- Email: [email protected]
Green tea and liver cancer risk: A meta-analysis of prospective cohort studies in Asian populations
Accepted Manuscript Green tea and liver cancer risk: a meta-analysis of prospective cohort studies in Asian populations Ya-Qing Huang, M.D., Xin Lu, M.D., Han Min, M.D., Qian-Qian Wu, M.D., Xiao-Ting Shi, M.D., Kang-Qi Bian, M.D., Xiao-Ping Zou, M.D., Ph.D. PII:
S0899-9007(15)00237-3
DOI:
10.1016/j.nut.2015.05.021
Reference:
NUT 9546
To appear in:
Nutrition
Received Date: 9 February 2015 Revised Date:
13 May 2015
Accepted Date: 20 May 2015
Please cite this article as: Huang Y-Q, Lu X, Min H, Wu Q-Q, Shi X-T, Bian K-Q, Zou X-P, Green tea and liver cancer risk: a meta-analysis of prospective cohort studies in Asian populations, Nutrition (2015), doi: 10.1016/j.nut.2015.05.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Green tea and liver cancer risk: a meta-analysis of prospective cohort studies in Asian
2
populations
3
Ya-Qing Huang M.D. a, Xin Lu M.D. b, Han Min M.D. a, Qian-Qian Wu M.D. a,
5
Xiao-Ting Shi M.D. a, Kang-Qi Bian M.D. a, Xiao-Ping Zou M.D., Ph.D. a, *
RI PT
4
6 7
a
8
Nanjing Medical University, Nanjing, China
9
b
Department of Cardiology, the First Affiliated Hospital of Nanjing Medical
SC
10
Department of Gastroenterology, Nanjing Drum Tower Hospital Clinical College of
University, Nanjing, China
M AN U
11 12
*corresponding author. Tel./fax: +86 13770771661.
13
E-mail address: [email protected] (X-P. Zou)
14
Y-QH and X-PZ designed the research. Y-QH and XL performed the literature
16
research and extracted the data. HM contributed analytical tools. Y-QH, X-TS and
17
Q-QW analyzed the data. All of the authors helped interpret the results and write and
18
revise the manuscript. The authors declare that they have no competing interests.
19
EP
TE D
15
Key words: Green tea
21
populations
22 23 24 25
Liver cancer
Meta-analysis
AC C
20
26 27 28 29 30 1
Cohort studies
Asian
ACCEPTED MANUSCRIPT Abstract
2
Objectives: The aim of this meta-analysis was to investigate whether an association
3
existed between green tea consumption and the risk of liver cancer in prospective
4
cohort studies in Asian populations.
5
Methods: Relevant studies were identified by searching PUBMED, EMBASE, ISI
6
web of science and Chinese Bio-medicine Database published before April 2015.
7
Study-specific risk estimates for the highest versus non/lowest and increment of 1
8
cup/day green tea consumption levels were combined based on fixed or random
9
effects models. STATA 11.0 software was used for statistical analysis.
SC
RI PT
1
Results: A total of nine prospective cohort articles involving 465,274 participants and
11
3,694 cases of liver cancer from China, Japan and Singapore were included. The
12
summary relative risk (RR) indicated a significant association between the highest
13
green tea consumption and reduced risk of liver cancer [summary RR: 0.88; 95%
14
confidence interval (CI): 0.81, 0.97]. However, no statistically significant association
15
was observed through the analysis of an increment of 1 cup/day (summary RR: 0.97;
16
95% CI: 0.95, 1.00). When stratified analysis by gender, the protective effect of green
17
tea consumption on liver cancer risk was observed in women group (summary RR:
18
0.78; 95% CI: 0.64, 0.96), but not in men group (summary RR: 0.89; 95% CI: 0.79,
19
1.00).
20
Conclusions: The present analysis indicated the preventive effects of green tea intake
21
on the risk of liver cancer in Asian women populations. However, additional studies
22
are needed to make a convincing case for this association.
24 25
TE D
EP
AC C
23
M AN U
10
26 27 28 29 30 2
ACCEPTED MANUSCRIPT 1
Introduction
2
Primary liver cancer is the sixth most common cancer and, owing to its poor
3
prognosis, the second most common cause of cancer death [1]. An estimated 782,000
4
new primary liver cancer cases and 745,000 deaths occurred worldwide in 2012. In
5
addition
6
infections, which are major aetiologic determinants, other possible environmental risk
7
factors such as dietary aflatoxin, excessive alcohol intake, cigarette smoking, obesity
8
and diabetes play a key role in the development of liver cancer [2].
hepatitis
B
virus
(HBV)/hepatitis
C
virus
(HCV)
RI PT
to
Brewed from the plant Camellia sinensis, tea is one of the most popular beverages
10
consumed around the world. According to the level of fermentation, tea can be mainly
11
classified as three major commercial types: green tea (non-fermented), black tea
12
(fermented) and oolong tea (half-fermented). Among them, green tea contains a much
13
higher level of polyphenols known as catechins, which have been shown to have
14
antimutagenic, antigenotoxic, and anticarcinogenic activities [3]. A major polyphenol
15
of green tea, epigallocatechin-3-gallate (EGCG), has generated interest for its
16
anti-tumor effects [4]. Numerous animal and in vitro experiments have been carried
17
out to support the possibility that green tea has preventive effects against liver cancer
18
[5, 6]. EGCG, as the major polyphenol of green tea, is proven to inhibit the growth of
19
all liver cancer cell lines [7]. However, the molecular mechanisms of the
20
chemopreventive effects of green tea are still uncertain.
EP
TE D
M AN U
SC
9
Over the last two decades, several epidemiological studies have been conducted to
22
investigate the association between green tea consumption and liver cancer risk, but
23
with varying results. One review [8] declared very limited support for green tea
24
consumption in reducing the risk of liver cancer. A previous meta-analysis [9]
25
suggested that high green tea intake might be associated with lower risk of liver
26
cancer, resulting from the combination of cohort and case-control studies. However, it
27
was limited because true differences in the level and range of intake between studies
28
were not taken into account. Considering the results of subsequent published cohort
29
studies tended to show a weak correlation between green tea intake and liver cancer
30
risk, we therefore conducted this exploratory meta-analysis to update and summarize
AC C
21
3
ACCEPTED MANUSCRIPT 1
the results from prospective studies for deriving a more precise estimation of this
2
topic and to further evaluate whether there is a dose-response relationship between
3
green tea intake and the risk for liver cancer.
4
Materials and methods
6
Data sources and searches
7
Pertinent studies were identified by a computerized Medline, EMBASE, ISI web of
8
science and Chinese Bio-medicine Database search that included the years 1966
9
through April 2015. We used the search terms “green tea,” “tea,” “polyphenol,” or
10
“catechin” combined with “liver neoplasm,” “liver cancer,” and “hepatocellular
11
carcinoma” in the full-text option. The search was done on studies conducted on
12
human species, without restriction on language. The reference lists of reviews and
13
retrieved articles were handsearched at the same time. The titles and abstracts were
14
scanned to exclude any studies that were clearly irrelevant. The full-texts and tables of
15
the remaining articles were retrieved and perused to determine the relevancy of the
16
study design and data, according to the inclusion criteria detailed below. Furthermore,
17
references from the retrieved articles were reviewed to make sure that all the relevant
18
bibliographies published were reviewed. Two authors (Y-QH and XL) conducted all
19
searches independently.
SC
M AN U
TE D
EP
20
22
Inclusion and exclusion criteria Following criteria were used to identify relevant studies for the meta-analysis: (1)
AC C
21
RI PT
5
23
participants were from Asian countries; (2) the exposure of interest was green tea
24
consumption; (3) the outcome of interest was total liver cancer incidence or mortality;
25
(4) RR or hazard ratio (HR) with their 95% CI (or sufficient information to calculate
26
them) were reported; (5) prospective cohort study design; and (6) some adjustments
27
were made for potential liver cancer risk. If cohorts were duplicated in more than one
28
study, only the most recent and complete publication was included in the
29
meta-analysis.
30
Our literature search identified 10 potentially relevant articles [10-19] concerning 4
ACCEPTED MANUSCRIPT 1
green tea consumption and liver cancer risk. One publication [10] was excluded due
2
to insufficient information on numbers of cases, controls and RRs with their
3
corresponding 95% CIs. Thus, the remaining 9 studies were included in the
4
meta-analysis.
6
RI PT
5
Data extraction
For each study, we extracted the following data: the name of first author, year of
8
publication, country where the study was conducted, recruitment time, participant age,
9
sex, years of follow-up, number of participants (cases and cohort size), type of
10
outcome (primary liver cancer or HCC, incidence or mortality), the RR estimates with
11
corresponding 95% CIs for each category of green tea consumption and adjustment
12
factors. Two reviewers (Y-QH and HM) independently inspected the data for
13
eligibility and quality. The results were compared and any questions or discrepancies
14
were resolved by a third reviewer (Q-QW). For example, if the data were insufficient
15
or missing, we attempted to contact the authors of the articles in order to request the
16
relevant data.
