Integrative Cancer Therapies

http://ict.sagepub.com/

The Effects of Green Tea Consumption on Incidence of Breast Cancer and Recurrence of Breast Cancer: A
Systematic Review and Meta-analysis
Dugald Seely, Edward J. Mills, Ping Wu, Shailendra Verma and Gordon H. Guyatt
Integr Cancer Ther 2005 4: 144
DOI: 10.1177/1534735405276420

The online version of this article can be found at:

http://ict.sagepub.com/content/4/2/144

Published by:

http://www.sagepublications.com

Additional services and information for Integrative Cancer Therapies can be found at:

Email Alerts: http://ict.sagepub.com/cgi/alerts

Subscriptions: http://ict.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://ict.sagepub.com/content/4/2/144.refs.html

>> Version of Record - May 23, 2005

What is This?

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

 Seely et al
Seely et
10.1177/1534735405276420
Green Teaal and Breast Cancer: Systematic Review
The Effects of Green Tea Consumption on Incidence
of Breast Cancer and Recurrence of Breast Cancer:
A Systematic Review and Meta-analysis
Dugald Seely, ND, MSc Cand, Edward J. Mills, DPH, MSc, Ping Wu, MBBS, MSc Cand, Shailendra
Verma, MD, FRCP(C), and Gordon H. Guyatt, MD, MSc, FRCP(C)
Background: Green tea is widely used by women for the pre-
vention and treatment of breast cancer. The authors aimed
to determine the efficacy of green tea ingestion on the risk
of breast cancer development and the risk of breast cancer
recurrence. Methods: The authors conducted a systematic re-
view and meta-analyses of observational studies from sys-
tematic searches of 8 electronic data sources and contact
with authors. They included studies assessing breast cancer
incidence and recurrence. Results from cohort studies and
case-control studies were pooled separately using a random
effects model with testing of a priori hypotheses to explain
heterogeneity. Results: The pooled relative risk (RR) of de-
veloping breast cancer for the highest levels of green tea
consumption in cohort studies was 0.89 (95% confidence
2
interval [CI], 0.71-1.1; P = .28; I = 0%), and in case control
studies, the odds ratio was 0.44 (95% CI, 0.14-1.31; P = .14;
2
I = 47%). The pooled RR of cohort studies for breast cancer
recurrence in all stages was 0.75 (95% CI, 0.47-1.19; P = .22;
2 The evidence for green tea’s use as an anticancer
I = 37%). A subgroup analysis of recurrence in stage I and II
disease showed a pooled RR in cohort studies of 0.56 (95%
2
CI, 0.38-0.83; P = .004; I = 0%). Dose-response relationships
were evident in only 3 of the 7 studies. Conclusion: To date,
the epidemiological data indicates that consumption of 5 or
more cups of green tea a day shows a non-statistically signifi-
cant trend towards the prevention of breast cancer develop-
ment. Evidence indicates that green tea consumption may
possibly help prevent breast cancer recurrence in early stage
(I and II) cancers. However, conclusions as to the potential
therapeutic application of green tea are currently impossi-
ble to make due to the small number of studies conducted,
the lack of any clinical trial evidence, the lack of a consistent
dose-response relationship, and the potential for interaction
with standard care.
Keywords: green tea; Camellia sinensis; breast cancer; system-
atic review; meta-analysis
The use of green tea (Camellia sinensis) as a traditional
cancer treatment has a long history in Asian cultures.
DOI: 10.1177/1534735405276420
144
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
The use of green tea as an adjunct cancer therapy in
Western culture is increasing, possibly due to a grow-
ing enthusiasm for complementary therapies to com-
bat cancer.1 Background information on green tea is
provided in Table 1.
Numerous animal and in vitro studies have
explored the mechanisms of green tea’s potential
anticancer activity. In vitro studies have demonstrated
that the green tea catechins, especially epigallocatechin
gallate (EGCG), have antioxidant properties,2-5 have
6-10
antiangiogenic properties, suppress neoplastic cell
11,12
proliferation, inhibit the formation of N-nitroso
13
compounds, inhibit cell division by telomerase inhi-
14
bition, upregulate intercellular gap junction com-
munication,
15-17
interfere with estrogen metabo-
18,19
lism, and increase apoptosis in cancer cells.
20-24
agent from human clinical studies includes several
case-control and cohort studies that have shown vari-
able results; some have suggested that green tea con-
sumption may decrease cancer risk, while others
25-29
report no benefit. To date, no randomized con-
trolled trials have addressed outcomes of green tea
consumption on breast cancer. In this article, we syn-
thesize the current epidemiological evidence to deter-
mine the effects of green tea consumption on breast
cancer incidence and recurrence.
DS, EJM, and PW are at the Department of Clinical Epidemiology,
Canadian College of Naturopathic Medicine, Toronto, Ontario,
Canada; DS is at the Institute of Medical Science, University of To-
ronto, Ontario, Canada; EJM and GHG are at the Department of
Clinical Epidemiology and Biostatistics, McMaster University,
Hamilton, Ontario, Canada; PW is at the London School of Hygiene
and Tropical Medicine, London, United Kingdom; and SV is at the
Ottawa Cancer Centre, University of Ottawa, Ontario, Canada.
Correspondence: Dugald Seely, 1255 Sheppard Ave East, To-
ronto, Ontario M2K 1E2, Canada. E-mail: dseely@ccnm.edu.
INTEGRATIVE CANCER THERAPIES 4(2); 2005 pp. 144-155
 Green Tea and Breast Cancer: Systematic Review
Table 1. Background Information on Green Tea
Green tea: Tea made from the dried leaves of Camellia
sinensis (L.) Kuntze (Theaceae) is one of the
most widely consumed beverages in the world.
Approximately 20% of the world’s tea is con-
sumed as green tea. Green tea is produced by
briefly heating or steaming the shrub’s leaves
immediately after harvest. This denatures
enzymes in the leaves, resulting in minimal oxida-
tion of the herb’s naturally occurring polyphenols,
commonly referred to as catechins. As well as for
its claimed benefit in cardiovascular health, den-
tal health, weight loss, and cognitive function, this
herb is used for its purported beneficial effects on
cancer.
Synonyms: Camellia sinensis, Thea sinensis L., Camellia
assamica (J W Mast) Hung T Chang, Chinese
tea, Japanese tea, Gruner tea, green tea extract,
GTE, Matsu-cha tea.
Source: Leaves of the perennial shrub Camellia sinensis
Constituents: The principal constituents are the flavonol
polyphenols, known as catechins: 10%-15%
(-)-epigallocatechin-3-gallate, 6%-10%
(-)-epigallocatechin, 2%-3% (-)-epicatechin
gallate, 2% (-)-epicatechin, 2%-3% caffeine
(percentages shown are derived from the dry
weight of the leaf). Other constituents include
theophylline, theobromine, protein (15%-20%),
fiber, sugars (5%), B vitamins, ascorbic acid, L-
theanine, gallic acid, malic acid, oxalic acid,
and fats.
