Advertisement

Breast Cancer Research and Treatment

, Volume 134, Issue 2, pp 479–493 | Cite as

Fruits, vegetables and breast cancer risk: a systematic review and meta-analysis of prospective studies

  • D. Aune
  • D. S. M. Chan
  • A. R. Vieira
  • D. A. Navarro Rosenblatt
  • R. Vieira
  • D. C. Greenwood
  • T. Norat
Review

Abstract

Evidence for an association between fruit and vegetable intake and breast cancer risk is inconclusive. To clarify the association, we conducted a systematic review and meta-analysis of the evidence from prospective studies. We searched PubMed for prospective studies of fruit and vegetable intake and breast cancer risk until April 30, 2011. We included fifteen prospective studies that reported relative risk estimates and 95 % confidence intervals (CIs) of breast cancer associated with fruit and vegetable intake. Random effects models were used to estimate summary relative risks. The summary relative risk (RR) for the highest versus the lowest intake was 0.89 (95 % CI: 0.80–0.99, I 2 = 0 %) for fruits and vegetables combined, 0.92 (95 % CI: 0.86–0.98, I 2 = 9 %) for fruits, and 0.99 (95 % CI: 0.92–1.06, I 2 = 20 %) for vegetables. In dose–response analyses, the summary RR per 200 g/day was 0.96 (95 % CI: 0.93–1.00, I 2 = 2 %) for fruits and vegetables combined, 0.94 (95 % CI: 0.89–1.00, I 2 = 39 %) for fruits, and 1.00 (95 % CI: 0.95–1.06, I 2 = 17 %) for vegetables. In this meta-analysis of prospective studies, high intake of fruits, and fruits and vegetables combined, but not vegetables, is associated with a weak reduction in risk of breast cancer.

Keywords

Fruits Vegetables Breast cancer Systematic review Meta-analysis 

Introduction

Breast cancer is the most common cause of cancer in women, with 1.38 million new cases diagnosed in 2008 worldwide, accounting for about 23 % of all cancer cases and 14 % of all cancer deaths among women [1]. The large international variation in breast cancer rates, coupled with the rapidly increasing rates observed in secular trend studies [2, 3] and migration studies [4, 5], suggest the importance of modifiable risk factors in breast cancer etiology.

Although dietary factors have long been suspected to be implicated in breast cancer etiology, few convincing dietary risk factors have been identified [6]. Fruits and vegetables contain numerous constituents that may reduce breast cancer risk, including fiber which can bind estrogens during the enterohepatic circulation [7], and antioxidants and several vitamins which can prevent oxidative DNA damage [8]. However, epidemiological studies of fruit and vegetable intake and breast cancer risk have provided inconsistent results. Case–control studies have generally found reduced breast cancer risk with high intake of fruits and vegetables [9], however, the interpretation of these studies, which may have been affected by recall bias and selection bias, have made conclusions difficult. This, in particular because most [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23], but not all [24] prospective studies (which are less prone to such biases) in contrast have found no statistically significant association between fruit or vegetable intake and breast cancer risk. In the World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) report from 2007, it was stated that the evidence for an association between fruit and vegetable intake and breast cancer risk was too limited or inconsistent for a conclusion to be made. At least seven prospective studies have reported results for fruit and vegetable intake and breast cancer risk since that report [18, 19, 20, 21, 22, 23, 24], and this should provide even more statistical power to detect an association. Thus, we aimed to clarify the evidence by conducting a systematic review and meta-analysis of the evidence from prospective studies.

Methods

Search strategy

As part of the Continuous Update Project of the WCRF/AICR, we updated the systematic literature review published in 2007 [6] and searched the PubMed database up to April 30, 2011 for studies of fruit and vegetable intake and breast cancer risk. We followed a prespecified protocol, which includes details of the search terms used, for the review (http://www.dietandcancerreport.org/downloads/SLR_Manual.pdf). The reference lists of all the included studies and the reference lists of the published systematic reviews and meta-analyses were also searched for any additional studies [6, 9, 25, 26, 27]. We followed standard criteria for conducting and reporting meta-analyses [28].