17 18
Statistical analysis
TE D
M AN U
SC
7
The measure of effect of interest was the RR and the corresponding 95% CI and the
20
HR was considered as RR directly. Study-specific RRs and 95% CIs for highest
21
versus non/lowest green tea consumption level were extracted from each study, and
22
then logRRs were weighted by the inverse of their variances to obtain summary RRs
23
and 95% CIs. If the study provided separate RR estimates for men and women, we
24
treated them as two different studies in statistical analysis. Studies were combined by
25
using the DerSimonian and Laird random-effects model, which incorporates both
26
within- and between-study variability [20].
AC C
EP
19
27
Statistical heterogeneity among studies was evaluated using the Q and I2 statistics
28
[21]; P<0.10 and I2>50% were considered statistically significant [22]. When there is
29
significant heterogeneity among study results, the random effect model was used to
30
calculate summary RR while the fixed effect model was used to calculate summary 5
ACCEPTED MANUSCRIPT 1
RR among studies with homogeneous results. In addition, subgroup analysis was
2
carried out to investigate the influence of geographic area (China versus Japan),
3
gender, the number of liver cancer cases (<200 vs. ≥200) and adjusted factor of
4
alcohol drinking. To normalize the variation between studies for the difference in exposure categories,
6
dose-response analysis was conducted to estimate study-specific slopes from the
7
natural logarithm of the RR across categories of green tea consumption, using the
8
method described by Greenland and Longnecher [23] and Orsini et al [24]. Because
9
this method requires the risk estimates with their variance estimates for three or more
10
quantitative exposure categories, so studies with only 2 categories [11,15,18,19] were
11
not included in this analysis. In studies [13,16] that did not provide the number of
12
cases and controls in each exposure category, the variance-weighted least-squares
13
regression model was used to estimate the slopes. Midpoint of upper and lower
14
boundaries was considered as the dose of each category if the study reported only the
15
range of green tea consumption. Because the highest category of consumption was
16
usually open, we considered it of the same amplitude as the preceding category. Then
17
summary RR estimates were obtained by pooling the study-specific slopes and taking
18
the inverse of their variances as weighted.
TE D
M AN U
SC
RI PT
5
Furthermore, Begg’s funnel plot and Egger’s regression test were used to evaluate
20
whether publication bias might affect the statistical results [25,26]. The funnel plot
21
would be asymmetrical if publication bias existed and a P value of less than 0.1 was
22
considered representative of significant statistical publication bias. All statistical
23
analysis was performed by the STATA version 11.0 software (Stata Corporation,
24
College Station, TX). All statistical tests were two-sided.
AC C
25
EP
19
26
Results
27
Search results and study characteristics
28
The details of literature search are shown in Figure 1. A total of nine articles and
29
eleven studies involving 465,274 participants and 3,694 cases of liver cancer were
30
included in the meta-analysis. And the characteristics of the included studies are listed 6
ACCEPTED MANUSCRIPT 1
in Table 1. These studies were published between 2000 and 2012. Seven of the eleven
2
studies were conducted in Japan [11-14,16,17], three in China [15,19], and the
3
remaining one in Singapore [18].
4
Highest versus non/lowest drinkers
RI PT
5
Relative risk estimates of liver cancer for highest green tea consumption compared
7
with non/lowest green tea consumption for individual studies and all studies
8
combined were shown in Figure 2. Overall, we found that highest green tea
9
consumption was statistically significantly associated with reduced risk of liver cancer
10
(summary RR: 0.88; 95% CI: 0.81, 0.97). No statistically significant heterogeneity
11
was found among the studies (Pheterogeneity=0.199, I2=25.7%). In a subgroup analysis
12
conducted by gender, only five studies were included in the analysis. The overall
13
results in women showed a significant 22% reduction in liver cancer risk with high
14
intake of green tea (summary RR: 0.78; 95% CI: 0.64, 0.96; Pheterogeneity=0.193;
15
I2=34.2%). No such effect was noted in men (summary RR: 0.89; 95% CI: 0.79, 1.00;
16
Pheterogeneity=0.35; I2=8.7%) (Figure 3, Table 2). We also stratified the various studies
17
by study country, liver cancer cases, outcome and adjustment for alcohol consumption
18
(Table 2). Similar but with borderline statistically significant association was found
19
for all these stratified analyses.
EP
TE D
M AN U
SC
6
20
22
Dose-response analysis
Given the wide array of measurement categories reported among studies, we also
AC C
21
23
conduct dose-response analysis. However, the data did not show any inverse
24
relationship between an increase in green tea consumption of 1 cup/day and the liver
25
cancer risk (summary RR: 0.97; 95% CI: 0.95, 1.00; Pheterogeneity=0.203; I2=24.5%)
26
(Table 2).
27 28
Publication bias
29
There was no indication of publication bias in the reporting of results on green tea
30
consumption and liver cancer risk, from either visualization of the Begg’s funnel plot 7
ACCEPTED MANUSCRIPT 1
(P=0.27) or the Eggers’s test (P=0.38) (Figure 4).
2 3
Discussion To our knowledge, the present meta-analysis is the first quantitative systematic
5
analysis of the association between green tea consumption and liver cancer risk based
6
on published prospective cohort studies in Asian populations. The results suggest that
7
high green tea consumption is associated with a decreased risk of liver cancer in Asian
8
women populations. However, no statistically significant association was observed
9
through the analysis of an increment of 1 cup/day. This result is consistent with a
10
recent dose–response meta-analysis [27] of 3 prospective studies, in which no inverse
11
association (RR: 0.93; 95% CI: 0.75, 1.17) was found between an increase in green
12
tea consumption of 3 cup/day and risk of liver cancer.
M AN U
SC
RI PT
4
Although the exact mechanisms are unclear, a protective role of green tea in
14
digestive system cancer prevention is biologically plausible. EGCG is the most
15
abundant polyphenol, accounting for 40% of the total polyphenol in green tea extract,
16
and is considered to be the most active constituent [28,29]. EGCG may reduce DNA
17
damage by reactive oxygen species [30,31] and inhibit tumor cell growth, invasion
18
and angiogenesis [32-35]. Several studies have reported that the incidence and
19
recurrence rate of animal tumors is significantly lower in animals receiving EGCG
20
than in control animals [36,37]. However, the effect of green tea on liver cancer is not
21
well understood. Although a few studies [38,39] linked it to hepatotoxicity, the
22
majority experimental studies have suggested green tea, especially green tea contracts,
23
have a hepatoprotective effect and anticarcinogenic potential in the liver [40].
EP
AC C
24
TE D
13
In this meta-analysis, a more obvious inverse association between green tea
25
consumption and liver cancer was found in women (RR=0.78) than in men (RR=0.89).
26
However, this result was derived from only five studies with a small number of cases,
27
so we could not draw a firm conclusion. Alcohol consumption is known to cause liver
28
cancer, and after adjustment for this, a significant inverse association was showed
29
between green tea consumption and liver cancer. Possible explanations may be due to
30
the anti-oxidative effects of tea polyphenols. Previous evidence showed green tea may 8
ACCEPTED MANUSCRIPT 1
protect against alcohol-related oxidative modification [41-43]. A history of liver disease may be a risk factor for liver cancer. Individuals who have
3
a history of liver disease are likely to have been infected with HBV or HCV [44]. It is
4
expected that the inverse association would be more pronounced among such
5
individuals [45, 46]. Nevertheless, in our study, only one article [17] discussed the
6
inverse association between green tea consumption and the risk of liver cancer among
7
individuals with/without a history of liver disease. A better understanding of this issue
8
will require further research.
RI PT
2
Our study has several potential limitations. Firstly, the possibility of bias and
10
confounding cannot be excluded in observational studies. Our meta-analysis enrolled
11
only cohort studies, which are less susceptible to bias because of the prospective
12
design. However, individual studies may have failed to adjust for potential known or
13
unknown confounders. Secondly, all of the studies included in the analysis had been
14
conducted in Asian populations (mainly China and Japan) because of the popularity of
15
green tea in East Asia. Therefore, the results should be cautiously extrapolated to
16
other populations. Finally, publication bias could be of concern because small studies
17
with null results tend not to be published. We had limited statistical power to
18
conclusively reject the null hypothesis of no publication bias because of the relative
19
small number of studies.
EP
TE D
M AN U
SC
9
In conclusion, our findings of this meta-analysis indicated an inverse association
21
between high green tea consumption and reduced liver cancer risk in Asian women
22
populations. Due to the limited available data, further large prospective studies should
23
be taken to confirm the results.
