Colorectal cancer and adenomatous polyps in relation to allium vegetables intake: A meta-analysis of observational studies

Mol. Nutr. Food Res. 2014, 00, 1–8 1DOI 10.1002/mnfr.201400169

RESEARCH ARTICLE

Colorectal cancer and adenomatous polyps in relation to

allium vegetables intake: A meta-analysis of

observational studies

Federica Turati1, Valentina Guercio2,3, Claudio Pelucchi1, Carlo La Vecchia3

and Carlotta Galeone1

1 Department of Epidemiology, IRCCS—Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy2 Department of Surgery, University of Siena, Siena, Italy3 Department of Clinical Sciences and Community Health, University of Milan, Milan, Italy

Received: March 10, 2014Revised: April 29, 2014Accepted: May 5, 2014

Scope: To provide updated quantitative estimates of the associations between allium vegetablesintake and risk of colorectal cancer and colorectal adenomatous polyps.Methods and results: We combined all published data on the issue, using a meta-analyticapproach. Pooled relative risks (RRs) were calculated using random-effects models. Sixteenstudies (13 333 cases) were included in the meta-analyses of colorectal cancer. Seven studiesprovided information on garlic, six on onion, and four on total allium vegetables. The pooledRRs of colorectal cancer for the highest versus the lowest category of intake were 0.85 (95%confidence interval; CI, 0.72–1.00) for garlic (0.76 for case-control, 0.99 for cohort studies), 0.85(95% CI, 0.70–1.04) for onion (0.74 for case-control, 1.04 for cohort studies), and 0.78 (95%CI, 0.56–1.08) for total allium vegetables. Significant heterogeneity was found for the threemeta-analyses. The pooled RR of colorectal adenomatous polyps for the highest versus thelowest category of total allium vegetables intake was 0.88 (95% CI, 0.80–0.98, three studies),with no heterogeneity.Conclusion: High garlic intake may reduce the risk of colorectal cancer. However, evidence ofsuch protection derived mainly from case-control studies. High intake of total allium vegetablesmay be associated with a risk reduction of colorectal adenomatous polyps.

Keywords:

Allium vegetables / Colorectal adenomatous polyps / Colorectal neoplasm / Diet /Garlic

� Additional supporting information may be found in the online version of this article atthe publisher’s web-site

1 Introduction

The hypothesis of a possible favorable role of allium veg-etables in cancer prevention is based on numerous phys-iological properties of these vegetables and their chemical

Correspondence: Dr. Claudio Pelucchi, Department of Epidemiol-ogy, IRCCS—Istituto di Ricerche Farmacologiche Mario Negri, ViaGiuseppe La Masa 19, 20156 Milan, ItalyE-mail: [email protected]: +39-0233200231

Abbreviations: CI, confidence interval; FFQs, food frequencyquestionnaires; MOOSE, Meta-analysis of Observational Studiesin Epidemiology; OR, odds ratio; OSCs, organosulfur compounds;PLCO, prostate, lung, colorectal, and ovarian; RR, relative risk

constituents, particularly organosulfur compounds (OSCs),which have been shown to inhibit carcinogenesis in severalanimal studies [1–3]. The association between allium veg-etables and the risk for various cancer sites has been consid-ered in several epidemiological studies, but the issue remainsunsettled [4]. According to the Second Expert Report of theAmerican Institute for Cancer Research, high consumptionof allium vegetables may reduce the risk of stomach cancerwith a dose–response relationship, whereas a potential fa-vorable role on colorectal cancer is limited to high intake ofgarlic [3]. Several reviews on the role of garlic intake on therisk of colorectal cancer have been published [1,4]. However,to our knowledge only a meta-analysis published over 10 yearsago provided a quantitative overall estimate of such associa-tion [5]. The pooled relative risk (RR) of colorectal cancer for

C© 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.mnf-journal.com

2 F. Turati et al. Mol. Nutr. Food Res. 2014, 00, 1–8

the highest level of garlic intake as compared to the lowestone was 0.66 (95% confidence interval; CI, 0.48–0.91), basedon eight studies, with an indication of possible presence ofpublication bias. During the last decade, new data on thisissue have been made available. In particular, results fromthe Shanghai Men’s Health Study (with about 4000 cases) [6],the American Cancer Society CPS-II cohort (with more than1000 cancer cases) [7], two case-control studies (with a totalof around 3000 cancer cases) [8, 9], and from the update ofthe Health Professionals Follow-up Study and Nurses’ HealthStudy were published [10].