Methods
Literature Search
We searched Medline (1966-November 2004),
AMED (1985-November 2004), AltHealthWatch
(1990-November 2004), CancerLit (1963-November
2004), CHKI (Chinese database, inception to October
2004), CinAhl (1982-November 2004), EMBASE
(1980-November 2004), and the Cochrane Library
(inception to November 2004) for relevant studies in
all languages. We supplemented this search by review-
ing reference sections of relevant articles and by
searching for unpublished trials on the National Re-
search Register (United Kingdom; October 1998-No-
vember 2004) and Clinical Trials.Gov (February 2000-
November 2004). In addition to the systematic data-
base searches, we contacted authors of relevant stud-
ies.
Study Selection and Characteristics
Two reviewers (D.S. and P.W.) independently evalu-
ated the abstracts identified by our search. To be in-
cluded in the final analysis, studies had to use
observational or randomized trial designs to examine
the effect of green tea consumption on breast cancer
incidence or recurrence in men and women of all
ages. Studies were excluded if they did not examine
green tea’s effect in isolation from other teas.
INTEGRATIVE CANCER THERAPIES 4(2); 2005
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
Data Abstraction and Validity Assessment
Two authors (D.S. and P.W.) independently extracted
data and appraised the characteristics of the studies.
In the case of the Chinese language database, P.W., in
consultation with D.S., evaluated and extracted the in-
formation from the 1 article retrieved via this data-
base.30 Population profile, response rate, confounding
factors, green tea dosage, and dose-response relation-
ships in the studies were determined by consensus. A
third reviewer (E.M.) dealt with any discrepancies as
they arose.
Statistics
Kappa scores were calculated to determine chance-
adjusted interobserver agreement in the study identi-
fication process. We calculated relative risk (RR) for
cohort studies and odds ratio (OR) for case-control
studies. The outcomes of primary interest were inci-
dence of breast cancer in previously undiagnosed
women and recurrence of breast cancer in those with a
prior diagnosis. In each of the studies, all dosage levels
of green tea were examined to determine if a dose-
response relationship existed. In the final analysis,
highest consumption levels were used to determine
pooled RRs and ORs. Highest consumption levels
were chosen a priori to maximize the likelihood of de-
tecting an association if indeed it was present.
Pooled analysis was conducted using a random
effects model. The outcomes of breast cancer inci-
dence and recurrence are clearly disparate and consti-
tute 2 separate populations and dictate 2 separate
analyses. We tested for homogeneity using the Zalen
test and the I2 statistic.31 A priori explanations of heter-
ogeneity include cancer stage (stages I and II vs stages
III and IV), study quality, and green tea dosage. Publi-
cation bias was tested using both the Egger test with
funnel plot and Kendall’s test on standardized effect
versus variance. StatsDirect was used for all meta-
analytic procedures (StatsDirect, Manchester, UK).
Results
The systematic literature search yielded 72 abstracts,
of which 10 were identified for further examination
and the full articles collected (see Figure 1). Three of
the 10 studies were excluded: 1 study because it exam-
ined catechin consumption rates based on overall diet
and catechin excretion and did not look at green tea
in isolation,32 and another because it looked at the in-
cidence of all cancers without isolating the incidence
33
of breast cancer alone (the authors declined to pro-
34
vide breast cancer–specific data), and the third study
because it was a follow-up to a study already included
35
in our review (this provided genotype examination
of women with breast cancer according to their green
145
 Seely et al
• 72 articles screened for inclusion
• 10 articles retrieved for further
analysis
• 7 studies included for systematic review and meta-
analysis
• 3 cohort studies examining the risk of breast cancer
occurrence in relation to green tea consumption
(table 2)
• 2 case-control studies examining the risk of breast
cancer occurrence in relation to green tea
consumption (table 3)
• 2 cohort studies examining the risk of recurrence of
breast cancer in relation to green tea consumption
(table 4)
Figure 1 Flowchart detailing study selection and exclusion for systematic review.
34
tea consumption levels ). Publication bias was as-
sessed, but too few studies were included for a
meaningful interpretation of the tests.
Thus, 7 studies were included for analysis. The κ
value for inclusion of the articles was 1.0, indicating
complete agreement between the initial reviewers
(D.S. and P.W.). Tables 2, 3, and 4 display the study
characteristics and results. Figure 1 displays the flow
diagram of study selection.
Of the 7 studies, 5 were prevention studies that
assessed risk of developing initial breast cancer, and 2
studies addressed risk of breast cancer recurrence in
relation to green tea consumption. For the 5 that dealt
with risk of breast cancer incidence, 3 were cohort
studies (N = 115 601; see Table 2)36-38 and 2 were case-
control studies (n for cases = 857, n for controls = 1519;
146
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
• 62 articles excluded as irrelevant,
review, cell line, or animal research
• 3 articles excluded from the
systematic review
• 1 study did not examine the effects of
green tea in isolation
• 1 study was a follow-up study which
hooked at genotype rather than breast
cancer incidence
• 1 study did not distinguish breast
cancer from other cancers
30,35
see Table 3). For the 2 studies that assessed the risk
of breast cancer recurrence, both were cohort studies
39,40
(N = 1632; see Table 4).
The pooled RR of cohort studies examining risk of
breast cancer incidence, using a random effects
model, is 0.89 (95% confidence interval [CI], 0.71-1.1;
P = .28; I2 = 0%; P = .8; see Figure 2). The pooled OR of
case-control studies examining risk of breast cancer
2
incidence is 0.44 (95% CI, 0.14-1.31; P = .14; I = 47%;
P = .16; see Figure 3). The small number of partici-
30
pants in the Tao study who fit the criteria of highest
green tea consumption explains the heterogeneity
between the case-controls studies.
The pooled RR of all-stage breast cancer recur-
rence in included cohort studies is 0.75 (95% CI, 0.47-
1.19; P = .22; I2 = 37%; P = .17; see Figure 4). The pooled
(text continues on page 150)
INTEGRATIVE CANCER THERAPIES 4(2); 2005
 Table 2. Cohort Studies That Prospectively Examine the Development of Breast Cancer in Relation to Green Tea Consumption
Confounding Adjusted Relative
Green Tea Consumption Factors Accounted Risk (95% CI) for All Dose-Response
Reference n (Response Rate) Population Profile Levels Analyzed and Adjusted for Consumption Levels Relationship