Study selection

To be included, the study had to have a prospective cohort, case-cohort, or nested case–control design and to investigate the association between the intake of fruits and vegetables and breast cancer incidence. Estimates of the relative risk (RR) (such as hazard ratio or risk ratio) and 95 % confidence intervals (CIs) had to be available in the publication. For the dose–response analysis, a quantitative measure of intake and the total number of cases and person-years had to be available in the publication. When multiple publications from the same study were available, we used the publication with the largest number of cases. We identified 26 potentially eligible full-text publications [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39]. We excluded three publications on breast cancer mortality [29, 30, 31], six duplicate publications, [33, 34, 35, 36, 37, 38] and two studies of childhood [32] and adolescent dietary intake [39]. One study was excluded from the dose–response analysis because the comparison was provided only for the highest versus the lowest intake [19]. In total, 15 publications were included in the analyses (Fig. 1; Table 1).
Fig. 1

Flow-chart of study selection

Table 1

Prospective studies of fruits, vegetables intake and breast cancer risk

Author, publication year, country/region

Study name

Follow-up period

Study size, gender, age, number of cases

Dietary assessment

Exposure

Quantity

RR (95 % CI)

Adjustment for confounders

Lof M et al., 2011, Sweden

Swedish Women’s Lifestyle and Health Cohort Study

1991–1992–2006, 14 years

44,848 pre- & postm. w., age 30–49 years: 1,067 cases

Validated FFQ, ~80 items

Fruits and vegetables

Per 200 g/day

0.94 (0.86–1.03)

Age, education, BMI, smoking, energy intake, alcohol

Brasky TM et al., 2010, USA

VITamins And Lifestyle (VITAL) Cohort

2000–2002–2007, 6 years

35,016 postm. w., age 50–76 years: 880 cases

Validated FFQ, 120 items

Fruits

>2.14 vs. ≤1.04 serv/day

0.86 (0.73–1.02)

Age

Vegetables

>2.85 vs. ≤1.73 serv/day

0.97 (0.82–1.15)

Boggs DA et al., 2010, USA

Black Women’s Health Study

1995–2007, 12 years

51,928 pre- & postm. w., age 21–69 years: 1,268 cases

Validated FFQ, 68/85 items

Fruits and vegetables

≥4 vs. <1 serv/day

0.87 (0.71–1.07)

Age, energy intake, age at menarche, BMI at age 18 years, FH-BC, education, geographic location, parity, age at 1st birth, OC use, menopausal status, age at menopause, menopausal hormone use, vigorous activity, smoking status, alcohol intake, multivitamin use

Total vegetables

≥2 serv/day vs. <4/week

0.87 (0.73–1.05)

Total fruits

≥2 serv./day vs. <2/week

0.91 (0.74–1.11)

Butler LM et al., 2010, Singapore

Singapore Chinese Health Study

1993–1998–2005, 10.7 years

34,028 postm. w., age 45–74 years: 629 cases

Validated FFQ, 165 items

Total vegetables

173.7 vs. 51.0 g/day

0.86 (0.63–1.16)

Age, dialect group, interview year, education, parity, BMI, 1st degree relative with BC, total energy

Total fruits

357.0 vs. 39.0 g/day

1.03 (0.77–1.38)

Jayalekshmi P et al., 2009, India

Karunagappally Cohort

1990–2004, 14 years

792 pre-& postm. controls, age ≥20 years: 264 cases

FFQ

Vegetables

Occasional vs. regular

0.71 (0.49–1.06)

Age, religion, place of residence

George SM et al., 2009, USA

NIH-AARP Diet and Health Study

1995–1996–2003, 8 years

195,229 postm. w., age 50–71 years: 5,815 cases

Validated FFQ, 124 food items

Total fruit

≥1.90 vs. ≤0.60 cup equiv/day

0.91 (0.84–1.00)

Age, smoking, energy intake, BMI, alcohol, physical activity, education, race, marital status, FH-cancer, menopausal HT, mutual adjustment between fruit and vegetables

Total vegetables

≥1.43 vs. ≤0.56 cup equiv/day

1.08 (1.00–1.18)

Sonestedt E et al., 2008, Sweden

Malmo Diet and Cancer Study

1991–1996–2004, 10.3 years

15,773 pre- & postm. w, age 46–75 years: 544 cases

Validated assessment; 7 day menu book, 168 item FFQ and 1 h interview

Fruits, berries,

629 vs. 118 g/day

0.78 (0.59–1.03)

Age, season of data collection, diet interviewer, method version, total energy, weight, height, educational status, smoking habits, leisure-time physical activity, hours of household activities, alcohol, age at menopause, parity, current use of HRT

vegetables

Van Gils CH et al., 2005, Europe

European Prospective Investigation into Cancer and Nutrition

1992–2001, 5.4 years

285,526 pre- & postm. w., age 25–70 years: 3,659 cases

Validated FFQs, ≤350 items, dietary interview, diet history, 7 day menu book, 7 day record

Total vegetables

245.95 vs. 122.22 g/day

0.98 (0.84–1.14)

Age, center, energy intake (divided into fat and nonfat sources), alcohol intake, SFA intake, height, weight, age at menarche, parity, current OC use, current HRT use, menopausal status, smoking status, physical activity, education

Total fruits

372.17 vs. 115.39 g/day

1.09 (0.94–1.25)