24 25
AC C
20
26 27 28 29 30 9
ACCEPTED MANUSCRIPT 1
References
2
[1] Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M et al. Cancer
3
incidence and mortality worldwide:sources, methods and major patterns in
4
GLOBOCAN 2012. Int J Cancer 2015; 136: E359-86.
6 7 8
[2] Yu MC, Yuan JM. Environmental factors and risk for hepatocellular carcinoma.
RI PT
5
Gastroenterology 2004; 127: S72-S78.
[3] Kuroda Y, Hara Y. Antimutagenic and anticarcinogenic activity of tea polyphenols. Mutat Res 1999; 436: 69-97.
[4] Landis-Piwowar KR, Huo C, Chen D, Milacic V, Shi G, Chan TH et al. A novel
10
prodrug of the green tea polyphenol (-)-epigallocatechin-3-gallate as a potential
11
anticancer agent. Cancer Res 2007; 67: 4303-4310 .
M AN U
12
SC
9
[5] Qin G, Gopalan-Kriczky P, Su J, Ning Y, Lotolikar PD. Inhibition of aflatoxin
13
B1-induced initiation of hepatocarcinogenesis in the rat by green tea. Cancer Lett
14
1997; 112: 149–154.
15
[6] Sakata R, Ueno T, Nakamura T, Sakamoto M, Torimura T, Sata M. Green tea polyphenol epigallocatechin-3-gallate inhibits platelet-derived growth
17
factor-induced proliferation of human hepatic stellate cell line LI90. J Hepatol
18
2004; 40: 52-59.
[7] Nishikawa T, Nakajima T, Moriguchi M, Jo M, Sekoguchi S, Ishii M et al. A green
EP
19
TE D
16
tea polyphenol, epigalocatechin-3-gallate, induces apoptosis of human
21
hepatocellular carcinoma, possibly through inhibition of Bcl-2 family proteins. J
22
Hepatol 2006; 44: 1074-1082.
23 24 25
AC C
20
[8] Yuan JM. Cancer prevention by green tea: evidence from epidemiologic studies. Am J Clin Nutr 2013; 98: 1676S-1681S.
[9] Fon Sing M, Yang WS, Gao S, Gao J, Xiang YB. Epidemiological studies of the
26
association between tea drinking and primary liver cancer: a meta-analysis. Eur J
27
Cancer Prev 2011; 20: 157-165.
28
[10] Inoue M, Yoshimi I, Sobue T, Tsugane S, JPHC Study Group. Influence of coffee
29
drinking on subsequent risk of hepatocellular carcinoma: a prospective study in
30
Japan. J Natl Cancer Inst 2005; 97: 293-300. 10
ACCEPTED MANUSCRIPT 1
[11] Nechuta S, Shu XO, Li HL, Yang G, Ji BT, Xiang YB et al. Prospective cohort
2
study of tea consumption and risk of digestive system cancers: results from the
3
Shanghai Women's Health Study. Am J Clin Nutr 2012; 96: 1056-1063.
4
[12] Johnson S, Koh WP, Wang R, Govindarajan S, Yu MC, Yuan JM. Coffee consumption and reduced risk of hepatocellular carcinoma: findings from the
6
Singapore Chinese Health Study. Cancer Causes Control 2011; 22: 503-510.
7
[13] Inoue M, Kurahashi N, Iwasaki M, Shimazu T, Tanaka Y, Mizokami M et al.
RI PT
5
Effect of coffee and green tea consumption on the risk of liver cancer: cohort
9
analysis by hepatitis virus infection status. Cancer Epidemiol Biomarkers Prev
11
2009; 18: 1746-1753.
[14] Ui A, Kuriyama S, Kakizaki M, Sone T, Nakaya N, Ohmori-Matsuda K et al.
M AN U
10
SC
8
12
Green tea consumption and the risk of liver cancer in Japan: the Ohsaki Cohort
13
study. Cancer Causes Control 2009; 20: 1939-1945.
14 15
[15] Wang N, Zheng Y, Jiang Q, Yu X, Chen Y. Tea and reduced liver cancer mortality. Epidemiology 2008; 19: 761.
[16] Iso H, Kubota Y. Nutrition and disease in the Japan Collaborative Cohort study
17
for evaluation of cancer (JACC). Asian Pac J Cancer Prev 2007; 8: 35-80.
18
TE D
16
[17] Shimazu T, Tsubono Y, Kuriyama S, Ohmori K, Koizumi Y, Nishino Y et al. Coffee consumption and the risk of primary liver cancer: pooled analysis of two
20
prospective studies in Japan. Int J Cancer 2005; 116: 150-154.
22 23 24 25
[18] Nagano J, Kono S, Preston DL, Mabuchi K. A prospective study of green tea consumption and cancer incidence, Hiroshima and Nagasaki (Japan).
AC C
21
EP
19
Cancer Causes Control 2001; 12: 501-508.
[19] Nakachi K, Matsuyama S, Miyake S, Suganuma M, Imai K. Preventive effects of drinking green tea on cancer and cardiovascular disease: epidemiological
26
evidence for multiple targeting prevention. Biofactors 2000; 13: 49-54.
27
[20] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
28 29 30
1986; 7: 177-188. [21] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557. 11
ACCEPTED MANUSCRIPT 1 2
[22] Hedges LV, Pigott TD. The power of statistical tests in meta-analysis. Psychol Methods 2001; 6: 203-217. [23] Greenland S, Longnecker MP. Methods for trend estimation from summarized
4
dose–response data, with applications to meta-analysis. Am J Epidemiol 1992;
5
135: 1301–1309.
RI PT
3
6
[24] Orsini N, Bellocco R, Greenland S. Generalized least squares for trend
7
estimation of summarized dose–response data. Stata J 2006; 6: 40–57.
8
[25] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629-634.
10
[26] Thornton A, Lee P. Publication bias in meta-analysis: its causes and
12
consequences. J Clin Epidemiol 2000; 53:207-216.
M AN U
11
SC
9
[27] Yu F, Jin Z, Jiang H, Xiang C, Tang J, Li T et al. Tea consumption and the risk of
13
five major cancers: a dose–response meta-analysis of prospective studies. BMC
14
Cancer 2014; 14:197.
15
[28] Fujiki H, Suganuma M, Okabe S, Komori A, Sueoka E, Sueoka N et al. Japanese green tea as a cancer preventive in humans. Nutr Rev 1996; 54: S67-S70.
17
[29] Ahmad N, Mukhtar H. Green tea polyphenols and cancer: biologic mechanisms
19 20
and practical implications. Nutr Rev 1999; 57: 78-83. [30] Higdon JV, Frei B. Tea catechins and polyphenols: health effects, metabolism,
EP
18
TE D
16
and antioxidant functions. Crit Rev Food Sci Nutr 2003; 43: 89-143. [31] Hakim IA, Harris RB, Brown S, Chow HH, Wiseman S, Agarwal S et al. Effect
22
of increased tea consumption on oxidative DNA damage among smokers: a
23 24 25
AC C
21
randomized controlled study. J Nutr 2003; 133: 3303S-3309S.
[32] Yamane T, Takahashi T, Kuwata K, Oya K, Inagake M, Kitao Y et al. Inhibition of N-methyl-N′-nitro-N-nitrosoguanidine-induced carcinogenesis by
26
(-)-epigallocatechin gallate in the rat glandular stomach. Cancer Res 1995; 55:
27
2081-2084.
28
[33] Ju J, Hong J, Zhou JN, Pan Z, Bose M, Liao J et al. Inhibition of intestinal
29
tumorigenesis in Apcmin/+ mice by (−)-epigallocatechin-3-gallate, the major
30
catechin in green tea. Cancer Res 2005; 65: 10623-10631. 12
ACCEPTED MANUSCRIPT 1
[34] Carter O, Dashwood RH, Wang R, Dashwood WM, Orner GA, Fischer KA et al.
2
Comparison of white tea, green tea, epigallocatechin-3-gallate, and caffeine as
3
inhibitors of PhIP-induced colonic aberrant crypts. Nutr Cancer 2007; 58: 60-65.
5 6
[35] Yang CS, Wang X, Lu G, Picinich SC. Cancer prevention by tea: animal studies, molecular mechanisms and human relevance. Nat Rev Cancer 2009; 9: 429-439.
RI PT
4
[36] Dreosti IE, Wargovich MJ, Yang CS. Inhibition of carcinogenesis by tea: the
evidence from experimental studies. Crit Rev Food Sci Nutr 1997; 37: 761-770.
8
[37] Lu YP, Lou YR, Lin Y, Shih WJ, Huang MT, Yang CS et al. Inhibitory Effects of
9
Orally Administered Green Tea, Black Tea, and Caffeine on Skin Carcinogenesis
SC
7
in Mice Previously Treated with Ultraviolet B Light (High-Risk Mice)
11
Relationship to Decreased Tissue Fat. Cancer Res 2001; 61: 5002-5009.