Key et al37 34 759 (62%) Urban Japanese women Frequency of consump- Age, calendar period, city, Set as 1.0 for lowest Not present
(age unspecified) who tion per day: ≤1 time, 2- age at time of atomic consumption; for 2-4
have been exposed to 4 times, ≥5 times bomb explosion, and times daily, RR = 1.02
ionizing radiation radiation dose (0.76-1.36); for 5 or more
times daily, RR = 0.86
(0.62-1.21)

Nagano et al36 38 540 (72%) Urban Japanese men Average number of cups City, age, gender, radia- Set as 1.0 for lowest Not present
(mean age = 52.8) and consumed daily: ≤1 tion exposure, smoking consumption; for 2-4
women (mean age = cup, 2-4 cups, ≥5 cups status, alcohol drinking, cups/d, RR = 1.2 (0.86-
56.8) who have been body mass index, edu- 1.8); for ≥5 cups/d, RR =
exposed to ionizing cation, and calendar 1.0 (0.67-1.60)
radiation time

Suzuki et al38 2 cohorts: 17 533 Rural Japanese women: Average number of cups Age, health insurance, Set as 1.0 for lowest Not present
from cohort 1 cohort 1 ≥40 y old; consumed daily by both age at menarche, consumption; RR = 0.87
(94%) and 24 cohort 2 40-64 y old cohorts: <1 cup, 1-2 menopausal status, (0.57-1.32) for 1 to 2
769 from cohort cups, 3-4 cups, ≥5 cups age at first birth, parity, cups daily, RR = 1.07

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

2 (93%) family history, smoking,
alcohol drinking, body
mass index, black tea
and coffee
consumption
(0.73-1.57) for 3 to 4
cups daily, and RR =
0.84 (0.57-1.24) for ≥5
cups daily

CI = confidence interval; RR = relative risk.

147
 148
Table 3. Case-Control Studies That Retrospectively Examine the Development of Breast Cancer in Relation to Green Tea Consumption
Confounding Adjusted OR
Green Tea Consumption Factors Accounted (95% CI) for All Dose-Response
Reference n (Response Rate) Population Profile Levels Analyzed and Adjusted for Consumption Levels Relationship

Tao et al30 356 cases (87%), Urban Chinese women
925 controls aged 30-74 y; women in
(90%) case group were an
average of 5 y younger
than in the control group
Average weight (g) of dry Age, body mass index, Set as 1.0 for zero con-
tea leaf steeped per age of menopause,
mo: none, <120 g, birth history, menarche,
>120-300 g, >300 g education, family his-
tory, economic status,
diet (31 types of foods
accounted for including
fruits, veggies, soy, and
meats); living and
working conditions
Positive trend
sumption; for <120 g/mo,
OR = 1.11 (0.47-2.60);
for 120-300 g/mo, OR =
0.67 (0.25-1.82); for
>300 g/mo, OR = 0.17
(0.03-1.04)