The Shanghai Men’s Health study found no relation be-tween allium vegetables intake and the risk of colorectal, colonand rectal cancers [6]; the American Cancer Society CPS-II co-hort found a weak inverse association of colorectal cancer riskwith dietary garlic consumption in women, but a possible in-creased risk in men [7]; the case-control study from Australiafound a nonsignificant decreased risk of colorectal cancer forhigh intakes of garlic [8]; the Italian and Swiss case-controlstudy showed a consistent protection on colorectal cancerfor both high frequencies of use of garlic and onion [9], andthe update of the Health Professionals Follow-up Study andNurses’ Health Study did not support an important role ofgarlic intake in colorectal carcinogenesis [10].

Using an accurate bibliographic search, including gen-eral key words as “diet” or “food” besides specific terms forallium vegetables, we collected sufficient information to re-view the association of colorectal cancer risk not only withgarlic intake but also with other individual types of alliumvegetables, including onion and leek, and garlic supplementintake. To provide quantitative overall estimates on these is-sues, we combined all published data using a meta-analyticapproach. Moreover, we summarized evidences on the roleof allium vegetable intake on the risk of colorectal adenoma-tous polyps. Colorectal adenomatous polyps are precursors ofcolorectal cancer, and environmental factors, including diet,seem to have an important influence on the multiple-stageprocess that leads from precursor lesions to cancer [11].

2 Materials and methods

2.1 Search strategy

This meta-analysis followed the Meta-analysis of ObservationalStudies in Epidemiology (MOOSE) guidelines [12]. In April2014, two authors (C. G. and V. G.) independently performeda systematic literature search in the Medline database, usingthe string “(case-control OR cohort OR prospective OR trialOR retrospective OR epidemiology) and (colorectal cancer ORcolon cancer OR rectum cancer OR large bowel cancer ORcolorectal adenomatous polyps), and (allium vegetables ORgarlic OR allium sativum OR garlic supplements OR allicingarlic OR garlic sulfur compounds OR ajoene garlic OR onionOR allium cepa OR quercetin OR leek OR scallion OR chivesOR diet OR food).” The same authors retrieved and indepen-

dently assessed potentially relevant articles reporting resultson the association between allium vegetables consumptionand colorectal cancer or colorectal adenomatous polyps oc-currence from observational studies. They also checked thereference list of all papers of interest, as well as that of some re-views on the issue [1,2,4–6] to retrieve other pertinent publica-tions. Abstracts and unpublished studies were not included.No studies were excluded a priori for weakness of design ordata quality. A flow chart of the procedure to select studies isshown in Supporting Information Figure 1.

2.2 Study selection

Studies were included if they met the following criteria: (i) hada case-control or cohort study design, (ii) the outcome was col-orectal (or colon, or rectal) cancer or colorectal adenomatouspolyps incidence/death, (iii) examined the association with al-lium vegetables (including garlic, onions, leeks, and others),(iv) provided the RR estimates with their CIs, or data nec-essary to calculate them. One study conducted as part of theprostate, lung, colorectal, and ovarian (PLCO) Cancer Screen-ing trial—a multiple investigation of the effectiveness of earlydetection of PLCO cancers—in which no dietary interventionwas assigned, was also included [13]. When multiple reportswere published on the same study population, we includedin the meta-analysis only the most informative one; threereports were therefore excluded [14–16]. Four studies wereexcluded because they did not report sufficient informationfor CI calculation [17–20].

2.3 Data extraction

All data were extracted independently in a standard formatand cross-checked by three reviewers (C. G., V. G., F. T.). Dis-agreement was resolved by discussion. The following datawere extracted from each study: last name of the first au-thor, year of publication, study design, country, number ofsubjects (cases and controls / noncases / cohort size), tu-mor site (for studies on colorectal cancer), gender, period ofenrolment (for case-control studies) or follow-up (for cohortstudies), type of allium vegetables examined (i.e., total alliumvegetables and/or individual allium vegetable items, such asgarlic, onion, and leek), consumption categories considered,estimates of the RR and their corresponding 95% CIs and, ifavailable, the number of cases and noncases/person at risk foreach category of exposure, and covariates adjusted for in theanalysis. Whenever available, we recorded the RR estimatefor garlic supplement use compared to no use. An attemptwas made to contact authors if data presentation was incom-plete or if it was necessary to resolve an apparent conflict orinconsistency in the article.

For one study reporting the adjusted OR but not the cor-responding 95% CI, we used the standard error of the corre-sponding crude OR, computed on the basis of the distribution


Mol. Nutr. Food Res. 2014, 00, 1–8 3

of cases and controls, to obtain the approximate CI for the re-ported adjusted OR [21]. In the study by Hu et al. [22], wechanged the reference category from “garlic use” to “no gar-lic use.” If a study provided more than one RR estimate, theone adjusted for the largest number of confounding factorswas included in the analysis. For colorectal cancer, availabledata allowed to assess the associations with consumption oftwo individual allium vegetable items, i.e. garlic and onion,as well as of total allium vegetables as defined in each study(total allium vegetables, hereafter).