Olsen A et al., 2003, Denmark

Diet, Cancer and Health

1993–1997–2000, 4.7 years

23,798 postm. w., age 50–64 years: 425 cases

Validated FFQ, 192 food items

Total fruits, vegetables and juice

Per 100 g/day

1.02 (0.98–1.06)

Age, time under study, parity, previous benign breast tumor surgery, education, HRT use and duration, alcohol, BMI

Zhang S et al., 1999, USA

Nurses’ Health Study

1980–1994, 14 years

83,234 pre- & postm. w., age 33–60 years: 2,697 cases

Validated FFQ, 61/126 items

Prem: Fruits

≥5.0 vs. <2 serv/day

0.74 (0.45–1.24)

Age, length of follow-up, energy intake, age at 1st birth, age at menarche, FH-BC, benign breast disease, alcohol, BMI at age 18 years, weight change from age 18 years, height. Postm.women: age at menopause and HRT

Vegetables

≥5.0 vs. <2 serv/day

0.64 (0.43–0.95)

Fruits and vegetables

≥5.0 vs. <2 serv/day

0.77 (0.58–1.02)

Postm: Fruits

≥5.0 vs. <2 serv/day

0.84 (0.64–1.09)

Vegetables

≥5.0 vs. <2 serv/day

1.02 (0.85–1.24)

Fruits and vegetables

≥5.0 vs. <2 serv/day

1.03 (0.81–1.31)

Key TJ et al., 1999, Japan

Life Span Study

1969–1970, 1979–1980–1993, 14 years

34,759 pre- & postm. w: 427 cases

FFQ, 19 items

Fruits

≥5/week vs. ≤1/week

0.95 (0.71–12.7)

Age, calendar period, city, age at time of bombing and radiation dose

Verhoeven DTH et al., 1997, Netherlands

Netherlands Cohort Study

1986–1990, 4.3 years

1812 postm. w., age 55–69 years: 650 cases

Validated FFQ, 150 food items

Vegetables

303 vs. 108 g/day

0.94 (0.67–1.31)

Age, energy intake, alcohol intake, benign breast disease, maternal breast cancer, breast cancer in sister(s), age at menarche, age at menopause, age at first birth, parity

Fruits

343.1 vs. 64.9 g/day

0.76 (0.54–1.08)

Byrne C et al., 1996, USA

National Health Epidemiologic Follow-up Study

1982–1984 –NA, 3.9 years

6156 pre- & postm. w., age 32–86 years: 53 cases

FFQ, 93 food items

Fruits and vegetables

>3 vs. ≤3 serv/day

0.7 (0.4–1.5)

Age

Rohan T et al., 1993, Canada

National Breast Screening Study

1982–1987, ~5 years

56,837 pre- & postm. w., age 40–59 years: 519 cases

Validated FFQ, 86 food items

Fruit

≥491 vs. <189 g/day

0.81 (0.57–1.14)

Age, age at menarche, FH-BC, surgical menopause, age at 1st livebirth, years of education, benign breast disease, other contributors to total food intake

Vegetables

≥433 vs. <203 g/day

0.86 (0.61–1.23)

Shibata et al., 1992, USA

Leisure World Cohort study

1981–1985–1989, 6 years

~7,299 postm. w., age 65–84 years: 219 cases

FFQ, 59 food items

Vegetables and fruit

10.06 vs. 4.54 serv/day

0.87 (0.63–1.21)

Age, smoking

Vegetables

5.98 vs. 2.34 serv/day

0.96 (0.69–1.34)

Fruit

4.58 vs. 1.66 serv/day

0.82 (0.60–1.12)

BMI  Body Mass Index, FFQ  food frequency questionnaire, FH-BC Family history of breast cancer, HRT/HT hormone therapy, MET metabolic equivalent task, OC use oral contraceptive use, prem premenopausal, postm. postmenopausal, w women, SFA saturated fatty acids, yrs years

Data extraction

We extracted the following data from each study: first author’s last name, publication year, country where the study was conducted, study name, follow-up period, sample size, gender, age, number of cases, dietary assessment method (type, number of items and whether it was validated), exposure, frequency or quantity of intake, RRs, and 95 % CIs and variables adjusted for in the analysis. The search and data extraction of articles published up to December 30, 2005 was conducted by several reviewers at the Istituto Nazionale Tumori Milan during the systematic literature review for the WCRF/AICR report (http://www.dietandcancerreport.org/downloads/SLR/Breast_SLR.pdf). The search from January 2006 to April 30, 2011 was conducted by two of the authors (D. S. M. C. and A. R. V). Data were extracted into a database by two authors (D. S. M. C., and A. R. V.) and were checked for accuracy by two authors (D. A. and T. N). We did not assess study quality using a quality score, but investigated whether specific study characteristics such as duration of follow-up, number of cases, menopausal status, and adjustment for confounders, which are indicators of study quality, influenced the results in subgroup analyses.