12
M AN U
10
[38] Mazzanti G, F Menniti-Ippolito, PA Moro, F Cassetti, R Raschetti, C Santuccio et
13
al. Hepatotoxicity from green tea: a review of the literature and two unpublished
14
cases. Eur J Clin Pharmacol 2009; 65: 331-341.
16
[39] Patel S S, S Beer, D L Kearney, G Phillips, B A Carter. Green tea extract: a potential cause of acute liver failure. World J Gastroenterol 2013;19: 5174-5177.
TE D
15
17
[40] He P, Noda Y, Sugiyama K. Green tea suppresses lipopolysaccharide induced
18
liver injury in D-galactosamine–sensitized rats. J Nutr 2001; 131: 1560-1567. [41] Dobrzynska I, Sniecinska A, Skrzydlewska E, Figaszewski Z. Green tea
EP
19 20
modulation of the biochemical and electric properties of rat liver cells that were
21
affected by ethanol and aging. Cell Mol Biol Lett 2004; 9: 709–21. [42] Skrzydlewska E, Ostrowska J, Stankiewicz A, Farbiszewski R. Green tea as a
23
potent antioxidant in alcohol intoxication. Addict Biol 2002; 7: 307–314.
24 25 26
AC C
22
[43] Augustyniak A, Waszkiewicz E, Skrzydlewska E. Preventive action of green tea from changes in the liver antioxidant abilities of different aged rats intoxicated with ethanol. Nutrition 2005; 21: 925–932.
27
[44] Fattovich G, Bortolotti F, Donato F. Natural history of chronic hepatitis B: special
28
emphasis on disease progression and prognostic factors. J Hepatol 2008; 48:
29
335-352.
30
[45] Patrick L. Hepatitis C: epidemiology and review of complementary/alternative 13
ACCEPTED MANUSCRIPT medicine treatments. Altern Med Rev 1999; 4: 220-238.
1 2
[46] Xu J, Wang J, Deng F, Hu Z, Wang H. Green tea extract and its major component
3
epigallocatechin gallate inhibits hepatitis B virus in vitro. Antiviral Res 2008; 78:
4
242-249.
RI PT
5 6 7 8
SC
9
Figure legends
11
Fig. 1. Flow diagram of identification of relevant studies.
12
Fig. 2. Relative risks of green tea consumption (highest versus non/lowest) and liver
13
cancer risk. Squares indicate study-specific risk estimates; horizontal lines indicate
14
95% CIs; diamond indicates summary RR estimate with its corresponding 95% CI;
15
RR, relative risk; 95 % CI, 95 % confidence interval.
16
Fig. 3. Relative risks of green tea consumption (highest versus non/lowest) and liver
17
cancer risk, stratified by gender. Squares indicate study-specific risk estimates;
18
horizontal lines indicate 95% CIs; diamond indicates summary RR estimate with its
19
corresponding 95% CI; RR, relative risk; 95 % CI, 95 % confidence interval.
20
Fig. 4. Begg’s funnel plot for publication bias of studies on green tea consumption and
21
liver
cancer
AC C
EP
TE D
M AN U
10
14
risk.
ACCEPTED MANUSCRIPT
Characteristics of cohort studies of green tea consumption and liver cancer risk Recruitment
Follow-up
Age
Country
time
(years)
(years)
Nechuta et al. [11]
1996-2000
11
40-70
Sex
Cohort
cases
size
133
67230
Outcome
Green tea consumption
Relative risk
Adjustments
(95% CI)
<3 times/week
1.00 (reference)
age, marital status, education, occupation, BMI,
(2012)
liver cancer
≥3 times/week
0.89 (0.58, 1.38)
exercise, fruit and vegetable intake, meat intake,
China
incidence
14
45-74
M/F
256
(2011) Singapore 1994.10-12
9
40-79
M/F
(2009)
(2009)
41761
1993-1994
14
40-69
AC C
Japan
Inoue et al. [14]
247
EP
Ui et al. [13]
39970
M/F
110
18815
diabetes, and family history of digestive system cancer
HCC
Non-drinkers
1.00 (reference)
age, gender, dialect group, year of recruitment,
incidence
Drinkers
0.98 (0.76, 1.26)
BMI, level of education, alcohol consumption,
TE D
1993-1998
M AN U
Primary
Johnson et al. [12]
F
No. of
SC
Study (year)
RI PT
Table 1
cigarette smoking, and history of diabetes
Primary
<1 cup/day
1.00 (reference)
age, sex, alcohol consumption, smoking status,
liver cancer
1-2 cups/day
0.78 (0.54, 1.12)
coffee consumption, vegetable consumption,
incidence
3-4 cups/day
0.98 (0.69, 1.37)
dairy products consumption, fruit consumption,
≥5 cups/day
0.58 (0.41, 0.83)
fish consumption, soybean consumption
Primary
<3 cups/day
1.00 (reference)
sex, age, area, smoking status, alcohol intake,
liver cancer
3-4 cups/day
1.62 (0.97, 2.69)
BMI, history of diabetes mellitus, coffee
15
ACCEPTED MANUSCRIPT
1992
15
25-69
M
1803
89789
China
Iso et al. [16]
No
mortality
Yes
F
1988-1990
13
40-79
M
age, surface antigen of HBV, occupation, history
0.91 (0.79, 1.06)
of liver cancer, smoking, and alcohol drinking
602
1.00 (reference)
Yes
0.51 (0.27, 0.96) 1.00 (reference)
(2007)
liver cancer
1-3/day
0.68 (0.49, 0.94)
Japan
mortality
≥4/day
0.89 (0.69, 1.16)
≤4/week
1.00 (reference)
1-3/day
0.69 (0.44, 1.08)
≥4/day
0.85 (0.59, 1.03)
Japan
1990.07-08
7
40-64
16
55
for cohort II Nagano et al. [18]
1979-1981
M/F
EP
for cohort I
≥40
70+47
M/F
391
age, area of study
22404+
Primary
≤2 cups/day
1.00 (reference)
age, gender, history of liver cancer, alcohol
38703
liver cancer
3-4 cups/day
1.20 (0.75, 1.94)
consumption, and smoking status
incidence
≥5 cups/day
0.90 (0.56, 1.44)
Primary
0-1 time/day
1.00 (reference)
AC C
(2005)
9
TE D
F
99510
No
≤4/week
1984.01
status, and HBV infection status
1.00 (reference)
Primary
Shimazu et al. [17]
consumption, serum ALT level, HCV infection
SC
(2008)
HCC
1.44 (0.84, 2.45)
M AN U
Wang et al. [15]
≥5 cups/day
RI PT
incidence
Japan
38540
16
city, age, gender, radiation exposure, smoking
ACCEPTED MANUSCRIPT
2-4 times/day
incidence
≥5 times/day
Primary
≤3 cups/day
(2000)
liver cancer
≥10 cups/day
Japan
incidence
Japan Nakachi et al. [19]
1986
11
>40
M/F
35
8552
1.10 (0.80, 1.40)
RI PT
liver cancer
status, alcohol drinking, BMI, education level,
0.95 (0.69, 1.30)
and calendar time
1.00 (reference)
age, cigarette smoking, alcohol consumption
0.53 (0.17, 1.57)
SC
(mean)
(2001)
M AN U
ALT, alanine aminotransferase; BMI, body mass index; CI, confidence interval; F, female; HCC, hepatocellular carcinoma; HBV, hepatitis B virus; HCV, hepatitis C virus;
AC C
EP
TE D
M, male
17
ACCEPTED MANUSCRIPT Table 2 Summary risk estimates and 95% CIs for green tea consumption (highest versus non/lowest) and liver cancer risk. No. of studies
Relative risk
Heterogeneity test P
I2 (%)
12
0.88 (0.81-0.97)
0.199
25.7
China
3
0.88 (0.77-1.01)
Japan
7
0.86 (0.75-0.98)
Singapore
1
0.98 (0.76-1.26)
Male
4
Female
5
Total population
6
Overall Country
34.2
0.144
37.3
NA
NA
0.89 (0.79-1.00)
0.350
8.7
0.78 (0.64-0.96)
0.193
34.2
0.89 (0.76-1.04)
0.066
51.6
0.98 (0.75-1.27)
0.324
13.7
SC
0.219
M AN U
Gender
No. of liver cancer cases
RI PT
(95% CI)
4
>200
7
0.87 (0.79-0.96)
0.154
35.9
7
0.89 (0.77-1.03)
0.112
41.9
4
0.88 (0.78-0.99)
0.375
3.5
3
0.87 (0.73-1.04)
0.969
0.0
8
0.89 (0.80-0.98)
0.064
47.6
5
0.97(0.95-1.00)
0.203
24.5
Outcome Liver cancer incidence
EP
Liver cancer mortality
TE D
<200
Adjusted for alcohol consumption
AC C
No Yes
Increament of 1cup/day
CI, confidence interval; NA, not available
18
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights 1. Green tea intake has preventive effects on liver cancer risk in Asian populations. 2. No association between green tea intake and liver cancer risk was observed in men. 3. One cup/day increment of green tea intake was not associated with liver cancer
AC C
EP
TE D
M AN U
SC
RI PT
risk.