The study by Iscovich et al. [23] examined together con-sumption of garlic, onion, and pepper, and was included inthe analysis of total allium vegetables as well. For colorectaladenomatous polyps, identified studies allowed the investiga-tion of the association with total allium vegetables only, sinceonly one study provided data for individual allium vegetableitems [24]. The study by Smith-Warner et al. [25], consideringtogether consumption of garlic, leeks, and onion, and that byMillen et al. [13], considering together consumption of gar-lic and onion, were included in the analysis of total alliumvegetables as well.

2.4 Statistical analysis

We pooled RR estimates for the comparison between thestudy-specific highest category of consumption versus thelowest using the DerSimonian and Laird random-effectsmodel, which incorporates both within- and between-studyvariability [26]. Heterogeneity among studies was assessedusing the �2 test (results were defined heterogeneous for p <

0.10) and the I2 statistic, i.e. the proportion of total variationcontributed by between-study variation [27].

In the meta-analyses on colorectal cancer, we includedstudies providing RR estimates for colorectal cancer, as well asthose providing RR estimates for colon and/or rectal cancer.When a study reported RRs for colon and rectal cancer onlyseparately, both estimates were included in the meta-analysis.The same criterion was used for studies reporting RRs formen and women separately.

Stratified analyses were conducted according to cancersite, gender, and geographic area, in order to identify possiblesources of heterogeneity. Presence of publication bias wasassessed by examination of funnel plots [28] and by applyingthe tests proposed by Begg and Mazumdar [29] and by Egger[30]. All the statistical analyses were performed using theSTATA software (version 10; StataCorp, College Station, TX,USA).

3 Results

3.1 Colorectal cancer

Sixteen studies published from 1980 to 2013, for a total of13 333 cases, were included in the meta-analyses of col-

orectal cancer [6–10, 21–23, 31–38] (Supporting InformationTable 1). Of the 16 studies, seven were population-basedcase-control studies [8, 23, 31–35], three were hospital-basedcase-control studies [9, 21, 22], and six were cohort studies[6, 7, 10, 36–38]. Four studies were conducted in Europe, fivein North America, one in South America, two in Australia,and four in Asia. The number of cases varied between 109[22] and 2280 [9] among case-control studies, and between212 [36] and 2368 [10] among cohort studies. All studies usedfood frequency questionnaires (FFQs) to assess the diet of thestudy participants. In seven out of ten case-control studies, theFFQs were administered in person by an interviewer, whereasin most cohort studies participants completed the FFQs bythemselves.

Seven studies provided information on garlic [7–10, 22,32, 36], six on onion [9, 31, 32, 35, 37, 38], and four on totalallium vegetables [6, 23, 33, 34]. Two studies reported RRs ofcolorectal cancer according to leek consumption [21, 37] andthree according garlic supplement use [7, 10, 37].

No evidence of publication bias emerged for the analysisof garlic (p for Egger’s test = 0.214; p for Begg’s test = 0.368),onion (p for Egger’s test = 0.106; p for Begg’s test = 0.260)and total allium vegetable intake (p for Egger’s test = 0.07; pfor Begg’s test = 0.174). Figure 1 shows the study-specific andpooled RRs and 95% CIs of colorectal cancer for the highestversus the lowest category of consumption of garlic, onion,and total allium vegetables. For the meta-analysis of garlic in-take (panel A), six RR estimates (from five studies) were belowunity (two were significant) and three (from two studies) wereequal to or above unity (nonsignificant), resulting in a pooledRR of 0.85 (95% CI, 0.72–1.00). There was significant hetero-geneity (p = 0.017, I2 = 57.2%). The pooled RRs were 0.76(95% CI, 0.67–0.85) for case-control (p for heterogeneity =0.413, I2 = 0.0%), and 0.99 (95% CI, 0.80–1.23) for cohortstudies (p for heterogeneity = 0.127, I2 = 47.4%). No associ-ation was found for garlic supplement use, with a pooled RRof 1.05 (95% CI, 0.84–1.32, 3 studies; data not shown).

For the analysis of onion intake, among the 11 RR esti-mates, eight (from five studies) were below unity (two weresignificant) and three (from two studies) were above unity(nonsignificant; panel B). The pooled RR for the highest ver-sus the lowest category of onion consumption was 0.85 (95%CI, 0.70–1.04), with significant heterogeneity (p = 0.046, I2 =46.1%). The corresponding pooled RRs for case-control andcohort studies were 0.74 (95% CI, 0.56–0.98, p for hetero-geneity = 0.061, I2 = 52.6%), and 1.04 (95% CI, 0.86–1.26 pfor heterogeneity = 0.562, I2 = 0.0%), respectively.

The meta-analysis of total allium vegetables intake wasbased on three case-control studies and one cohort study, andgave a pooled RR of colorectal cancer of 0.78 (95% CI, 0.56–1.08), with significant heterogeneity (p = 0.008, I2 = 70.9%;panel C).