Statistical methods

To take into account heterogeneity between studies, we used random effects models to calculate summary RRs and 95 % CIs for the highest versus the lowest level of fruit and vegetable intake and for the dose–response analysis [40]. The average of the natural logarithm of the RRs was estimated and the RR from each study was weighted by the inverse of its variance. A two-tailed p < 0.05 was considered statistically significant.

For the dose–response analysis, we used the method described by Greenland and Longnecker [41] to compute linear trends and 95 % CIs from the natural logs of the RRs and the CIs across categories of fruit and vegetable intake. The method requires that the distribution of cases and person-years, or non-cases and the RRs with the variance estimates for at least three quantitative exposure categories are known. We estimated the distribution of cases or person-years in studies that did not report these, but reported the total number of cases/person-years. For example, if the total number of person-years was provided and the exposure variable was categorized by quintiles, we divided the number of person-years by five. The median or mean level of fruit and vegetable intake in each category of intake was assigned to the corresponding RR for each study when provided in the paper. For studies that reported fruit and vegetable intake by ranges of intake we estimated the mean intake in each category by calculating the average of the lower and upper bound. When the highest or lowest category was open-ended we assumed the open-ended interval length to be the same as the adjacent interval. Consistent with previous meta-analyses of fruit and vegetable intake and cancer risk [26, 42], we used 80 g as a serving size for recalculation of the intakes to a common scale [grams per day (g/day)] in studies that reported intakes as frequency. We contacted the authors of one study that reported results in cup equivalents to retrieve intakes in g/day [24]. The linear dose–response results are presented for a 200 g/day increment. We examined a potential nonlinear dose–response relationship between fruit and vegetable intakes and breast cancer using fractional polynomial models [43]. We determined the best fitting second order fractional polynomial regression model, defined as the one with the lowest deviance. A likelihood ratio test was used to assess the difference between the nonlinear and linear models to test for nonlinearity [43]. In the analysis of total fruits and vegetables combined, we used 100 g/day as a reference category because there were no studies with zero intake in the reference.

Heterogeneity between studies was assessed using Q and I 2 statistics [44]. Potential sources of heterogeneity were investigated in subgroup and meta-regression analyses by menopausal status, duration of follow-up, number of cases, geographic location, and adjustment for confounding factors. Small study bias, such as publication bias, was assessed using a funnel plot and Egger’s test [45] with results considered to indicate potential small study bias when p < 0.10. In a sensitivity analysis, we examined the impact of including studies of breast cancer mortality on the results as well.

Stata version 10.1 software (StataCorp, College Station, TX, USA) was used for the statistical analyses.

Results

We identified 14 cohort studies [10, 11, 12, 13, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24] and one nested case–control study [19] that were included in the analysis of fruit and/or vegetable intake and breast cancer risk (Table 1; Fig. 1). Five of the studies were from Europe, seven from America, and three from Asia (Table 1).

Total fruits and vegetables

High versus low analysis

Seven cohort studies [10, 12, 14, 16, 18, 21, 23] investigated the association between total fruit and vegetable intakes and breast cancer risk, and included 6,273 cases among 233,036 participants. Six of these studies [10, 12, 14, 16, 18, 21] were included in the high versus low analysis (one study reported only continuous results [23]). The summary RR for high versus low intake was 0.89 (95 % CI: 0.80–0.99), with no heterogeneity, I 2 = 0 % and p heterogeneity = 0.67 (n = 6) (Fig. 2a). There was no evidence of publication bias with Egger’s test, p = 0.44.
Fig. 2

Fruits and vegetables and breast cancer

Dose–response analysis

Six cohort studies [10, 14, 16, 18, 21, 23] were included in the dose–response analysis. The summary RR per 200 g/day was 0.96 (95 % CI: 0.93–1.00, p for association = 0.045), with no evidence of heterogeneity, I 2 = 2 % and p heterogeneity = 0.41 (n = 6) (Fig. 2b). There was no evidence of a nonlinear association between total fruits and vegetables and breast cancer risk, p nonlinearity = 0.20 (Fig. 3).
Fig. 3

Fruits, vegetables and breast cancer, nonlinear dose–response analysis

Fruits

High versus low analysis

Ten cohort studies [10, 11, 13, 14, 15, 17, 20, 21, 22, 24] were included in the analysis of fruit intake and breast cancer risk, including 16,763 cases among 785,668 participants. The summary RR for high versus low intake was 0.92 (95 % CI: 0.86–0.98), with little heterogeneity, I 2 = 9 %, and p heterogeneity = 0.36 (n = 10) (Fig. 4a). There was no evidence of publication bias with Egger’s test, p = 0.41.
Fig. 4