Wu et al35 501 cases (53%), Chinese, Japanese, and
594 controls Filipino women aged 25-
(72%) 74 y who were urban
residents in the United
States
Average volume con- Age, ethnicity, birthplace, Set as 1.0 for nondrinkers; Positive trend
sumed per day: none, age at menarche, par- for 0-85.7 mL, OR = 0.74
0-85.7 mL, ≥85.7 mL ity, menopausal status, (0.52-1.04); for ≥85.7
use of menopausal mL, OR = 0.61 (0.40-
hormones, body size, 0.93)
caloric intake, smoking,
alcohol, coffee, black
tea intake, family his-

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

tory of breast cancer,
physical activity, educa-
tion, soy intake, and
dark green vegetable
intake

OR = odds ratio; CI = confidence interval.

 Table 4. Cohort Studies That Prospectively Examine Recurrence of Breast Cancer With Respect to Stage in Relation to Green Tea Consumption
Adjusted RR (95% CI)
Confounding for Different Stage
Green Tea Consumption Factors Accounted Cancers According to Dose-Response
Reference n (Response Rate) Population Profile Levels Analyzed and Adjusted For Green Tea Consumption Relationship

39
Nakachi et al 472 female breast Pre- and postmeno- Average number of Age at surgery; clinical staging; Stage I and II: RR set as Stage I and II: not
cancer patients pausal Japanese cups consumed tumor volume; histological typing; 1.0 for <4 cups daily, enough data
(100%); 117 with women outpatients daily: <4 cups, ≥5 numbers of metastasized lymph RR = 0.56 (0.35-0.91) points to
stage I, 273 with with previously diag- cups nodes; blood vessel invasion; lym- for consumption of ≥5 assess; stage
stage II; 82 with nosed breast cancer phatic vessel invasion; expression cups; stage III: RR set III: not enough
stage III levels of estrogen and progester- as 1.0 for <4 cups daily, data points to
one receptor; use of chemother- RR = 1.88 (0.79-4.54) assess.
apy, radiotherapy, and tamoxifen; for ≥5 cups daily
height; weight; occupation; coffee
consumption; soy consumption;
nuts, fruits, green and yellow veg-
etables, seaweed, fish, meat,
eggs, and milk consumption; pas-
sive and direct exposure to ciga-
rette smoke, alcohol consumption;
experience of pregnancy, child-
birth, and abortion

Inoue et al40 1160 female breast Japanese women out- Average number of Age, year, season at first hospital Stage I: set at 1.0 for ≤2 Stage I: not pres-
cancer patients patients with an aver- cups consumed visit, family history of breast can- cups. For 3-5 cups, ent; stage II:
(97%); 751 with age age of 51.5 y daily: 0-2 cups, 3-5 cer, body mass index, physical RR = 0.37 (0.17-0.8); positive trend;
stage I, 257 with with previously diag- cups, ≥6 cups exercise, age at menarche, age at for ≥6 cups, RR = 0.59 stage III and IV:
stage II, 152 with nosed breast cancer pregnancy, parity, menopausal (0.23-1.52). Stage II: not present

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

stage III and IV status, tofu intake, fruit intake, and set at 1.0 for ≤2 cups.
coffee intake For 3-5 cups, RR = 0.8
(0.38-1.69); for ≥6 cups,
RR = 0.51 (0.18-1.46).
Stage III and IV: Set at
1.0 for ≤2 cups. For 3-5
cups, RR = 1.06 (0.51-
2.17); for ≥6 cups, RR =
0.87 (0.33-2.27)

RR = relative risk; CI = confidence interval.

149
 Seely et al

Key et al. (1999) (37) 0.86 (0.62, 1.21)

Nagano et al. (2001) (36) 1.00 (0.67, 1.60)

Suzuki et al. (2004) (38) 0.84 (0.57, 1.24)

combined 0.89 (0.71, 1.10)

0.5 1 2
relative risk (95% confidence interval)

Figure 2 Meta-analysis of cohort studies examining the risk of breast cancer incidence according to highest consumption of green tea
levels (see Table 2).

0.61 (0.40, 0.93)

Wu et al. (2003) (35)

0.17 (0.03, 1.04)

Tao et al. (2002) (30)

combined 0.44 (0.14, 1.31)

0.01 0.1 0.2 0.5 1 2

odds ratio (95% confidence interval)

Figure 3 Meta-analysis of case-control studies examining the risk of breast cancer incidence according to highest consumption levels of
green tea (see Table 3).

RR of stage I and II breast cancer recurrence in
included cohort studies is 0.56 (95% CI, 0.38-0.83; P =
2
.004; I = 0%; P = .9; see Figure 5) and 1.31 (95% CI,
2
0.62-2.79; P = .3; I = 23%; P = .2) for stage III and IV
breast cancer recurrence in cohort studies (z test, P =
.049).
With respect to a dose-response relationship
between green tea consumption and risk reduction

150 INTEGRATIVE CANCER THERAPIES 4(2); 2005

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

 Green Tea and Breast Cancer: Systematic Review

Nakachi et al. (1998) – Stage I and II (39) 0.56 (0.35, 0.91)

Nakachi et al. (1998) – Stage III (39) 1.88 (0.79, 4.54)

Inoue et al. (2001) – Stage I (40) 0.59 (0.23, 1.52)

Inoue et al. (2001) – Stage II (40) 0.51 (0.18, 1.46)

Inoue et al. (2001) – Stage III and IV (40) 0.87 (0.33, 2.27)

combined 0.75 (0.47, 1.19)

0.1 0.2 0.5 1 2 5

relative risk (95% confidence interval)

Figure 4 Meta-analysis of risk of recurrence of all-stage breast cancer in cohort studies according to highest consumption levels of green
tea (see Table 4).