Table 1 shows pooled RRs for colorectal cancer for thehighest versus the lowest consumption of garlic and onionintake in strata of relevant factors. The association with gar-lic, but not with onion, intake was, if anything, stronger for



Figure 1. Forest plots for the associ-ations between colorectal cancer riskand intake of garlic (panel A), onion(panel B), and total allium vegetables(panel C). Summary RRs were for thehighest versus the lowest category ofintake. CC, colon cancer; CRC, colorec-tal cancer; M, men; RC, rectal cancer;W, women.


Mol. Nutr. Food Res. 2014, 00, 1–8 5

Table 1. Summary RRs of colorectal cancer and corresponding 95% CIs for the highest versus the lowest category of garlic or onion intakein strata of selected factors

Garlic Onion

n studies RR (95% CI) p heterogeneity; I2 n studies RR (95% CI) p heterogeneity; I2

SubtypeColon 4 0.90 (0.75–1.08) 0.067; 51.5% 3 0.72 (0.44–1.19) 0.021; 65.3%Rectum 4 0.76 (0.59–0.98) 0.224; 29.6% 2 0.70 (0.39–1.25) 0.164; 44.7%

SexF 6 0.76 (0.55–1.03) 0.001; 74.5% 5 0.82 (0.52–1.29) 0.002; 74.4%M 4 0.89 (0.73–1.08) 0.224; 31.4% 5 0.84 (0.68–1.03) 0.793; 0.0%

Geographic areaNorth America 4 0.90 (0.73–1.10) 0.058; 53.1% 2 0.98 (0.82–1.17) 0.769; 0.0%Asia 1 0.22 (0.04–0.98) Not calculable - - -Europe/Australia 2 0.78 (0.68–0.89) 0.297; 7.9% 4 0.79 (0.58–0.1.09) 0.032; 54.2%

rectal than for colon cancer. For both garlic and onion, the in-verse association from studies from Europe/Australia wassomewhat stronger than that from studies from NorthAmerica, and particularly stronger in the only study fromChina [22].

3.2 Colorectal adenomatous polyps

Four studies (4873 cases) [13, 24, 25, 39] on colorectal adeno-matous polyps and allium vegetable consumption, includingthe one based on the PLCO Cancer Screening trial [13], satis-fied the inclusion criteria. All of these studies were conductedin the USA and were colonoscopy- or sigmoscopy-based case-control studies, in which cases were subjects with a positiveresult for colorectal adenoma at colonoscopy or sigmoscopy,and controls those with negative results. The MinnesotaCancer Prevention Research Unit case-control study recruitedboth colonoscopy-negative and community controls.

One study reported only results for garlic use and was notincluded the meta-analysis on total allium vegetables [24]. Noevidence of publication bias emerged (p for Egger’s test =0.756; p for Begg’s test �1).

Figure 2 shows the funnel plot for the meta-analysis oftotal allium vegetables consumption and colorectal adenoma-tous polyps. The pooled RR for the highest versus the lowestcategory of consumption was 0.88 (95% CI, 0.80–0.98), withno heterogeneity among studies (p = 0.812, I2 = 0.0%).

4 Discussion

This systematic review of epidemiological data indicates thata high intake of garlic is associated with a 15% reduction incolorectal cancer risk. There was, however, relevant hetero-geneity across studies, particularly according to study design,with cohort studies providing essentially null results. Theresults were in the same direction, but not significant, foronion. High intake of total allium vegetables is associatedwith a risk reduction of colorectal adenomatous polyps too,but this result is based on three case-control studies only.

Observational studies included in our meta-analysis, inparticular the retrospective ones, may be affected by varioussources of bias. Recall bias in studies with a retrospective de-sign may, at least in part, contribute to the discrepant resultsbetween case-control and cohort studies, given the conceiv-able influence of a recent diagnosis of colorectal cancer onthe recall of dietary habits. The over-reporting of the health-conscious cases of their consumption of “healthy foods”, suchas vegetables (and hence allium vegetables), in case-controlstudies may however hide a possible real association.

Another possible explanation of the different associationsobserved in case-control and cohort studies is that all the lat-ter were conducted in North America where diet is differentfrom Europe. Indeed, in Europe, particularly in southern Eu-rope, garlic and onion are most often eaten and cooked incombination with foods as tomatoes and olive oil in saladsand sauces of pasta [40], and some studies reported that thesynergistic action of garlic and tomato could have a preven-tive effect against the carcinogenic process [41, 42]. US percapita use of garlic has substantially risen in the last decades,starting from 0.9 pounds per person per year in 1980 to 2.3in 2012, with a peak of 3.3 pounds in 1999. Interest in garlicreflects several factors, such as the growing popularity of eth-nic foods, the perceived health-enhancing benefits, as well asthe demand from the health supplements industry. Similarly,onion use in US has almost doubled in the last two decades:per capita use of fresh onion was around 11 pounds in 1980and around 20 in 2012 [43].