Fruits and breast cancer

Dose–response analysis

The summary RR per 200 g/day was 0.94 (95 % CI: 0.89–1.00), with low heterogeneity, I 2 = 39 %, and p heterogeneity = 0.10 (n = 10) (Fig. 4b). There was no evidence of a nonlinear association between fruit intake and breast cancer risk, p nonlinearity = 0.60 (Fig. 5a).
Fig. 5

Fruits and vegetables and breast cancer, nonlinear dose–response analyses

Vegetables

High versus low analysis

Nine cohort studies [10, 11, 13, 14, 17, 20, 21, 22, 24] and one nested case–control study [19] were included in the analysis of high versus low vegetable intake and breast cancer, including 16,600 cases among 751,965 participants. The summary RR was 0.99 (95 % CI: 0.92–1.06). There was little evidence of heterogeneity, I 2 = 20 %, and p heterogeneity = 0.26 (Fig. 6a). There was no evidence of small-study bias with Egger’s test, p = 0.23.
Fig. 6

Vegetables and breast cancer

Dose–response analysis

Nine cohort studies [10, 11, 13, 14, 17, 19, 20, 21, 22, 24] were included in the dose–response analysis. The summary RR per 200 g/day was 1.00 (95 % CI: 0.95–1.06) with little evidence of heterogeneity, I 2 = 17 %, and p heterogeneity = 0.29 (Fig. 6b). There was no evidence of a nonlinear association between vegetable intake and breast cancer risk, p nonlinearity = 0.33 (Fig. 5b).

Subgroup and sensitivity analyses

In stratified analyses (Table 2), the association between high versus low fruit intake and breast cancer risk was inverse in most strata, although usually not statistically significant. There was marginally significant heterogeneity (p = 0.06) in the results for vegetables among pre- and postmenopausal women, with a significant inverse association among premenopausal, but not postmenopausal women, however, there was only two studies among premenopausal women (Table 2). Too few studies reported results stratified by hormone receptor status to conduct subgroup analyses of these. For fruits, or fruits and vegetables combined, there was no evidence of a difference in the results by menopausal status, although the inverse association with fruit intake only was significant among postmenopausal women. There was a suggestion of a difference in the results between studies of fruit intake that adjusted or not for oral contraceptive use, p for heterogeneity = 0.07, with no association among the two studies that adjusted for oral contraceptive use, but an inverse association among the studies which did not. For vegetables there was suggestion of heterogeneity between studies that adjusted or not for age at menarche or age at 1st birth, p for heterogeneity = 0.07 for both, with a suggestive inverse association among the studies that made these adjustments, but not for those which did not.
Table 2

Subgroup analyses of fruit and vegetable intakes and breast cancer, high versus low intake

 

Total fruit and vegetables

Fruits

n

RR (95 % CI)

I 2 (%)

P h a

P h b

n

RR (95 % CI)

I 2 (%)

P h a

P h b

All studies

6

0.89 (0.80–0.99)

0

0.71

 

10

0.92 (0.86–0.98)

9.4

0.36

 

Duration of follow-up

 <10 years follow-up

3

0.94 (0.77–1.16)

0

0.43

0.54

6

0.91 (0.83–1.01)

39.9

0.14

0.98

 ≥10 yrs follow-up

3

0.87 (0.77–0.98)

0

0.66

4

0.91 (0.80–1.03)

0

0.67

Menopausal status

 Premenopausal

2

0.82 (0.67–1.02)

0

0.47

0.82/0.43c

2

0.92 (0.71–1.20)

0.3

0.32

0.12/0.79c

 Pre- and post-menopausal

2

0.77 (0.59–0.99)

0

0.77

3

1.00 (0.85–1.17)

27.2

0.25

 Post-menopausal

4

0.93 (0.79–1.08)

20.5

0.29

7

0.89 (0.83–0.95)

0

0.85

Geographic location

 Europe

2

0.91 (0.67–1.23)

56.3

0.13

0.83

2

0.94 (0.67–1.33)

71.9

0.06

0.31

 America

4

0.88 (0.78–1.00)

0

0.89

6

0.89 (0.83–0.95)

0

0.91

 Asia

0

   

2

0.99 (0.80–1.22)

0

0.70

Number of cases

 Cases <500

3

0.94 (0.77–1.16)

0

0.43

0.86

2

0.89 (0.72–1.10)

0

0.50

0.41

 Cases 500–<1500

2

0.84 (0.71–0.99)

0

0.54

5

0.88 (0.79–0.98)

0

0.70

 Cases ≥1500

1

0.91 (0.76–1.09)

  

3

0.95 (0.82–1.09)

66.8

0.05

Adjustment for confounders

 Hormone therapy

  Yes

4

0.90 (0.80–1.00)

0

0.49

0.67

4

0.94 (0.84–1.05)