S0899-9007(15)00237-3
DOI:
10.1016/j.nut.2015.05.021
Reference:
NUT 9546
To appear in:
Nutrition
Received Date: 9 February 2015 Revised Date:
13 May 2015
Accepted Date: 20 May 2015
Please cite this article as: Huang Y-Q, Lu X, Min H, Wu Q-Q, Shi X-T, Bian K-Q, Zou X-P, Green tea and liver cancer risk: a meta-analysis of prospective cohort studies in Asian populations, Nutrition (2015), doi: 10.1016/j.nut.2015.05.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1
Green tea and liver cancer risk: a meta-analysis of prospective cohort studies in Asian
2
populations
3
Ya-Qing Huang M.D. a, Xin Lu M.D. b, Han Min M.D. a, Qian-Qian Wu M.D. a,
5
Xiao-Ting Shi M.D. a, Kang-Qi Bian M.D. a, Xiao-Ping Zou M.D., Ph.D. a, *
RI PT
4
6 7
a
8
Nanjing Medical University, Nanjing, China
9
b
Department of Cardiology, the First Affiliated Hospital of Nanjing Medical
SC
10
Department of Gastroenterology, Nanjing Drum Tower Hospital Clinical College of
University, Nanjing, China
M AN U
11 12
*corresponding author. Tel./fax: +86 13770771661.
13
E-mail address: [email protected] (X-P. Zou)
14
Y-QH and X-PZ designed the research. Y-QH and XL performed the literature
16
research and extracted the data. HM contributed analytical tools. Y-QH, X-TS and
17
Q-QW analyzed the data. All of the authors helped interpret the results and write and
18
revise the manuscript. The authors declare that they have no competing interests.
19
EP
TE D
15
Key words: Green tea
21
populations
22 23 24 25
Liver cancer
Meta-analysis
AC C
20
26 27 28 29 30 1
Cohort studies
Asian
ACCEPTED MANUSCRIPT Abstract
2
Objectives: The aim of this meta-analysis was to investigate whether an association
3
existed between green tea consumption and the risk of liver cancer in prospective
4
cohort studies in Asian populations.
5
Methods: Relevant studies were identified by searching PUBMED, EMBASE, ISI
6
web of science and Chinese Bio-medicine Database published before April 2015.
7
Study-specific risk estimates for the highest versus non/lowest and increment of 1
8
cup/day green tea consumption levels were combined based on fixed or random
9
effects models. STATA 11.0 software was used for statistical analysis.
SC
RI PT
1
Results: A total of nine prospective cohort articles involving 465,274 participants and
11
3,694 cases of liver cancer from China, Japan and Singapore were included. The
12
summary relative risk (RR) indicated a significant association between the highest
13
green tea consumption and reduced risk of liver cancer [summary RR: 0.88; 95%
14
confidence interval (CI): 0.81, 0.97]. However, no statistically significant association
15
was observed through the analysis of an increment of 1 cup/day (summary RR: 0.97;
16
95% CI: 0.95, 1.00). When stratified analysis by gender, the protective effect of green
17
tea consumption on liver cancer risk was observed in women group (summary RR:
18
0.78; 95% CI: 0.64, 0.96), but not in men group (summary RR: 0.89; 95% CI: 0.79,
19
1.00).
20
Conclusions: The present analysis indicated the preventive effects of green tea intake
21
on the risk of liver cancer in Asian women populations. However, additional studies
22
are needed to make a convincing case for this association.
24 25
TE D
EP
AC C
23
M AN U
10
26 27 28 29 30 2
ACCEPTED MANUSCRIPT 1
Introduction
2
Primary liver cancer is the sixth most common cancer and, owing to its poor
3
prognosis, the second most common cause of cancer death [1]. An estimated 782,000
4
new primary liver cancer cases and 745,000 deaths occurred worldwide in 2012. In
5
addition
6
infections, which are major aetiologic determinants, other possible environmental risk
7
factors such as dietary aflatoxin, excessive alcohol intake, cigarette smoking, obesity
8
and diabetes play a key role in the development of liver cancer [2].
hepatitis
B
virus
(HBV)/hepatitis
C
virus
(HCV)
RI PT
to
Brewed from the plant Camellia sinensis, tea is one of the most popular beverages
10
consumed around the world. According to the level of fermentation, tea can be mainly
11
classified as three major commercial types: green tea (non-fermented), black tea
12
(fermented) and oolong tea (half-fermented). Among them, green tea contains a much
13
higher level of polyphenols known as catechins, which have been shown to have
14
antimutagenic, antigenotoxic, and anticarcinogenic activities [3]. A major polyphenol
15
of green tea, epigallocatechin-3-gallate (EGCG), has generated interest for its
16
anti-tumor effects [4]. Numerous animal and in vitro experiments have been carried
17
out to support the possibility that green tea has preventive effects against liver cancer
18
[5, 6]. EGCG, as the major polyphenol of green tea, is proven to inhibit the growth of
19
all liver cancer cell lines [7]. However, the molecular mechanisms of the
20
chemopreventive effects of green tea are still uncertain.
EP
TE D
M AN U
SC
9
Over the last two decades, several epidemiological studies have been conducted to
22
investigate the association between green tea consumption and liver cancer risk, but
23
with varying results. One review [8] declared very limited support for green tea
24
consumption in reducing the risk of liver cancer. A previous meta-analysis [9]
25
suggested that high green tea intake might be associated with lower risk of liver
26
cancer, resulting from the combination of cohort and case-control studies. However, it
27
was limited because true differences in the level and range of intake between studies
28
were not taken into account. Considering the results of subsequent published cohort
29
studies tended to show a weak correlation between green tea intake and liver cancer
30
risk, we therefore conducted this exploratory meta-analysis to update and summarize
AC C
21
3
ACCEPTED MANUSCRIPT 1
the results from prospective studies for deriving a more precise estimation of this
2
topic and to further evaluate whether there is a dose-response relationship between
3
green tea intake and the risk for liver cancer.
4
Materials and methods
6
Data sources and searches
7
Pertinent studies were identified by a computerized Medline, EMBASE, ISI web of
8
science and Chinese Bio-medicine Database search that included the years 1966
9
through April 2015. We used the search terms “green tea,” “tea,” “polyphenol,” or
10
“catechin” combined with “liver neoplasm,” “liver cancer,” and “hepatocellular
11
carcinoma” in the full-text option. The search was done on studies conducted on
12
human species, without restriction on language. The reference lists of reviews and
13
retrieved articles were handsearched at the same time. The titles and abstracts were
14
scanned to exclude any studies that were clearly irrelevant. The full-texts and tables of
15
the remaining articles were retrieved and perused to determine the relevancy of the
16
study design and data, according to the inclusion criteria detailed below. Furthermore,
17
references from the retrieved articles were reviewed to make sure that all the relevant
18
bibliographies published were reviewed. Two authors (Y-QH and XL) conducted all
19
searches independently.
SC
M AN U
TE D
EP
20
22
Inclusion and exclusion criteria Following criteria were used to identify relevant studies for the meta-analysis: (1)
AC C
21
RI PT
5
23
participants were from Asian countries; (2) the exposure of interest was green tea
24
consumption; (3) the outcome of interest was total liver cancer incidence or mortality;
25
(4) RR or hazard ratio (HR) with their 95% CI (or sufficient information to calculate
26
them) were reported; (5) prospective cohort study design; and (6) some adjustments
27
were made for potential liver cancer risk. If cohorts were duplicated in more than one
28
study, only the most recent and complete publication was included in the
29
meta-analysis.
30
Our literature search identified 10 potentially relevant articles [10-19] concerning 4
ACCEPTED MANUSCRIPT 1
green tea consumption and liver cancer risk. One publication [10] was excluded due
2
to insufficient information on numbers of cases, controls and RRs with their
3
corresponding 95% CIs. Thus, the remaining 9 studies were included in the
4
meta-analysis.
6
RI PT
5
Data extraction
For each study, we extracted the following data: the name of first author, year of
8
publication, country where the study was conducted, recruitment time, participant age,
9
sex, years of follow-up, number of participants (cases and cohort size), type of
10
outcome (primary liver cancer or HCC, incidence or mortality), the RR estimates with
11
corresponding 95% CIs for each category of green tea consumption and adjustment
12
factors. Two reviewers (Y-QH and HM) independently inspected the data for
13
eligibility and quality. The results were compared and any questions or discrepancies
14
were resolved by a third reviewer (Q-QW). For example, if the data were insufficient
15
or missing, we attempted to contact the authors of the articles in order to request the
16
relevant data.