Many “in vitro” and “in vivo” studies have suggesteda protective effect of allium vegetables against differenttypes of cancer, and in particular gastrointestinal can-cer. The protective effect seems to be related to theOSCs, which are either lipid or water soluble, and mostlyallyl derivates, even if the presence of other compo-nents, including flavonoids and selenium, seems to con-tribute to the antitumor effect of allium vegetables [2, 44].Garlic contains oil-soluble OSCs such as ajoene, dial-lyl sulfide, diallyl disulfide, and diallyltrisulfide, whereasonion mainly contains S-propenylcysteinesulfoxide [45], butalso other sulfoxides, including S-propylcysteinesulfoxide



Figure 2. Forest plot for the association between colorectal adenomatous polyps risk and intake of total allium vegetables. Summary RRswere for the highest versus the lowest category of intake.

and S-methylcysteinesulfoxide. Preclinical studies foundgarlic to be protective when administered during the initi-ation phase of aberrant crypt foci (preneosplastic lesions forcolon cancer) induction [46,47], but only diallyl disulfide waseffective in postinitiation phase studies [47]. Also, onion sig-nificantly inhibited chemically induced aberrant crypt fociformation in preclinical studies, both when onion was givenduring the initiation phase and when it was given during thepostinitiation phase of carcinogenesis [48].

In addition to the allyl sulfur contents, the presence ofseveral other factors, including selenium and flavonoids, mayalso influence several cellular processes that have been relatedexperimentally to cancer incidence and prognosis [49, 50].

Still, the mechanisms of the cancer-preventive effectsof allium vegetables, in particular of OSCs, are not clear,although several hypotheses have been proposed. In partic-ular, the most convincing mechanisms are the inhibitionof carcinogen-induced DNA adduct formation [51], theinhibition of tumor cell proliferation by apoptosis andalteration of the cell cycle [52], the modulation of enzymeactivities by enhancement of carcinogen-detoxifying en-zymes (glutathione S-transferase) [53], or the suppressionof carcinogen-activating enzymes (cytochrome P450) [54].Furthermore, the presence of selenium content seems tocontribute to the anticarcinogenic effect of allium vegetablesby multiple mechanisms. These include effects on geneexpression, DNA damage and repair, signaling pathways,regulation of cell cycle and apoptosis, metastasis andangiogenesis, and oxidative stress [55].

In our meta-analysis, we pooled RR estimates for the high-est versus the lowest categories of exposure in each study,although some reports did not provide cutoff levels and therewas great heterogeneity of frequency categories among thehighest ones, since the pattern of allium vegetable consump-tion appreciably differs by geographic area. However, this ap-proach has been previously adopted by other meta-analyses ofdifferent dietary exposures in relation to cancer [56–58], andcombining RRs for the highest versus the lowest allium veg-etable categories has previously shown inverse associationswith cancers of the colorectum and stomach [5, 59]. Anotherlimitation is that data from original studies did not allow to

conduct a dose-risk analysis, given the difficulties in convert-ing in a standard unit of measurement the doses of expo-sure reported in each study (e.g., “servings” and “portions”).Further, results from stratified analyses including very fewstudies are of questionable interpretation, and any inferenceremains therefore speculative.

It is possible that residual confounding accounts for somedegree of the association observed with garlic and onion, sincea high allium vegetables intake may be an indirect indicatorof a favorable diet for colorectal cancer. Indeed, total vegetableconsumption has been favorably related to colorectal cancer[3, 60] and correlates with allium vegetables intake.

Another possible limitation is that we have no informa-tion on the modalities of garlic and onion consumption, andthere are important varietal differences in the composition,concentration, and beneficial activities of these bioactive com-pounds, e.g. by modalities of cooking.

The major strengths of our study are the fact that we fol-lowed the recommendations for reporting developed by theMOOSE group [12] and that we used a search string notexclusively relying to terms specifically related to allium veg-etables (e. g., “allium vegetable,” “onion,” “garlic”) but whichincluded also the concept terms “diet” and “food.” This al-lowed to identify not only papers providing information onallium vegetables (and/or its individual items) in the titleor abstract, but also papers with information on allium veg-etables in the main text or tables only. Through this searchstrategy, more than 200 full texts were examined.

In conclusion, a high consumption of garlic appears to re-duce the risk of colorectal cancer. However, evidence of suchprotection derived mainly from case-control studies, with co-hort studies finding essentially null results. Therefore, theseresults have to be taken with due caution. Moreover, highintake of total allium vegetables may be associated with a riskreduction of colorectal adenomatous polyps.

