were proposed to be leading risk factors for men (28, 29). Both factors were adjusted for in our analyses, and to avoid residual confounding we limited the analyses only to male noncurrent smokers/noncurrent alcohol consumers, and the results still revealed a negative relation between tea consumption and stroke risk. For women, it has been reported that 2 factors, atrial fibrillation and stress, were more prevalent (28, 29). Nevertheless, owing to lack of related information, these 2 effect factors were not adjusted for, which might somewhat modify the results among women.
Increasing evidence has shown that the protective effects of tea on stroke risk might have different biological bases (5-7). Tea polyphenols, which are the most important components in tea, especially green tea, could exert a wide spectrum of beneficial effects, including modulating the plasma lipid profile, decreasing plasma glucose, and reducing the risk of atherosclerosis, all of which are vital risk factors for stroke (8, 9, 30). Furthermore, catechins, a major category of polyphenols, might block increases in serum nitric oxide concentration and improve endothelial function, which could also reduce stroke risk (31, 32). In addition, tea could significantly reduce blood pressure, which is the most predominant risk factor for stroke (31, 32).
The results of this prospective study with 0.5 million participants from 10 geographic areas might be a good representation of the situation of some Chinese adults. Moreover, the grams of tea consumed indicator, was a vital analysis index in our study and might better reflect the amount of tea consumed than other indexes. In addition, based on daily grams per cup and year of tea drinking, we proposed new indexes to indicate the cumulative amount of tea consumed. This study also has some limitations. First, although we adjusted for many influencing factors in the analyses according to previously published studies (22, 33), residual confounding caused by other dietary patterns, total energy intake, blood cholesterol concentration, etc., still persisted owing to a lack of related information, which might affect our results to some extent. In order to achieve the real causality of associations, in future studies more detailed information about dietary factors should be gathered and randomized controlled trials should be considered. Second, coffee might confound our results as well. In China, apart from tea, the major beverage might be water, at least among middle-aged or elderly populations, with very little consumption of coffee based on the findings of other nationwide surveys during the early 2000s (34). In addition, both resurveys (one during 2008, the other during 2013–2014) showed that <2% of randomly chosen participants consumed coffee at least once a week (22, 35). Third, information on tea consumption at baseline was collected based on self-report by using a simple qualitative questionnaire. According to the previously published CKB study, the weighted κ coefficient for drinking tea was 0.77, which means that the consistency is quite satisfactory (22). However, further validation against gold standards including circulating biomarkers of tea metabolites could not be conducted. Fourth, although the P value for the trend test showed statistical significance, there was limited evidence of the dose-response relation between tea consumption and stroke risk. First, the protective effect of tea consumption was relatively modest; second, some other unavailable risk factors might have had a certain impact on the relation.
In summary, this large Chinese cohort study demonstrated that higher consumption of tea, especially green tea, was inversely related to the risks of total stroke and different stroke types, including ischemic and hemorrhagic stroke. Further experimental studies are required to explore the causal relation.
Supplementary Material
Acknowledgments
The most important acknowledgement is to the 10 regional centers and the Beijing and Oxford centers. The details of acknowledgment are listed in Supplemental Acknowledgments and the members of the steering committee and collaborative group are listed in the supplemental material as well.
The authors’ responsibilities were as follows—TT, HS, and ZH: conceived and designed the research; TT and GJ: performed the statistical analyses and drafted the manuscript; JL, HS, ZC, and LL: supervised the conduct of the research and had primary responsibility for the final content; and all authors: contributed to interpretation of the results, reviewed the manuscript for important intellectual content, and read and approved the final manuscript. The authors report no conflicts of interest.
Notes
Supported by National Natural Science Foundation of China grants 81390543 and 81390540 and National Key Research and Development Program of China grants 2016YFC0900500, 2016YFC0900501, and 2016YFC0900504. The China Kadoorie Biobank baseline survey and the first resurvey were supported by a grant from the Kadoorie Charitable Foundation in Hong Kong. The long-term follow-up is supported by UK Wellcome Trust grants 202922/Z/16/Z, 088158/Z/09/Z, and 104085/Z/14/Z and Chinese Ministry of Science and Technology grant 2011BAI09B01.
The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, and approval of the manuscript; or the decision to submit the manuscript for publication.
Supplemental Figure 1, Supplemental Tables 1–8, and Supplemental Acknowledgments are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/ajcn/.
TT and JL contributed equally to this work.
The China Kadoorie Biobank (CKB) data access policy and procedures are available at www.ckbiobank.org. All researchers can apply to use the CKB data by registering and applying at the website: http://www.ckbiobank.org/site/Data+Access.