51.0

0.11

0.35

  No

2

0.83 (0.62–1.12)

0

0.56

 

6

0.87 (0.78–0.97)

0

0.77

 OC use

  Yes

1

0.87 (0.71–1.07)

  

0.82

2

1.01 (0.85–1.20)

50.9

0.15

0.07

  No

5

0.90 (0.79–1.01)

0

0.57

 

8

0.89 (0.83–0.95)

0

0.84

 Age at menarche

  Yes

2

0.89 (0.78–1.02)

0

0.74

0.93

5

0.91 (0.78–1.05)

48.2

0.10

0.75

  No

4

0.88 (0.75–1.04)

0

0.42

 

5

0.90 (0.84–0.97)

0

0.81

 Age at menopause

  Yes

3

0.87 (0.77–0.98)

0

0.66

0.54

2

0.80 (0.66–0.97)

0

0.73

0.51

  No

3

0.94 (0.77–1.16)

0

0.43

 

8

0.93 (0.87–1.00)

8.0

0.37

 Age at 1st birth

  Yes

2

0.89 (0.78–1.02)

0

0.74

0.93

4

0.98 (0.85–1.12)

36.0

0.20

0.21

  No

4

0.88 (0.75–1.04)

0

0.42

 

6

0.89 (0.83–0.95)

0

0.90

 Parity

  Yes

3

0.89 (0.76–1.04)

15.6

0.31

1.00

4

0.98 (0.85–1.12)

36.0

0.20

0.09

  No

3

0.89 (0.76–1.04)

0

0.74

 

6

0.89 (0.83–0.95)

0

0.90

 Education

  Yes

3

0.89 (0.76–1.04)

15.6

0.31

1.00

5

0.96 (0.87–1.05)

30.9

0.22

0.14

  No

3

0.89 (0.76–1.04)

0

0.74

 

5

0.85 (0.76–0.94)

0

0.89

 Alcohol

  Yes

4

0.90 (0.80–1.00)

0

0.49

0.67

5

0.92 (0.83–1.03)

46.7

0.11

0.59

  No

2

0.83 (0.62–1.12)

0

0.56

 

5

0.89 (0.79–0.99)

0

0.76

 Smoking

  Yes

3

0.83 (0.72–0.96)

0

0.82

0.38

4

0.95 (0.85–1.06)

45.5

0.14

0.26

  No

3

0.94 (0.81–1.09)

0

0.46

 

6

0.87 (0.78–0.96)

0

0.75

 Body mass index, weight, WHR

  Yes

4

0.90 (0.80–1.00)

0

0.49

0.67

5

0.95 (0.86–1.05)

38.4

0.17

0.19

  No

2

0.83 (0.62–1.12)

0

0.56

 

5

0.85 (0.76–0.95)

0

0.89

 Physical activity

  Yes

2

0.84 (0.71–0.99)

0

0.54

0.41

3

0.97 (0.85–1.09)

57.3

0.10

0.18

  No

4

0.93 (0.81–1.06)

0

0.63

 

7

0.86 (0.78–0.95)

0

0.84

 Energy intake

  Yes

3

0.87 (0.77–0.98)

0

0.66

0.54

7

0.93 (0.85–1.01)

30.1

0.20

0.48

  No

3

0.94 (0.77–1.16)

0

0.43

 

3

0.87 (0.76–0.99)

0

0.78

 

Vegetables

n

RR (95 % CI)

I 2 (%)

P h a

P h b

All studies

10

0.99 (0.92–1.06

19.6

0.26

 

Duration of follow-up

 <10 years follow-up

6

1.03 (0.97–1.10)

0

0.60

0.27

 ≥10 yrs follow-up

4

0.95 (0.81–1.11)

42.4

0.16

Menopausal status

 Premenopausal

2

0.76 (0.60–0.95)

0.8

0.32

0.14/0.06c

 Pre- and post-menopausal

3

1.03 (0.82–1.29)

47.3

0.15

 Post-menopausal

7

1.03 (0.96–1.09)

0

0.53

Geographic location

 Europe

2

0.97 (0.85–1.12)

0

0.83

0.71

 America

6

0.98 (0.90–1.07)

28.5

0.22

 Asia

2

1.08 (0.67–1.76)

74.1

0.05

Number of cases

 Cases <500

2

1.15 (0.79–1.67)

54.2

0.14

0.71

 Cases 500–<1500

5

0.91 (0.82–1.01)

0

0.90

 Cases ≥1500

3

1.02 (0.94–1.12)

29.8

0.24

Adjustment for confounders

 Hormone therapy

  Yes

4

0.98 (0.89–1.09)

50.0

0.11

0.80

  No

6

0.97 (0.86–1.08)