Further epidemiological data, particularly from largeprospective cohorts, may help to elucidate and quantify thepossible protection exerted by allium vegetables on colorec-tal cancer risk. In particular, future investigations shouldconcentrate on identifying which are the types of alliumvegetables, the amounts and the modality of consumption


Mol. Nutr. Food Res. 2014, 00, 1–8 7

(i.e., preparation practices and cooking methods) effective forcancer prevention.

This work was conducted with the contribution of the ItalianFoundation for Cancer Research (FIRC). F. T. was supportedby a FIRC fellowship. C. G. was supported by Fondazione Um-berto Veronesi. The authors thank Mrs I. Garimoldi for editorialassistance.

The authors have declared no conflict of interest.

5 References

[1] Ngo, S. N., Williams, D. B., Cobiac, L., Head, R. J., Does garlicreduce risk of colorectal cancer? A systematic review. J. Nutr.2007, 137, 2264–2269.

[2] Bianchini, F., Vainio, H., Allium vegetables and organosulfurcompounds: do they help prevent cancer? Environ. HealthPerspect. 2001, 109, 893–902.

[3] World Cancer Research Fund and American Institute forCancer Research, Food, Nutrition, Physical Activity and thePrevention of Cancer: a Global Perspective, AICR, Washing-ton, DC 2007.

[4] Fleischauer, A. T., Arab, L., Garlic and cancer: a critical reviewof the epidemiologic literature. J. Nutr. 2001, 131, 1032S–1040S.

[5] Fleischauer, A. T., Poole, C., Arab, L., Garlic consumption andcancer prevention: meta-analyses of colorectal and stomachcancers. Am. J. Clin. Nutr. 2000, 72, 1047–1052.

[6] Vogtmann, E., Xiang, Y. B., Li, H. L., Levitan, E. B. et al. Fruitand vegetable intake and the risk of colorectal cancer: re-sults from the Shanghai Men’s Health Study. Cancer CausesControl 2013, 24, 1935–1945.

[7] McCullough, M. L., Jacobs, E. J., Shah, R., Campbell, P.T. et al., Garlic consumption and colorectal cancer risk inthe CPS-II nutrition cohort. Cancer Causes Control 2012, 23,1643–1651.

[8] Annema, N., Heyworth, J. S., McNaughton, S. A., Iacopetta,B. et al., Fruit and vegetable consumption and the risk ofproximal colon, distal colon, and rectal cancers in a case-control study in Western Australia. J. Am. Diet Assoc. 2011,111, 1479–1490.

[9] Galeone, C., Pelucchi, C., Levi, F., Negri, E. et al., Onion andgarlic use and human cancer. Am. J. Clin. Nutr. 2006, 84,1027–1032.

[10] Meng, S., Zhang, X., Giovannucci, E. L., Ma, J. et al., No as-sociation between garlic intake and risk of colorectal cancer.Cancer Epidemiol. 2013, 37, 152–155.

[11] Giovannucci, E., Wu, K., in: Schottenfeld, D., Fraumeni, J.F., Jr. (Eds.), Cancer Epidemiology and Prevention, OxfordUniversity Press, New York 2006, pp. 809–829.

[12] Stroup, D. F., Berlin, J. A., Morton, S. C., Olkin, I. et al., Meta-analysis of observational studies in epidemiology: a pro-posal for reporting. Meta-analysis Of Observational Studiesin Epidemiology (MOOSE) group. JAMA 2000, 283, 2008–2012.

[13] Millen, A. E., Subar, A. F., Graubard, B. I., Peters, U. et al., Fruitand vegetable intake and prevalence of colorectal adenomain a cancer screening trial. Am. J. Clin. Nutr. 2007, 86, 1754–1764.

[14] Levi, F., Pasche, C., La Vecchia, C., Lucchini, F., Franceschi, S.,Food groups and colorectal cancer risk. Br. J. Cancer 1999,79, 1283–1287.

[15] Sellers, T. A., Bazyk, A. E., Bostick, R. M., Kushi, L. H.et al., Diet and risk of colon cancer in a large prospectivestudy of older women: an analysis stratified on family his-tory (Iowa, United States). Cancer Causes Control 1998, 9,357–367.

[16] Giovannucci, E., Rimm, E. B., Stampfer, M. J., Colditz, G. A.et al., Intake of fat, meat, and fiber in relation to risk of coloncancer in men. Cancer Res. 1994, 54, 2390–2397.

[17] Kotzev, I., Mirchev, M., Manevska, B., Ivanova, I. et al. Ka, Riskand protective factors for development of colorectal polypsand cancer (Bulgarian experience). Hepatogastroenterology2008, 55, 381–387.

[18] Haenszel, W., Berg, J. W., Segi, M., Kurihara, M. et al., Large-bowel cancer in Hawaiian Japanese. J. Natl. Cancer Inst.1973, 51, 1765–1779.