Abbreviations used: CDC, Chinese Center for Disease Control and Prevention; CKB, China Kadoorie Biobank; MET, metabolic equivalent.
References
- 1. Kuriyama S, Shimazu T, Ohmori K, Kikuchi N, Nakaya N, Nishino Y, Tsubono Y, Tsuji I. Green tea consumption and mortality due to cardiovascular disease, cancer, and all causes in Japan: the Ohsaki study. JAMA. 2006;296(10):1255–65. [DOI] [PubMed] [Google Scholar]
- 2. Khan N, Mukhtar H. Tea and health: studies in humans. Curr Pharm Des. 2013;19(34):6141–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Zhang C, Qin YY, Wei X, Yu FF, Zhou YH, He J. Tea consumption and risk of cardiovascular outcomes and total mortality: a systematic review and meta-analysis of prospective observational studies. Eur J Epidemiol. 2015;30(2):103–13. [DOI] [PubMed] [Google Scholar]
- 4. Sato Y, Nakatsuka H, Watanabe T, Hisamichi S, Shimizu H, Fujisaku S, Ichinowatari Y, Ida Y, Suda S, Kato K et al.. Possible contribution of green tea drinking habits to the prevention of stroke. Tohoku J Exp Med. 1989;157(4):337–43. [DOI] [PubMed] [Google Scholar]
- 5. Curin Y, Andriantsitohaina R. Polyphenols as potential therapeutical agents against cardiovascular diseases. Pharmacol Rep. 2005;57(Suppl):97–107. [PubMed] [Google Scholar]
- 6. Arab L, Liebeskind DS. Tea, flavonoids and stroke in man and mouse. Arch Biochem Biophys. 2010;501(1):31–6. [DOI] [PubMed] [Google Scholar]
- 7. Larsson SC. Coffee, tea, and cocoa and risk of stroke. Stroke. 2014;45(1):309–14. [DOI] [PubMed] [Google Scholar]
- 8. Basu A, Lucas EA. Mechanisms and effects of green tea on cardiovascular health. Nutr Rev. 2007;65(8 Pt 1):361–75. [DOI] [PubMed] [Google Scholar]
- 9. Yung LM, Leung FP, Wong WT, Tian XY, Yung LH, Chen ZY, Yao XQ, Huang Y. Tea polyphenols benefit vascular function. Inflammopharmacology. 2008;16(5):230–4. [DOI] [PubMed] [Google Scholar]
- 10. Larsson SC, Mannisto S, Virtanen MJ, Kontto J, Albanes D, Virtamo J. Coffee and tea consumption and risk of stroke subtypes in male smokers. Stroke. 2008;39(6):1681–7. [DOI] [PubMed] [Google Scholar]
- 11. Tanabe N, Suzuki H, Aizawa Y, Seki N. Consumption of green and roasted teas and the risk of stroke incidence: results from the Tokamachi-Nakasato cohort study in Japan. Int J Epidemiol. 2008;37(5):1030–40. [DOI] [PubMed] [Google Scholar]
- 12. Larsson SC, Virtamo J, Wolk A. Black tea consumption and risk of stroke in women and men. Ann Epidemiol. 2013;23(3):157–60. [DOI] [PubMed] [Google Scholar]
- 13. Kokubo Y, Iso H, Saito I, Yamagishi K, Yatsuya H, Ishihara J, Inoue M, Tsugane S. 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. Stroke. 2013;44(5):1369–74. [DOI] [PubMed] [Google Scholar]
- 14. de Koning Gans JM, Uiterwaal CS, van der Schouw YT, Boer JM, Grobbee DE, Verschuren WM, Beulens JW. Tea and coffee consumption and cardiovascular morbidity and mortality. Arterioscler Thromb Vasc Biol. 2010;30(8):1665–71. [DOI] [PubMed] [Google Scholar]
- 15. Weisburger JH. Tea and health: a historical perspective. Cancer Lett. 1997;114(1–2):315–17. [DOI] [PubMed] [Google Scholar]
- 16. Ferrara L, Montesano D, Senatore A. The distribution of minerals and flavonoids in the tea plant (Camellia sinensis). Farmaco. 2001;56(5–7):397–401. [DOI] [PubMed] [Google Scholar]
- 17. Tanaka K, Miyake Y, Fukushima W, Sasaki S, Kiyohara C, Tsuboi Y, Yamada T, Oeda T, Miki T, Kawamura N et al.. Intake of Japanese and Chinese teas reduces risk of Parkinson's disease. Parkinsonism Relat Disord. 2011;17(6):446–50. [DOI] [PubMed] [Google Scholar]