17 18
Statistical analysis
TE D
M AN U
SC
7
The measure of effect of interest was the RR and the corresponding 95% CI and the
20
HR was considered as RR directly. Study-specific RRs and 95% CIs for highest
21
versus non/lowest green tea consumption level were extracted from each study, and
22
then logRRs were weighted by the inverse of their variances to obtain summary RRs
23
and 95% CIs. If the study provided separate RR estimates for men and women, we
24
treated them as two different studies in statistical analysis. Studies were combined by
25
using the DerSimonian and Laird random-effects model, which incorporates both
26
within- and between-study variability [20].
AC C
EP
19
27
Statistical heterogeneity among studies was evaluated using the Q and I2 statistics
28
[21]; P<0.10 and I2>50% were considered statistically significant [22]. When there is
29
significant heterogeneity among study results, the random effect model was used to
30
calculate summary RR while the fixed effect model was used to calculate summary 5
ACCEPTED MANUSCRIPT 1
RR among studies with homogeneous results. In addition, subgroup analysis was
2
carried out to investigate the influence of geographic area (China versus Japan),
3
gender, the number of liver cancer cases (<200 vs. ≥200) and adjusted factor of
4
alcohol drinking. To normalize the variation between studies for the difference in exposure categories,
6
dose-response analysis was conducted to estimate study-specific slopes from the
7
natural logarithm of the RR across categories of green tea consumption, using the
8
method described by Greenland and Longnecher [23] and Orsini et al [24]. Because
9
this method requires the risk estimates with their variance estimates for three or more
10
quantitative exposure categories, so studies with only 2 categories [11,15,18,19] were
11
not included in this analysis. In studies [13,16] that did not provide the number of
12
cases and controls in each exposure category, the variance-weighted least-squares
13
regression model was used to estimate the slopes. Midpoint of upper and lower
14
boundaries was considered as the dose of each category if the study reported only the
15
range of green tea consumption. Because the highest category of consumption was
16
usually open, we considered it of the same amplitude as the preceding category. Then
17
summary RR estimates were obtained by pooling the study-specific slopes and taking
18
the inverse of their variances as weighted.
TE D
M AN U
SC
RI PT
5
Furthermore, Begg’s funnel plot and Egger’s regression test were used to evaluate
20
whether publication bias might affect the statistical results [25,26]. The funnel plot
21
would be asymmetrical if publication bias existed and a P value of less than 0.1 was
22
considered representative of significant statistical publication bias. All statistical
23
analysis was performed by the STATA version 11.0 software (Stata Corporation,
24
College Station, TX). All statistical tests were two-sided.
AC C
25
EP
19
26
Results
27
Search results and study characteristics
28
The details of literature search are shown in Figure 1. A total of nine articles and
29
eleven studies involving 465,274 participants and 3,694 cases of liver cancer were
30
included in the meta-analysis. And the characteristics of the included studies are listed 6
ACCEPTED MANUSCRIPT 1
in Table 1. These studies were published between 2000 and 2012. Seven of the eleven
2
studies were conducted in Japan [11-14,16,17], three in China [15,19], and the
3
remaining one in Singapore [18].
4
Highest versus non/lowest drinkers
RI PT
5
Relative risk estimates of liver cancer for highest green tea consumption compared
7
with non/lowest green tea consumption for individual studies and all studies
8
combined were shown in Figure 2. Overall, we found that highest green tea
9
consumption was statistically significantly associated with reduced risk of liver cancer
10
(summary RR: 0.88; 95% CI: 0.81, 0.97). No statistically significant heterogeneity
11
was found among the studies (Pheterogeneity=0.199, I2=25.7%). In a subgroup analysis
12
conducted by gender, only five studies were included in the analysis. The overall
13
results in women showed a significant 22% reduction in liver cancer risk with high
14
intake of green tea (summary RR: 0.78; 95% CI: 0.64, 0.96; Pheterogeneity=0.193;
15
I2=34.2%). No such effect was noted in men (summary RR: 0.89; 95% CI: 0.79, 1.00;
16
Pheterogeneity=0.35; I2=8.7%) (Figure 3, Table 2). We also stratified the various studies
17
by study country, liver cancer cases, outcome and adjustment for alcohol consumption
18
(Table 2). Similar but with borderline statistically significant association was found
19
for all these stratified analyses.
EP
TE D
M AN U
SC
6
20
22
Dose-response analysis
Given the wide array of measurement categories reported among studies, we also
AC C
21
23
conduct dose-response analysis. However, the data did not show any inverse
24
relationship between an increase in green tea consumption of 1 cup/day and the liver
25
cancer risk (summary RR: 0.97; 95% CI: 0.95, 1.00; Pheterogeneity=0.203; I2=24.5%)
26
(Table 2).
27 28
Publication bias
29
There was no indication of publication bias in the reporting of results on green tea
30
consumption and liver cancer risk, from either visualization of the Begg’s funnel plot 7
ACCEPTED MANUSCRIPT 1
(P=0.27) or the Eggers’s test (P=0.38) (Figure 4).
2 3
Discussion To our knowledge, the present meta-analysis is the first quantitative systematic
5
analysis of the association between green tea consumption and liver cancer risk based
6
on published prospective cohort studies in Asian populations. The results suggest that
7
high green tea consumption is associated with a decreased risk of liver cancer in Asian
8
women populations. However, no statistically significant association was observed
9
through the analysis of an increment of 1 cup/day. This result is consistent with a
10
recent dose–response meta-analysis [27] of 3 prospective studies, in which no inverse
11
association (RR: 0.93; 95% CI: 0.75, 1.17) was found between an increase in green
12
tea consumption of 3 cup/day and risk of liver cancer.
M AN U
SC
RI PT
4
Although the exact mechanisms are unclear, a protective role of green tea in
14
digestive system cancer prevention is biologically plausible. EGCG is the most
15
abundant polyphenol, accounting for 40% of the total polyphenol in green tea extract,
16
and is considered to be the most active constituent [28,29]. EGCG may reduce DNA
17
damage by reactive oxygen species [30,31] and inhibit tumor cell growth, invasion
18
and angiogenesis [32-35]. Several studies have reported that the incidence and
19
recurrence rate of animal tumors is significantly lower in animals receiving EGCG
20
than in control animals [36,37]. However, the effect of green tea on liver cancer is not
21
well understood. Although a few studies [38,39] linked it to hepatotoxicity, the
22
majority experimental studies have suggested green tea, especially green tea contracts,
23
have a hepatoprotective effect and anticarcinogenic potential in the liver [40].
EP
AC C
24
TE D
13
In this meta-analysis, a more obvious inverse association between green tea
25
consumption and liver cancer was found in women (RR=0.78) than in men (RR=0.89).
26
However, this result was derived from only five studies with a small number of cases,
27
so we could not draw a firm conclusion. Alcohol consumption is known to cause liver
28
cancer, and after adjustment for this, a significant inverse association was showed
29
between green tea consumption and liver cancer. Possible explanations may be due to
30
the anti-oxidative effects of tea polyphenols. Previous evidence showed green tea may 8
ACCEPTED MANUSCRIPT 1
protect against alcohol-related oxidative modification [41-43]. A history of liver disease may be a risk factor for liver cancer. Individuals who have
3
a history of liver disease are likely to have been infected with HBV or HCV [44]. It is
4
expected that the inverse association would be more pronounced among such
5
individuals [45, 46]. Nevertheless, in our study, only one article [17] discussed the
6
inverse association between green tea consumption and the risk of liver cancer among
7
individuals with/without a history of liver disease. A better understanding of this issue
8
will require further research.
RI PT
2
Our study has several potential limitations. Firstly, the possibility of bias and
10
confounding cannot be excluded in observational studies. Our meta-analysis enrolled
11
only cohort studies, which are less susceptible to bias because of the prospective
12
design. However, individual studies may have failed to adjust for potential known or
13
unknown confounders. Secondly, all of the studies included in the analysis had been
14
conducted in Asian populations (mainly China and Japan) because of the popularity of
15
green tea in East Asia. Therefore, the results should be cautiously extrapolated to
16
other populations. Finally, publication bias could be of concern because small studies
17
with null results tend not to be published. We had limited statistical power to
18
conclusively reject the null hypothesis of no publication bias because of the relative
19
small number of studies.
EP
TE D
M AN U
SC
9
In conclusion, our findings of this meta-analysis indicated an inverse association
21
between high green tea consumption and reduced liver cancer risk in Asian women
22
populations. Due to the limited available data, further large prospective studies should
23
be taken to confirm the results.
24 25
AC C
20
26 27 28 29 30 9
ACCEPTED MANUSCRIPT 1
References
2
[1] Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M et al. Cancer
3
incidence and mortality worldwide:sources, methods and major patterns in
4
GLOBOCAN 2012. Int J Cancer 2015; 136: E359-86.
6 7 8
[2] Yu MC, Yuan JM. Environmental factors and risk for hepatocellular carcinoma.