[19] Tajima, K., Tominaga, S., Dietary habits and gastro-intestinalcancers: a comparative case-control study of stomach andlarge intestinal cancers in Nagoya, Japan. Jpn. J. CancerRes. 1985, 76, 705–716.

[20] Tuyns, A. J., Kaaks, R., Haelterman, M., Colorectal cancer andthe consumption of foods: a case-control study in Belgium.Nutr. Cancer 1988, 11, 189–204.

[21] Haenszel, W., Locke, F. B., Segi, M., A case-control study oflarge bowel cancer in Japan. J. Natl. Cancer Inst. 1980, 64,17–22.

[22] Hu, J. F., Liu, Y. Y., Yu, Y. K., Zhao, T. Z. et al., Diet and cancerof the colon and rectum: a case-control study in China. Int.J. Epidemiol. 1991, 20, 362–367.

[23] Iscovich, J. M., L’Abbe, K. A., Castelleto, R., Calzona, A. et al.,Colon cancer in Argentina. I: risk from intake of dietary items.Int. J. Cancer. 1992, 51, 851–857.

[24] Witte, J. S., Longnecker, M. P., Bird, C. L., Lee, E. R. et al.,Relation of vegetable, fruit, and grain consumption to col-orectal adenomatous polyps. Am. J. Epidemiol. 1996, 144,1015–1025.

[25] Smith-Warner, S. A., Elmer, P. J., Fosdick, L., Randall, B. et al.,Fruits, vegetables, and adenomatous polyps: the MinnesotaCancer Prevention Research Unit case-control study. Am. J.Epidemiol. 2002, 155, 1104–1113.

[26] DerSimonian, R., Laird, N., Meta-analysis in clinical trials.Control Clin. Trials 1986, 7, 177–188.

[27] Higgins, J. P., Thompson, S. G., Quantifying heterogeneityin a meta-analysis. Stat. Med. 2002, 21, 1539–1558.

[28] Sterne, J. A., Egger, M., Funnel plots for detecting bias inmeta-analysis: guidelines on choice of axis. J. Clin. Epi-demiol. 2001, 54, 1046–1055.

[29] Begg, C. B., Mazumdar, M., Operating characteristics of arank correlation test for publication bias. Biometrics 1994,50, 1088–1101.



[30] Egger, M., Davey Smith, G., Schneider, M., Minder, C., Biasin meta-analysis detected by a simple, graphical test. BMJ1997, 315, 629–634.

[31] Steinmetz, K. A., Potter, J. D., Food-group consumption andcolon cancer in the Adelaide Case-Control Study. I. Vegeta-bles and fruit. Int. J. Cancer 1993, 53, 711–719.

[32] Le Marchand, L., Hankin, J. H., Wilkens, L. R., Kolonel, L. N.et al., Dietary fiber and colorectal cancer risk. Epidemiology1997, 8, 658–665.

[33] Chiu, B. C., Ji, B. T., Dai, Q., Gridley, G. et al., Dietary fac-tors and risk of colon cancer in Shanghai, China. CancerEpidemiol. Biomarkers Prev. 2003, 12, 201–208.

[34] Diergaarde, B., van Geloof, W. L., van Muijen, G. N., Kok, F. J.et al., Dietary factors and the occurrence of truncating APCmutations in sporadic colon carcinomas: a Dutch population-based study. Carcinogenesis 2003, 24, 283–290.

[35] Theodoratou, E., Kyle, J., Cetnarskyj, R., Farrington, S.M. et al., Dietary flavonoids and the risk of colorec-tal cancer. Cancer Epidemiol. Biomarkers Prev. 2007, 16,684–693.

[36] Steinmetz, K. A., Kushi, L. H., Bostick, R. M., Folsom,A. R. et al., Vegetables, fruit, and colon cancer in theIowa Women’s Health Study. Am. J. Epidemiol. 1994, 139,1–15.

[37] Dorant, E., van den Brandt, P. A., Goldbohm, R. A., A prospec-tive cohort study on the relationship between onion and leekconsumption, garlic supplement use and the risk of colorec-tal carcinoma in The Netherlands. Carcinogenesis 1996, 17,477–484.

[38] Lin, J., Zhang, S. M., Wu, K., Willett, W. C. et al., Flavonoidintake and colorectal cancer risk in men and women. Am. J.Epidemiol. 2006, 164, 644–651.

[39] Wu, H., Dai, Q., Shrubsole, M. J., Ness, R. M. et al., Fruit andvegetable intakes are associated with lower risk of colorectaladenomas. J. Nutr. 2009, 139, 340–344.

[40] Pelucchi, C., Bosetti, C., Rossi, M., Negri, E. et al., Selectedaspects of Mediterranean diet and cancer risk. Nutr. Cancer2009, 61, 756–766.