Nakachi et al. (1998) – Stage I and II (39) 0.56 (0.35, 0.91)

Inoue et al. (2001) – Stage I (40) 0.59 (0.23, 1.52)

Inoue et al. (2001) – Stage II (40) 0.51 (0.18, 1.46)

combined 0.56 (0.38, 0.83)

0.1 0.2 0.5 1 2

relative risk (95% confidence interval)

Figure 5 Meta-analysis of early-stage (I and II) breast cancer recurrence in cohort studies according to highest consumption levels of
green tea (see Table 4).

for both breast cancer development and recurrence, 3 cohort studies did not show a dose-response trend
only 3 of the 7 studies demonstrated a positive rela- (Table 2); however, both the case-control studies did
tionship between increasing dose and reduced risk. In demonstrate a positive dose-response relationship
the studies examining prevention of breast cancer, the (Table 3). For the 2 studies looking at risk reduction of

INTEGRATIVE CANCER THERAPIES 4(2); 2005 151

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014

 Seely et al
breast cancer recurrence, only 1 study demonstrated a
dose-response relationship, and this was only for stage
40
II disease (Table 4). In the same study by Inoue et al,
analysis of stage I disease demonstrated an inverse
relationship, with the greater risk reduction being
observed at a consumption level of 3 to 5 cups daily
rather than at 6 or more cups daily (Table 4).
Discussion
This systematic review and meta-analysis offers tenta-
tive support that high levels of green tea consumption
are positively correlated with prevention of recur-
rence of early-stage breast cancer but not in the case of
more serious disease. The results of our meta-analysis
should be interpreted with caution considering the
small number of observational studies available. How-
ever, the combined data from all the studies do pro-
vide enough power to validate the meta-analysis. The
pooled analysis does not definitively support the use of
green tea in primary prevention of breast cancer; how-
ever, a trend toward risk reduction is demonstrated in
nearly all the case-control and cohort studies.
There are several limitations to be considered in
our meta-analysis. Primarily, no randomized con-
trolled trials have been conducted to date, and only
observational studies are available for this review. A
principal limitation of the observational studies is the
variability of response rate to questionnaires concern-
ing green tea intake. Questionnaire response rate in
the cohort studies was highly variable, ranging from
62% to 97%. In the case-control study by Tao et al30 and
the cohort studies by Key et al37 and Inoue et al,40 the
authors did not account for the important confound-
ing factors of smoking and alcohol consumption.
These confounders may have skewed the results, as
they are known risk factors in the development of
breast cancer. In addition, the cohort study by Inoue
et al40 did not report controlling for the primary treat-
ment the women had undergone for their breast can-
cer. In all studies, the confounding factors as listed in
Tables 2, 3, and 4 were adjusted for in the determina-
tion of ORs and RRs. Clear examples include the stud-
ies by Key et al and Inoue et al with respect to smoking
and alcohol. The major threats to validity in observa-
tional studies are the potential for confounding and
selection bias that could distort the findings. The sta-
tistical combination of these results could therefore
produce precise yet unreliable results.
41
Case-control studies are a lower level of evidence
than cohort studies and may provide spurious results
due to selection bias. In our meta-analysis, the case-
control studies gave a greater pooled difference than
the cohort studies did. The results, however, did not
achieve significance when using a random effects
152
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
model. In a post hoc analysis applying a fixed-effects
model, we observed a statistically significant RR of 0.57
(95% CI, 0.38-0.86; P = .007; I2 = 47%). The larger
number of outcomes in the study by Tao et al rather
30
than a smaller number contribute to this limitation, a
paradoxical outcome of using the random effects
model.
In the case of the 2 studies assessing risk for breast
cancer recurrence, the 2 populations have a high
degree of homogeneity. Both cohort groups are from
urban Japanese hospital outpatient populations and
are relatively similar in age range, history of exposure
to ionizing radiation, potential biases due to study
design, and geographic location. Ultimately, it is
unknown if different populations will have a different
response to green tea. Therefore, given the similarity
of the populations from the 2 studies assessing risk of
breast cancer recurrence, the pooled data are viewed
with a slightly greater degree of confidence.
The potential for publication bias must also be con-
sidered in every systematic review, especially among
42
observational studies. Eggers test and Kendall’s τ had
too small a sample size to conduct a robust evaluation,
yet our search strategy was very broad in terms of the
multiple sources accessed as well as the low threshold