RI PT
5
Gastroenterology 2004; 127: S72-S78.
[3] Kuroda Y, Hara Y. Antimutagenic and anticarcinogenic activity of tea polyphenols. Mutat Res 1999; 436: 69-97.
[4] Landis-Piwowar KR, Huo C, Chen D, Milacic V, Shi G, Chan TH et al. A novel
10
prodrug of the green tea polyphenol (-)-epigallocatechin-3-gallate as a potential
11
anticancer agent. Cancer Res 2007; 67: 4303-4310 .
M AN U
12
SC
9
[5] Qin G, Gopalan-Kriczky P, Su J, Ning Y, Lotolikar PD. Inhibition of aflatoxin
13
B1-induced initiation of hepatocarcinogenesis in the rat by green tea. Cancer Lett
14
1997; 112: 149–154.
15
[6] Sakata R, Ueno T, Nakamura T, Sakamoto M, Torimura T, Sata M. Green tea polyphenol epigallocatechin-3-gallate inhibits platelet-derived growth
17
factor-induced proliferation of human hepatic stellate cell line LI90. J Hepatol
18
2004; 40: 52-59.
[7] Nishikawa T, Nakajima T, Moriguchi M, Jo M, Sekoguchi S, Ishii M et al. A green
EP
19
TE D
16
tea polyphenol, epigalocatechin-3-gallate, induces apoptosis of human
21
hepatocellular carcinoma, possibly through inhibition of Bcl-2 family proteins. J
22
Hepatol 2006; 44: 1074-1082.
23 24 25
AC C
20
[8] Yuan JM. Cancer prevention by green tea: evidence from epidemiologic studies. Am J Clin Nutr 2013; 98: 1676S-1681S.
[9] Fon Sing M, Yang WS, Gao S, Gao J, Xiang YB. Epidemiological studies of the
26
association between tea drinking and primary liver cancer: a meta-analysis. Eur J
27
Cancer Prev 2011; 20: 157-165.
28
[10] Inoue M, Yoshimi I, Sobue T, Tsugane S, JPHC Study Group. Influence of coffee
29
drinking on subsequent risk of hepatocellular carcinoma: a prospective study in
30
Japan. J Natl Cancer Inst 2005; 97: 293-300. 10
ACCEPTED MANUSCRIPT 1
[11] Nechuta S, Shu XO, Li HL, Yang G, Ji BT, Xiang YB et al. Prospective cohort
2
study of tea consumption and risk of digestive system cancers: results from the
3
Shanghai Women's Health Study. Am J Clin Nutr 2012; 96: 1056-1063.
4
[12] Johnson S, Koh WP, Wang R, Govindarajan S, Yu MC, Yuan JM. Coffee consumption and reduced risk of hepatocellular carcinoma: findings from the
6
Singapore Chinese Health Study. Cancer Causes Control 2011; 22: 503-510.
7
[13] Inoue M, Kurahashi N, Iwasaki M, Shimazu T, Tanaka Y, Mizokami M et al.
RI PT
5
Effect of coffee and green tea consumption on the risk of liver cancer: cohort
9
analysis by hepatitis virus infection status. Cancer Epidemiol Biomarkers Prev
11
2009; 18: 1746-1753.
[14] Ui A, Kuriyama S, Kakizaki M, Sone T, Nakaya N, Ohmori-Matsuda K et al.
M AN U
10
SC
8
12
Green tea consumption and the risk of liver cancer in Japan: the Ohsaki Cohort
13
study. Cancer Causes Control 2009; 20: 1939-1945.
14 15
[15] Wang N, Zheng Y, Jiang Q, Yu X, Chen Y. Tea and reduced liver cancer mortality. Epidemiology 2008; 19: 761.
[16] Iso H, Kubota Y. Nutrition and disease in the Japan Collaborative Cohort study
17
for evaluation of cancer (JACC). Asian Pac J Cancer Prev 2007; 8: 35-80.
18
TE D
16
[17] Shimazu T, Tsubono Y, Kuriyama S, Ohmori K, Koizumi Y, Nishino Y et al. Coffee consumption and the risk of primary liver cancer: pooled analysis of two
20
prospective studies in Japan. Int J Cancer 2005; 116: 150-154.
22 23 24 25
[18] Nagano J, Kono S, Preston DL, Mabuchi K. A prospective study of green tea consumption and cancer incidence, Hiroshima and Nagasaki (Japan).
AC C
21
EP
19
Cancer Causes Control 2001; 12: 501-508.
[19] Nakachi K, Matsuyama S, Miyake S, Suganuma M, Imai K. Preventive effects of drinking green tea on cancer and cardiovascular disease: epidemiological
26
evidence for multiple targeting prevention. Biofactors 2000; 13: 49-54.
27
[20] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials
28 29 30
1986; 7: 177-188. [21] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557. 11
ACCEPTED MANUSCRIPT 1 2
[22] Hedges LV, Pigott TD. The power of statistical tests in meta-analysis. Psychol Methods 2001; 6: 203-217. [23] Greenland S, Longnecker MP. Methods for trend estimation from summarized
4
dose–response data, with applications to meta-analysis. Am J Epidemiol 1992;
5
135: 1301–1309.
RI PT
3
6
[24] Orsini N, Bellocco R, Greenland S. Generalized least squares for trend
7
estimation of summarized dose–response data. Stata J 2006; 6: 40–57.
8
[25] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629-634.
10
[26] Thornton A, Lee P. Publication bias in meta-analysis: its causes and
12
consequences. J Clin Epidemiol 2000; 53:207-216.
M AN U
11
SC
9
[27] Yu F, Jin Z, Jiang H, Xiang C, Tang J, Li T et al. Tea consumption and the risk of
13
five major cancers: a dose–response meta-analysis of prospective studies. BMC
14
Cancer 2014; 14:197.
15
[28] Fujiki H, Suganuma M, Okabe S, Komori A, Sueoka E, Sueoka N et al. Japanese green tea as a cancer preventive in humans. Nutr Rev 1996; 54: S67-S70.
17
[29] Ahmad N, Mukhtar H. Green tea polyphenols and cancer: biologic mechanisms
19 20
and practical implications. Nutr Rev 1999; 57: 78-83. [30] Higdon JV, Frei B. Tea catechins and polyphenols: health effects, metabolism,
EP
18
TE D
16
and antioxidant functions. Crit Rev Food Sci Nutr 2003; 43: 89-143. [31] Hakim IA, Harris RB, Brown S, Chow HH, Wiseman S, Agarwal S et al. Effect
22
of increased tea consumption on oxidative DNA damage among smokers: a
23 24 25
AC C
21
randomized controlled study. J Nutr 2003; 133: 3303S-3309S.
[32] Yamane T, Takahashi T, Kuwata K, Oya K, Inagake M, Kitao Y et al. Inhibition of N-methyl-N′-nitro-N-nitrosoguanidine-induced carcinogenesis by
26
(-)-epigallocatechin gallate in the rat glandular stomach. Cancer Res 1995; 55:
27
2081-2084.
28
[33] Ju J, Hong J, Zhou JN, Pan Z, Bose M, Liao J et al. Inhibition of intestinal
29
tumorigenesis in Apcmin/+ mice by (−)-epigallocatechin-3-gallate, the major
30
catechin in green tea. Cancer Res 2005; 65: 10623-10631. 12
ACCEPTED MANUSCRIPT 1
[34] Carter O, Dashwood RH, Wang R, Dashwood WM, Orner GA, Fischer KA et al.
2
Comparison of white tea, green tea, epigallocatechin-3-gallate, and caffeine as
3
inhibitors of PhIP-induced colonic aberrant crypts. Nutr Cancer 2007; 58: 60-65.
5 6
[35] Yang CS, Wang X, Lu G, Picinich SC. Cancer prevention by tea: animal studies, molecular mechanisms and human relevance. Nat Rev Cancer 2009; 9: 429-439.
RI PT
4
[36] Dreosti IE, Wargovich MJ, Yang CS. Inhibition of carcinogenesis by tea: the
evidence from experimental studies. Crit Rev Food Sci Nutr 1997; 37: 761-770.
8
[37] Lu YP, Lou YR, Lin Y, Shih WJ, Huang MT, Yang CS et al. Inhibitory Effects of
9
Orally Administered Green Tea, Black Tea, and Caffeine on Skin Carcinogenesis
SC
7
in Mice Previously Treated with Ultraviolet B Light (High-Risk Mice)
11
Relationship to Decreased Tissue Fat. Cancer Res 2001; 61: 5002-5009.
12
M AN U
10
[38] Mazzanti G, F Menniti-Ippolito, PA Moro, F Cassetti, R Raschetti, C Santuccio et
13
al. Hepatotoxicity from green tea: a review of the literature and two unpublished
14
cases. Eur J Clin Pharmacol 2009; 65: 331-341.