[41] Sengupta, A., Ghosh, S., Das, S., Modulatory influ-ence of garlic and tomato on cyclooxygenase-2 activity,cell proliferation and apoptosis during azoxymethane in-duced colon carcinogenesis in rat. Cancer Lett .2004, 208,127–136.

[42] Bhuvaneswari, V., Abraham, S. K., Nagini, S., Combinato-rial antigenotoxic and anticarcinogenic effects of tomatoand garlic through modulation of xenobiotic-metabolizingenzymes during hamster buccal pouch carcinogenesis. Nu-trition 2005, 21, 726–731.

[43] Economic Research Service, USDA. Vegetables and PulsesYearbook Data/#89011/May 31 2012. Available at: http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=1858 (Last accessed April 28, 2014).

[44] Wargovich, M. J., Uda, N., Woods, C., Velasco, M. et al.,Allium vegetables: their role in the prevention of cancer.Biochem. Soc. Trans. 1996, 24, 811–814.

[45] Block, E., The chemistry of garlic and onions. Sci. Am. 1985,252, 114–119.

[46] Wargovich, M. J., Chen, C. D., Jimenez, A., Steele, V. E. et al.,Aberrant crypts as a biomarker for colon cancer: evaluationof potential chemopreventive agents in the rat. Cancer Epi-demiol. Biomarkers Prev. 1996, 5, 355–360.

[47] Wargovich, M. J., Jimenez, A., McKee, K., Steele, V. E. et al.,Efficacy of potential chemopreventive agents on rat colonaberrant crypt formation and progression. Carcinogenesis2000, 21, 1149–1155.

[48] Tache, S., Ladam, A., Corpet, D. E., Chemoprevention of aber-rant crypt foci in the colon of rats by dietary onion. Eur. J.Cancer. 2007, 43, 454–458.

[49] Dong, Y., Lisk, D., Block, E., Ip, C., Characteriza-tion of the biological activity of gamma-glutamyl-Se-methylselenocysteine: a novel, naturally occurring anti-cancer agent from garlic. Cancer Res. 2001, 61, 2923–2928.

[50] Milner, J. A., Preclinical perspectives on garlic and cancer.J. Nutr. 2006, 136, 827S–831S.

[51] Amagase, H., Milner, J. A., Impact of various sources of gar-lic and their constituents on 7,12-dimethylbenz[a]anthracenebinding to mammary cell DNA. Carcinogenesis 1993, 14,1627–1631.

[52] Hosono, T., Fukao, T., Ogihara, J., Ito, Y. et al., Diallyl trisul-fide suppresses the proliferation and induces apoptosis ofhuman colon cancer cells through oxidative modification ofbeta-tubulin. J. Biol. Chem. 2005, 280, 41487–41493.

[53] Hatono, S., Jimenez, A., Wargovich, M. J., Chemopreventiveeffect of S-allylcysteine and its relationship to the detoxi-fication enzyme glutathione S-transferase. Carcinogenesis1996, 17, 1041–1044.

[54] Davenport, D. M., Wargovich, M. J., Modulation of cy-tochrome P450 enzymes by organosulfur compounds fromgarlic. Food Chem. Toxicol. 2005, 43, 1753–1762.

[55] Jackson, M. I., Combs, G. F., Jr., Selenium and anticar-cinogenesis: underlying mechanisms. Curr. Opin. Clin. Nutr.Metab. Care 2008, 11, 718–726.

[56] Turati, F., Galeone, C., Edefonti, V., Ferraroni, M. et al., Ameta-analysis of coffee consumption and pancreatic cancer.Ann. Oncol. 2012, 23, 311–318.

[57] Galeone, C., Augustin, L. S., Filomeno, M., Malerba, S. et al.,Dietary glycemic index, glycemic load, and the risk of en-dometrial cancer: a case-control study and meta-analysis.Eur. J. Cancer Prev. 2013, 22, 38–45.

[58] Bertuccio, P., Rosato, V., Andreano, A., Ferraroni, M. et al.,Dietary patterns and gastric cancer risk: a systematic reviewand meta-analysis. Ann. Oncol. 2013, 24, 1450–1458.

[59] Zhou, Y., Zhuang, W., Hu, W., Liu, G. J. et al., Consumption oflarge amounts of Allium vegetables reduces risk for gastriccancer in a meta-analysis. Gastroenterology 2011, 141, 80–89.

[60] Franceschi, S., Parpinel, M., La Vecchia, C., Favero, A. et al.,Role of different types of vegetables and fruit in the preven-tion of cancer of the colon, rectum, and breast. Epidemiology1998, 9, 338–341.


http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=1858



Documents

Colorectal cancer and adenomatous polyps in relation to allium vegetables intake: A meta-analysis of observational studies