for retrieval with no language restrictions. Of greatest
strength is the inclusion of the CHKD database, a Chi-
nese database encompassing Asian trials that are often
excluded from indexing in Western databases. In this
case, we identified 1 additional Chinese language
cohort study examining breast cancer incidence in
30
relation to green tea consumption.
Our subgroup analysis indicates that a high intake
of green tea may be associated with a relative risk
reduction in stage I and II breast cancer recurrence.
However, subgroup analyses should be interpreted
with caution, and several criteria should be applied in
determining the credibility of the within-study com-
43
parisons. Most important, we made the a priori
hypothesis to assess cancer stage, thus decreasing the
likelihood of a type I error. Overall, we found that the
magnitude of the protective effect is larger in stage I
and II cancer in comparison to the stage III and IV
cancer, with a difference that is statistically significant
(P = .049).
Because of the potential inhibitory effect of green
6-8
tea on angiogenesis and metastasis, there is some
rationale and biological plausibility that green tea is
more effective in preventing recurrence in less severe
stages of the disease. Given the fact that the presence
of angiogenesis and metastasis is much rarer, by defini-
tion, in early-stage cancer, green tea might effect a
reduction in recurrence in the early stages of disease,
more so than at a later stage. If metastatic spread has
INTEGRATIVE CANCER THERAPIES 4(2); 2005
 Green Tea and Breast Cancer: Systematic Review
already taken place, the likelihood of a subtle molecu-
lar agent having any real effect is remote.
Recently, investigators conducted a systematic
review of the evidence for green tea in risk reduction
of gastrointestinal cancer.44 As is the case for green tea
and breast cancer, the investigators did not identify
any randomized controlled trials. Findings of the
gastrointestinal/green tea review demonstrate that
green tea is associated with reduced adenomatous
polyp formation and reduced risk of chronic atrophic
gastritis. With respect to the development of stomach
and intestinal cancer, however, the authors found that
there is no supportive evidence for green tea con-
sumption as a preventative agent.44
If there actually is a protective effect, it is difficult to
determine what the optimal dose of green tea might
be since the methods of tea preparation differ from
person to person and study to study. Variables in prep-
aration include but are not limited to steeping time,
amount of tea leaves used, the reusing of tea leaves for
more than 1 cup of tea, cup size, quality of the tea,
freshness of the leaf, and steeping temperature. While
most of the studies did not exhibit a dose-response
relationship, the optimal effect was achieved at the
highest dose in all cases except for stage I breast can-
cer in the study by Inoue et al.40 Green tea consump-
tion was measured in a number of ways, with the pre-
dominant measure being number of cups. Nearly all
the studies indicate greatest risk reduction at con-
sumption levels equal to or more than 5 cups a day. We
roughly surmise that any protective effect is most likely
to result from exposure to green tea at consumption
levels greater than or equal to 5 cups a day.
The results of this review should be of interest as
green tea is widely used, has few documented adverse
effects, and is relatively inexpensive. An important
concern, however, is the possibility for green tea to
interact negatively with conventional cancer treat-
ment. One pharmacokinetic clinical trial examining
green tea’s impact on cytochrome P450 3A4 found no
significant interaction with the probe drugs
dextromethorphan and alprazolam.45 Green tea has
antioxidant properties, and as such, there is a concern
that it may interfere with the activity of chemotherapy
and radiation therapy.46-49 However, there is also evi-
dence that some antioxidants can enhance the effect
of chemotherapeutic agents.50-52 With regard to green
tea, a single in vitro study implies that green tea may
promote tumor survival pathways associated with
hypoxia-inducible factor–1 that are targeted by some
chemotherapies.53
Another consideration is the potential for antioxi-
dants and specifically green tea to reduce the
cytotoxicity of chemotherapy.54 Evidence also exists to
INTEGRATIVE CANCER THERAPIES 4(2); 2005
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
support green tea and specifically the catechin EGCG
as reducing the negative effects of ionizing radiation
55
without curtailing benefit.
In summary, our systematic review and meta-analysis
did not find a significant effect of green tea on breast
cancer prevention. Green tea consumption, however,
may be associated with a reduced risk of recurrence of
stage I and II breast cancer. Given the encouraging
findings of this review for prevention of breast cancer
recurrence, randomized controlled trials are war-
ranted to show any true causality.
In women recovering from breast cancer who are
receiving conventional treatment, green tea should be