0

0.46

 OC use

  Yes

2

0.93 (0.83–1.05)

0

0.33

0.36

  No

8

1.01 (0.93–1.09)

15.5

0.31

 Age at menarche

  Yes

5

0.93 (0.85–1.02)

0

0.88

0.07

  No

5

1.04 (0.93–1.15)

26.4

0.25

 Age at menopause

  Yes

4

0.93 (0.85–1.02)

0

0.81

0.10

  No

6

1.02 (0.92–1.13)

24.8

0.25

 Age at 1st birth

  Yes

4

0.90 (0.81–1.01)

0

0.93

0.07

  No

6

1.03 (0.95–1.11)

18.2

0.30

 Parity

  Yes

4

0.92 (0.83–1.03)

0

0.75

0.16

  No

6

1.02 (0.93–1.12)

24.7

0.25

 Education

  Yes

5

0.97 (0.87–1.08)

43.8

0.13

0.90

  No

5

0.98 (0.88–1.09)

0

0.44

 Alcohol

  Yes

5

0.99 (0.90–1.08)

35.5

0.19

0.89

  No

5

0.97 (0.85–1.11)

13.7

0.33

 Smoking

  Yes

4

1.00 (0.90–1.10)

41.6

0.16

0.58

  No

6

0.96 (0.87–1.06)

0

0.44

 Body mass index, weight, WHR

  Yes

5

0.97 (0.89–1.07)

43.6

0.13

0.88

  No

5

0.99 (0.87–1.11)

0

0.41

 Physical activity

  Yes

3

0.99 (0.88–1.12)

59.8

0.08

0.53

  No

7

0.96 (0.87–1.05)

0

0.57

 Energy intake

  Yes

7

0.97 (0.90–1.05)

25.0

0.24

0.52

  No

3

1.05 (0.85–1.28)

36.5

0.21

aThe p for heterogeneity within each subgroup

bThe p for heterogeneity between subgroups with meta-regression analysis

cThe p for heterogeneity between premenopausal and postmenopausal women (excluding studies with mixed menopausal status)

When we further stratified the studies by the median range of intake, the summary RR was 0.88 (95 % CI: 0.73–1.07) and 0.89 (95 % CI: 0.78–1.02) for studies with a range of fruit and vegetable intake of ≥441 and <441 g/day, respectively. The summary RR was 0.87 (95 % CI: 0.77–0.99) and 0.93 (95 % CI: 0.85–1.02) for studies with a range of fruit intake of ≥275 and <275 g/day, respectively, and 1.03 (95 % CI: 0.96–1.10) and 0.92 (95 % CI: 0.82–1.02) for studies with a range of vegetable intake of ≥273 and <273 g/day, respectively (results not shown).

We also assessed the influence of including studies on breast cancer mortality with our results. Two additional studies were included in the high versus low analysis of fruit [30, 31] and one of these in the dose–response [31]. The summary RR for high versus low intake was 0.92 (95 % CI: 0.86–0.97) with low heterogeneity, I 2 = 33 %, and p heterogeneity = 0.42, and per 200 g/day was 0.94 (95 % CI: 0.89–0.99) with low heterogeneity, I 2 = 33 %, and p heterogeneity = 0.14 similar to the original analysis.

Discussion

In this meta-analysis, high versus low intake of fruits and fruits and vegetables combined, but not vegetables, were associated with small, but statistically significant reductions in breast cancer risk. In the dose–response analyses, fruits and fruits and vegetables combined, but not vegetables, were associated with reduced risk, although only marginally significantly so.

Our results are similar to those of a pooled analysis of eight prospective studies which found a non-significant reduction of ~7 % for high versus low intake of fruits and fruits and vegetables combined, but no association with intake of vegetables [46]. In the 2nd report from the WCRF/AICR, it was stated that the evidence for an association between intake of fruits and non-starchy vegetables and breast cancer risk was too limited or inconsistent for a conclusion, thus a downgrading of the judgement of the evidence for fruit since the 1st report [6]. However, with additional large prospective studies published after the report we found significant inverse associations between high versus low intake of fruits and fruits and vegetables combined and breast cancer risk. To our knowledge this is the first meta-analysis to have assessed a possible nonlinear association between fruit and vegetable intake and breast cancer risk, but the inverse association with fruit and fruit and vegetable intake combined appeared to be linear. This meta-analysis included a larger number of studies than previous meta-analyses and had more than twice as many cases and participants as the pooled analysis, thus we had statistical power to detect moderate associations, although the associations for fruits and vegetables and fruits were still only marginally significant in the dose–response analysis. This may partly be due to the range being larger in the high versus low analysis than in the linear dose–response analysis. In addition, gains in statistical power by increasing sample size are less when effect estimates are small and the sample size already is large.