16
[39] Patel S S, S Beer, D L Kearney, G Phillips, B A Carter. Green tea extract: a potential cause of acute liver failure. World J Gastroenterol 2013;19: 5174-5177.
TE D
15
17
[40] He P, Noda Y, Sugiyama K. Green tea suppresses lipopolysaccharide induced
18
liver injury in D-galactosamine–sensitized rats. J Nutr 2001; 131: 1560-1567. [41] Dobrzynska I, Sniecinska A, Skrzydlewska E, Figaszewski Z. Green tea
EP
19 20
modulation of the biochemical and electric properties of rat liver cells that were
21
affected by ethanol and aging. Cell Mol Biol Lett 2004; 9: 709–21. [42] Skrzydlewska E, Ostrowska J, Stankiewicz A, Farbiszewski R. Green tea as a
23
potent antioxidant in alcohol intoxication. Addict Biol 2002; 7: 307–314.
24 25 26
AC C
22
[43] Augustyniak A, Waszkiewicz E, Skrzydlewska E. Preventive action of green tea from changes in the liver antioxidant abilities of different aged rats intoxicated with ethanol. Nutrition 2005; 21: 925–932.
27
[44] Fattovich G, Bortolotti F, Donato F. Natural history of chronic hepatitis B: special
28
emphasis on disease progression and prognostic factors. J Hepatol 2008; 48:
29
335-352.
30
[45] Patrick L. Hepatitis C: epidemiology and review of complementary/alternative 13
ACCEPTED MANUSCRIPT medicine treatments. Altern Med Rev 1999; 4: 220-238.
1 2
[46] Xu J, Wang J, Deng F, Hu Z, Wang H. Green tea extract and its major component
3
epigallocatechin gallate inhibits hepatitis B virus in vitro. Antiviral Res 2008; 78:
4
242-249.
RI PT
5 6 7 8
SC
9
Figure legends
11
Fig. 1. Flow diagram of identification of relevant studies.
12
Fig. 2. Relative risks of green tea consumption (highest versus non/lowest) and liver
13
cancer risk. Squares indicate study-specific risk estimates; horizontal lines indicate
14
95% CIs; diamond indicates summary RR estimate with its corresponding 95% CI;
15
RR, relative risk; 95 % CI, 95 % confidence interval.
16
Fig. 3. Relative risks of green tea consumption (highest versus non/lowest) and liver
17
cancer risk, stratified by gender. Squares indicate study-specific risk estimates;
18
horizontal lines indicate 95% CIs; diamond indicates summary RR estimate with its
19
corresponding 95% CI; RR, relative risk; 95 % CI, 95 % confidence interval.
20
Fig. 4. Begg’s funnel plot for publication bias of studies on green tea consumption and
21
liver
cancer
AC C
EP
TE D
M AN U
10
14
risk.
ACCEPTED MANUSCRIPT
Characteristics of cohort studies of green tea consumption and liver cancer risk Recruitment
Follow-up
Age
Country
time
(years)
(years)
Nechuta et al. [11]
1996-2000
11
40-70
Sex
Cohort
cases
size
133
67230
Outcome
Green tea consumption
Relative risk
Adjustments
(95% CI)
<3 times/week
1.00 (reference)
age, marital status, education, occupation, BMI,
(2012)
liver cancer
≥3 times/week
0.89 (0.58, 1.38)
exercise, fruit and vegetable intake, meat intake,
China
incidence
14
45-74
M/F
256
(2011) Singapore 1994.10-12
9
40-79
M/F
(2009)
(2009)
41761
1993-1994
14
40-69
AC C
Japan
Inoue et al. [14]
247
EP
Ui et al. [13]
39970
M/F
110
18815
diabetes, and family history of digestive system cancer
HCC
Non-drinkers
1.00 (reference)
age, gender, dialect group, year of recruitment,
incidence
Drinkers
0.98 (0.76, 1.26)
BMI, level of education, alcohol consumption,
TE D
1993-1998
M AN U
Primary
Johnson et al. [12]
F
No. of
SC
Study (year)
RI PT
Table 1
cigarette smoking, and history of diabetes
Primary
<1 cup/day
1.00 (reference)
age, sex, alcohol consumption, smoking status,
liver cancer
1-2 cups/day
0.78 (0.54, 1.12)
coffee consumption, vegetable consumption,
incidence
3-4 cups/day
0.98 (0.69, 1.37)
dairy products consumption, fruit consumption,
≥5 cups/day
0.58 (0.41, 0.83)
fish consumption, soybean consumption
Primary
<3 cups/day
1.00 (reference)
sex, age, area, smoking status, alcohol intake,
liver cancer
3-4 cups/day
1.62 (0.97, 2.69)
BMI, history of diabetes mellitus, coffee
15
ACCEPTED MANUSCRIPT
1992
15
25-69
M
1803
89789
China
Iso et al. [16]
No
mortality
Yes
F
1988-1990
13
40-79
M
age, surface antigen of HBV, occupation, history
0.91 (0.79, 1.06)
of liver cancer, smoking, and alcohol drinking
602
1.00 (reference)
Yes
0.51 (0.27, 0.96) 1.00 (reference)
(2007)
liver cancer
1-3/day
0.68 (0.49, 0.94)
Japan
mortality
≥4/day
0.89 (0.69, 1.16)
≤4/week
1.00 (reference)
1-3/day
0.69 (0.44, 1.08)
≥4/day
0.85 (0.59, 1.03)
Japan
1990.07-08
7
40-64
16
55
for cohort II Nagano et al. [18]
1979-1981
M/F
EP
for cohort I
≥40
70+47
M/F
391
age, area of study
22404+
Primary
≤2 cups/day
1.00 (reference)
age, gender, history of liver cancer, alcohol
38703
liver cancer
3-4 cups/day
1.20 (0.75, 1.94)
consumption, and smoking status
incidence
≥5 cups/day
0.90 (0.56, 1.44)
Primary
0-1 time/day
1.00 (reference)
AC C
(2005)
9
TE D
F
99510
No
≤4/week
1984.01
status, and HBV infection status
1.00 (reference)
Primary
Shimazu et al. [17]
consumption, serum ALT level, HCV infection
SC
(2008)
HCC
1.44 (0.84, 2.45)
M AN U
Wang et al. [15]
≥5 cups/day
RI PT
incidence
Japan
38540
16
city, age, gender, radiation exposure, smoking
ACCEPTED MANUSCRIPT
2-4 times/day
incidence
≥5 times/day
Primary
≤3 cups/day
(2000)
liver cancer
≥10 cups/day
Japan
incidence
Japan Nakachi et al. [19]
1986
11
>40
M/F
35
8552
1.10 (0.80, 1.40)
RI PT
liver cancer
status, alcohol drinking, BMI, education level,
0.95 (0.69, 1.30)
and calendar time
1.00 (reference)
age, cigarette smoking, alcohol consumption
0.53 (0.17, 1.57)
SC
(mean)
(2001)
M AN U
ALT, alanine aminotransferase; BMI, body mass index; CI, confidence interval; F, female; HCC, hepatocellular carcinoma; HBV, hepatitis B virus; HCV, hepatitis C virus;
AC C
EP
TE D
M, male
17
ACCEPTED MANUSCRIPT Table 2 Summary risk estimates and 95% CIs for green tea consumption (highest versus non/lowest) and liver cancer risk. No. of studies
Relative risk
Heterogeneity test P
I2 (%)
12
0.88 (0.81-0.97)
0.199
25.7
China
3
0.88 (0.77-1.01)
Japan
7
0.86 (0.75-0.98)
Singapore
1
0.98 (0.76-1.26)
Male
4
Female
5
Total population
6
Overall Country
34.2
0.144
37.3
NA
NA
0.89 (0.79-1.00)
0.350
8.7
0.78 (0.64-0.96)
0.193
34.2
0.89 (0.76-1.04)
0.066
51.6
0.98 (0.75-1.27)
0.324
13.7
SC
0.219
M AN U
Gender
No. of liver cancer cases
RI PT
(95% CI)
4
>200
7
0.87 (0.79-0.96)
0.154
35.9
7
0.89 (0.77-1.03)
0.112
41.9
4
0.88 (0.78-0.99)
0.375
3.5
3
0.87 (0.73-1.04)
0.969
0.0
8
0.89 (0.80-0.98)
0.064
47.6
5
0.97(0.95-1.00)
0.203
24.5
Outcome Liver cancer incidence
EP
Liver cancer mortality
TE D
<200
Adjusted for alcohol consumption
AC C
No Yes
Increament of 1cup/day
CI, confidence interval; NA, not available
18
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights 1. Green tea intake has preventive effects on liver cancer risk in Asian populations. 2. No association between green tea intake and liver cancer risk was observed in men. 3. One cup/day increment of green tea intake was not associated with liver cancer
AC C
EP
TE D
M AN U
SC
RI PT
risk.