used with some caution as the interactions of green tea
with standard treatment such as chemotherapy or hor-
monal therapy have not been fully defined. Investiga-
tors should consider organizing, and granting agen-
cies supporting, both clinical trials examining
interactions between green tea and conventional
treatments for breast cancer and randomized trials
looking at the impact of green tea on early breast
cancer recurrence.
Acknowledgment
This study was supported by the Lotte and John Hecht
Memorial Foundation.
References
1. Boon H, Stewart M, Kennard MA, et al. Use of complemen-
tary/alternative medicine by breast cancer survivors in
Ontario: prevalence and perceptions. J Clin Oncol.
2000;18:2515-2521.
2. Yang CS. Tea and health. Nutrition. 1999;15:946-949.
3. Wiseman SA, Balentine DA, Frei B. Antioxidants in tea. Crit Rev
Food Sci Nutr. 1997;37:705-718.
4. Rietveld A, Wiseman S. Antioxidant effects of tea: evidence
from human clinical trials. J Nutr. 2003;133:3285S-3292S.
5. Frei B, Hidgon JV. Antioxidant activity of tea polyphenols in
vivo: evidence from animal studies. J Nutr. 2003;133:3275S-
3284S.
6. Sartippour MR, Shao ZM, Heber D, et al. Green tea inhibits
vascular endothelial growth factor (VEGF) induction in
human breast cancer cells. J Nutr. 2002;132:2307-2311.
7. Sartor L, Pezzato E, Dell’Aica I, Caniato R, Biggin S, Garbisa S.
Inhibition of matrix-proteases by polyphenols: chemical
insights for anti-inflammatory and anti-invasion drug design.
Biochem Pharmacol. 2002;64:229-237.
8. Garbisa S, Biggin S, Cavallarin N, Sartor L, Benelli R, Albini A.
Tumor invasion: molecular shears blunted by green tea. Nat
Med. 1999;5:1216.
9. Tang FY, Nguyen N, Meydani M. Green tea catechins inhibit
VEGF-induced angiogenesis in vitro through suppression of
VE-cadherin phosphorylation and inactivation of Akt mole-
cule. Int J Cancer. 2003;106:871-878.
10. Tang FY, Meydani M. Green tea catechins and vitamin E inhibit
angiogenesis of human microvascular endothelial cells
through suppression of IL-8 production. Nutr Cancer.
2001;41:119-125.
11. Liang YC, Lin-shiau SY, Chen CF, Lin JK. Suppression of
extracellular signals and cell proliferation through EGF
153
 Seely et al
receptor binding by (-)-epigallocatechin gallate in human
A431 epidermoid carcinoma cells. J Cell Biochem. 1997;67:55-
65.
12. Lin JK, Liang YC, Lin-Shiau SY. Cancer chemoprevention by
tea polyphenols through mitotic signal transduction blockade.
Biochem Pharmacol. 1999;58:911-915.
13. Wu YN, Wang HZ, Li JS, Han C. The inhibitory effect of Chi-
nese tea and its polyphenols on in vitro and in vivo N-
nitrosation. Biomed Environ Sci. 1993;6:237-258.
14. Naasani I, Seimiya H, Tsuruo T. Telomerase inhibition,
telomere shortening, and senescence of cancer cells by tea
catechins. Biochem Biophys Res Commun. 1998;249:391-396.
15. Trosko JE, Chang CC. Mechanism of up-regulated gap
j un ct ion a l in t ercellula r commun ica t ion durin g
chemoprevention and chemotherapy of cancer. Mutat Res.
2001;480-481:219-229.
16. Sai K, Kanno J, Hasegawa R, Trosko JE, Inoue T. Prevention of
the down-regulation of gap junctional intercellular communi-
cation by green tea in the liver of mice fed pentachlorophenol.
Carcinogenesis. 2000;21:1671-1676.
17. Sigler K, Ruch RJ. Enhancement of gap junctional
intercellular communication in tumor promoter-treated cells
by components of green tea. Cancer Lett. 1993;69:15-19.
18. Nagata C, Kabuto M, Shimizu H. Association of coffee, green
tea, and caffeine intakes with serum concentrations of
es t ra diol a n d s ex h ormon e-bin din g globulin in
premenopausal Japanese women. Nutr Cancer. 1998;30:21-24.
19. Komori A, Yatsunami J, Okabe S, et al. Anticarcinogenic activ-
ity of green tea polyphenols. Jpn J Clin Oncol. 1993;23:186-190.
20. Ahmad N, Feyes DK, Nieminen AL, Agarwal R, Mukhtar H.
Green tea constituent epigallocatechin-3-gallate and induc-
tion of apoptosis and cell cycle arrest in human carcinoma
cells. J Natl Cancer Inst. 1997;89:1881-1886.
21. Hibasami H, Achiwa Y, Fujikawa T, Komiya T. Induction of pro-
grammed cell death (apoptosis) in human lymphoid leukemia
cells by catechin compounds. Anticancer Res. 1996;16:1943-
1946.
22. Okabe S, Suganuma M, Hayashi M, Sueoka E, Komori A, Fujiki
H. Mechanisms of growth inhibition of human lung cancer
cell line, PC-9, by tea polyphenols. Jpn J Cancer Res. 1997;88:639-
643.
23. Yang GY, Liao J, Kim K, Yurkow EJ, Yang CS. Inhibition of
growth and induction of apoptosis in human cancer cell lines
by tea polyphenols. Carcinogenesis. 1998;19:611-616.
24. Hibasami H, Komiya T, Achiwa Y, et al. Induction of apoptosis
in human stomach cancer cells by green tea catechins. Oncol
Rep. 1998;5:527-529.
25. Tsubono Y, Nishino Y, Komatsu S, et al. Green tea and the risk
of gastric cancer in Japan. N Engl J Med. 2001;344:632-636.
26. Hoshiyama Y, Kawaguchi T, Miura Y, et al. A prospective study
of stomach cancer death in relation to green tea consumption
in Japan. Br J Cancer. 2002;87:309-313.
27. Bushman JL. Green tea and cancer in humans: a review of the
literature. Nutr Cancer. 1998;31:151-159.
28. Blot WJ, Chow WH, McLaughlin JK. Tea and cancer: a review