- 18. Yang G, Wang Y, Zeng Y, Gao GF, Liang X, Zhou M, Wan X, Yu S, Jiang Y, Naghavi M et al.. Rapid health transition in China, 1990–2010: findings from the Global Burden of Disease Study 2010. Lancet. 2013;381(9882):1987–2015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham J, Adair T, Aggarwal R, Ahn SY et al.. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095–128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Chen Z, Chen J, Collins R, Guo Y, Peto R, Wu F, Li L. China Kadoorie Biobank of 0.5 million people: survey methods, baseline characteristics and long-term follow-up. Int J Epidemiol. 2011;40(6):1652–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Tian T, Jin G, Yu C, Lv J, Guo Y, Bian Z, Yang L, Chen Y, Shen H, Chen Z et al.. Family history and stroke risk in China: evidence from a large cohort study. J Stroke. 2017;19(2):188–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Li X, Yu C, Guo Y, Bian Z, Si J, Yang L, Chen Y, Ren X, Jiang G, Chen J et al.. Tea consumption and risk of ischaemic heart disease. Heart. 2017;103(10):783–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sun L, Clarke R, Bennett D, Guo Y, Walters RG, Hill M, Parish S, Millwood IY, Bian Z, Chen Y et al.. Causal associations of blood lipids with risk of ischemic stroke and intracerebral hemorrhage in Chinese adults. Nat Med. 2019;25(4):569–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Shen L, Song LG, Ma H, Jin CN, Wang JA, Xiang MX. Tea consumption and risk of stroke: a dose-response meta-analysis of prospective studies. J Zhejiang Univ Sci B. 2012;13(8):652–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Liang W, Lee AH, Binns CW, Huang R, Hu D, Zhou Q. Tea consumption and ischemic stroke risk: a case-control study in southern China. Stroke. 2009;40(7):2480–5. [DOI] [PubMed] [Google Scholar]
- 26. Graham HN. Green tea composition, consumption, and polyphenol chemistry. Prev Med. 1992;21(3):334–50. [DOI] [PubMed] [Google Scholar]
- 27. Mineharu Y, Koizumi A, Wada Y, Iso H, Watanabe Y, Date C, Yamamoto A, Kikuchi S, Inaba Y, Toyoshima H et al.. Coffee, green tea, black tea and oolong tea consumption and risk of mortality from cardiovascular disease in Japanese men and women. J Epidemiol Community Health. 2011;65(3):230–40. [DOI] [PubMed] [Google Scholar]
- 28. Strozynska E, Fiszer U, Ryglewicz D, Zaborski J. The impact of risk burden differences between men and women on the clinical course of ischemic stroke. J Stroke Cerebrovasc Dis. 2016;25(4):843–7. [DOI] [PubMed] [Google Scholar]
- 29. Traylor M, Rutten-Jacobs LC, Holliday EG, Malik R, Sudlow C, Rothwell PM, Maguire JM, Koblar SA, Bevan S, Boncoraglio G et al.. Differences in common genetic predisposition to ischemic stroke by age and sex. Stroke. 2015;46(11):3042–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Babu PV, Liu D. Green tea catechins and cardiovascular health: an update. Curr Med Chem. 2008;15(18):1840–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Bhardwaj P, Khanna D. Green tea catechins: defensive role in cardiovascular disorders. Chin J Nat Med. 2013;11(4):345–53. [DOI] [PubMed] [Google Scholar]
- 32. Bhardwaj P, Khanna D, Balakumar P. Catechin averts experimental diabetes mellitus-induced vascular endothelial structural and functional abnormalities. Cardiovasc Toxicol. 2014;14(1):41–51. [DOI] [PubMed] [Google Scholar]
- 33. Du H, Li L, Bennett D, Guo Y, Key TJ, Bian Z, Sherliker P, Gao H, Chen Y, Yang L et al.. Fresh fruit consumption and major cardiovascular disease in China. N Engl J Med. 2016;374(14):1332–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Lee YH, Wang Z, Chiang TC, Liu CT. Beverage intake, smoking behavior, and alcohol consumption in contemporary China—a cross-sectional analysis from the 2011 China Health and Nutrition Survey. Int J Environ Res Public Health. 2017;14(5):493. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Shen Q, Yu C, Guo Y, Bian Z, Zhu N, Yang L, Chen Y, Luo G, Li J, Qin Y et al.. Habitual tea consumption and risk of fracture in 0.5 million Chinese adults: a prospective cohort study. Nutrients. 2018;10(11):1633. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.