of the epidemiological evidence. Eur J Cancer Prev. 1996;5:425-
438.
29. Zhang M, Binns CW, Lee AH. Tea consumption and ovarian
cancer risk: a case-control study in China. Cancer Epidemiol
Biomarkers Prev. 2002;11:713-718.
30. Tao M, Liu D, Gao L, Jin F. Association between green tea
drinking and breast cancer risk [in Chinese]. Tumor.
2002;22(3):11-15.
31. Higgins JP, Thompson SG. Quantifying heterogeneity in a
meta-analysis. Stat Med. 2002;21:1539-1558.
32. Arts IC, Jacobs DR Jr, Gross M, Harnack LJ, Folsom AR. Dietary
catechins and cancer incidence among postmenopausal
154
Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014
women: the Iowa Women’s Health Study (United States). Can-
cer Causes Control. 2002;13:373-382.
33. Imai K, Suga K, Nakachi K. Cancer-preventive effects of drink-
ing green tea among a Japanese population. Prev Med.
1997;26:769-775.
34. Wu AH, Tseng CC, Van Den Berg D, Yu MC. Tea intake, COMT
genotype, and breast cancer in Asian-American women. Cancer
Res. 2003;63:7526-7529.
35. Wu AH, Yu MC, Tseng CC, Hankin J, Pike MC. Green tea and
risk of breast cancer in Asian Americans. Int J Cancer.
2003;106:574-579.
36. Nagano J, Kono S, Preston DL, Mabuchi K. A prospective study
of green tea consumption and cancer incidence, Hiroshima
and Nagasaki (Japan). Cancer Causes Control. 2001;12:501-508.
37. Key TJ, Sharp GB, Appleby PN, et al. Soya foods and breast can-
cer risk: a prospective study in Hiroshima and Nagasaki, Japan.
Br J Cancer. 1999;81:1248-1256.
38. Suzuki Y, Tsubono Y, Nakaya N, Koizumi Y, Tsuji I. Green tea
and the risk of breast cancer: pooled analysis of two prospec-
tive studies in Japan. Br J Cancer. 2004;90:1361-1363.
39. Nakachi K, Matsuyama S, Miyake S, Suganuma M, Imai K. Pre-
ventive effects of drinking green tea on cancer and cardiovas-
cular disease: epidemiological evidence for multiple targeting
prevention. Biofactors. 2000;13(1-4):49-54.
40. Inoue M, Tajima K, Mizutani M, et al. Regular consumption of
green tea and the risk of breast cancer recurrence: follow-up
study from the Hospital-based Epidemiologic Research Pro-
gram at Aichi Cancer Center (HERPACC), Japan. Cancer Lett.
2001;167:175-182.
41. Egger M, Smith GD, Schneider M. Systematic reviews of obser-
vational studies. In: Altman D, ed. Systematic Reviews in
Healthcare. London, UK: BMJ; 2003:211-227.
42. Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR. Publica-
tion bias in clinical research. Lancet. 1991;337:867-872.
43. Oxman A, Guyatt GH. When to believe a subgroup analysis. In:
Guyatt GH, Rennie D, eds. The Users’ Guides to the Medical Litera-
ture: A Manual for Evidence-Based Clinical Practice. Chicago, Ill:
AMA Press; 2002:553-565.
44. Borrelli F, Capasso R, Russo A, Ernst E. Systematic review:
green tea and gastrointestinal cancer risk. Aliment Pharmacol
Ther. 2004;19:497-510.
45. Donovan JL, Devane CL, Chavin KD, et al. Green tea (Camellia
sinensis) extract does not alter cytochrome P-450 3A4 or 2D6
activity in healthy volunteers. Drug Metab Dispos. 2004;32:906-
908.
46. McCune JS, Hatfield AJ, Blackburn AA, Leith PO, Livingston
RB, Ellis GK. Potential of chemotherapy-herb interactions in
adult cancer patients. Support Care Cancer. 2004;12:454-462.
47. Labriola D, Livingston R. Possible interactions between dietary
antioxidants and chemotherapy. Oncology (Huntingt).
1999;13:1003-1008.
48. Lamson DW, Brignall MS. Antioxidants in cancer therapy:
their actions and interactions with oncologic therapies. Altern
Med Rev. 1999;4:304-329.
49. Whiteside MA, Heimburger DC, Johanning GL. Micronutri-
ents and cancer therapy. Nutr Rev. 2004;62:142-147.
50. Ohishi T, Kishimoto Y, Miura N, et al. Synergistic effects of (-)-
epigallocatechin gallate with sulindac against colon
carcinogenesis of rats treated with azoxymethane. Cancer Lett.
2002;177:49-56.
51. Prasad KN, Kumar R. Effect of individual and multiple antioxi-
dant vitamins on growth and morphology of human
nontumorigenic and tumorigenic parotid acinar cells in cul-
ture. Nutr Cancer. 1996;26:11-19.
52. Prasad KN, Kumar A, Kochupillai V, Cole WC. High doses of
multiple antioxidant vitamins: essential ingredients in
INTEGRATIVE CANCER THERAPIES 4(2); 2005
 Green Tea and Breast Cancer: Systematic Review

improving the efficacy of standard cancer therapy. J Am Coll
Nutr. 1999;18:13-25.
53. Zhou YD, Kim YP, Li XC, et al. Hypoxia-inducible factor-1 acti-
vation by (-)-epicatechin gallate: potential adverse effects of
cancer chemoprevention with high-dose green tea extracts. J
Nat Prod. 2004;67:2063-2069.
54. Conklin KA. Dietary antioxidants during cancer chemother-
apy: impact on chemotherapeutic effectiveness and develop-
ment of side effects. Nutr Cancer. 2000;37:1-18.
55. Annabi B, Lee YT, Martel C, Pilorget A, Bahary JP, Beliveau R.
Radiation induced-tubulogenesis in endothelial cells is antag-
onized by the antiangiogenic properties of green tea
polyphenol (-) epigallocatechin-3-gallate. Cancer Biol Ther.
2003;2:642-649.

INTEGRATIVE CANCER THERAPIES 4(2); 2005 155

Downloaded from ict.sagepub.com at UNIV MASSACHUSETTS BOSTON on August 21, 2014