Our meta-analysis may have several limitations that need to be discussed. We cannot exclude the possibility that the observed inverse association between fruit and vegetable intake and breast cancer risk could be due to unmeasured or residual confounding. Higher intake of fruits and vegetables is often associated with other lifestyle factors including higher levels of physical activity, lower prevalence of overweight/obesity, and lower intakes of alcohol and dietary fat. Many, but not all of the studies adjusted for these and other potential confounders. In subgroup and meta-regression analyses, there was a suggestion of a difference in the results between studies of fruit intake that adjusted or not for oral contraceptive use, p for heterogeneity = 0.07, and for vegetables among between studies that adjusted or not for age at menarche or age at 1st birth, p for heterogeneity = 0.07 for both. However, the few studies in some of these subgroups make the interpretation of these findings difficult. Because of the few studies published we were not able to examine the association between specific types of fruits and vegetables and breast cancer risk.

Measurement errors in the assessment of the exposure variable are known to bias effect estimates, however, bias toward the null is most likely because we included only prospective studies. Almost all the studies included in our meta-analysis used validated food frequency questionnaires, but only one of the studies corrected the risk estimates for measurement error. However, the results did not differ substantially before and after calibration [17]. Dietary changes during follow-up can obscure associations between dietary intake and disease risk if dietary intake only is assessed at baseline. One study reported a RR of 0.59 (95 % CI: 0.40–0.87) for high versus low intake of fruits, berries and vegetables among women without a dietary change in the past, while there was no association among persons who reported that they had changed their dietary intake, RR = 1.26 (95 % CI: 0.63–2.55) [37]. If the relevant exposure window is in the distant past or in adolescence it is possible that most studies may have missed an effect, because most of the studies published to date have been conducted primarily among middle-aged and older persons. In addition, measurement errors due to different dietary questionnaires or nutrient databases may have affected the results. Because some studies reported intakes in frequency, we had to convert the intakes to g/day based on a standard serving size (80 g). It is possible that this may have introduced some measurement error because different types of fruits and vegetables may have different serving sizes. Any further studies should report results in g/day to provide data that can more easily be incorporated in future meta-analyses. Considering the weak associations we observed, future studies might want to clarify whether improved exposure assessment using biomarkers of fruit and vegetable intake or by correcting for measurement error might lead to more conclusive results.

Although small study bias, such as publication bias can be a problem in meta-analyses of published studies, we found no statistical evidence of publication bias in this meta-analysis and there was also no asymmetry in the funnel plots when inspected visually.

Several potential mechanisms may explain an inverse association between fruits and vegetables and breast cancer risk. Fruits and vegetables are good sources of fiber which may prevent breast cancer by binding estrogens during the enterohepatic reabsorption of estrogens in the colon [47]. In addition, fruits and vegetables are good sources of various antioxidants, such as carotenoids [48, 49, 50], glucosinolates, indoles, and isothiocyanates [51], which may prevent breast cancer by inducing the activity of detoxifying enzymes, and reducing oxidative stress and inflammation. High intake of fruits and vegetables may also decrease the risk of overweight/obesity [52] which is an established risk factor for postmenopausal breast cancer.

Strengths of our meta-analysis include the prospective design of the included studies which minimize the possibility for recall and selection bias, and the large number of cases and participants (up to 780,000 participants and >16,000 cases), which provides statistical power to detect moderate associations. Our results for fruit and vegetable intake and breast cancer risk are relatively weak, but of similar size as our previously reported associations with colorectal cancer risk [53]. However, if consistent across cancer sites such a reduction in cancer risk could still have a moderate, but nevertheless important impact on overall cancer risk.

In conclusion, we found weak and linear inverse associations between intake of fruits and fruits and vegetables combined, but not vegetables, and breast cancer risk. Further studies of specific types of fruits and vegetables, with improved exposure assessment methods, adjustment for more confounding factors, and stratified by menopausal status and hormone receptor status are warranted.

Notes

Acknowledgments

We thank the systematic literature review team at the Istituto Nazionale Tumori Milan for their contributions to the breast cancer database. This work was funded by the World Cancer Research Fund (Grant No. 2007/SP01) as part of the Continuous Update Project. The views expressed in this review are the opinions of the authors. They may not represent the views of WCRF International/AICR and may differ from those in future updates of the evidence related to food, nutrition, physical activity, and cancer risk. All authors had full access to all of the data in the study. D. Aune takes responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest

The authors declare that there are no conflicts of interest.

Funding

The sponsor of this study had no role in the decisions about the design and conduct of the study, collection, management, analysis or interpretation of the data or the preparation, review or approval of the manuscript.

References

  1. 1.
    Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM (2010) Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 127:2893–2917PubMedCrossRefGoogle Scholar
  2. 2.
    Chia KS, Reilly M, Tan CS et al (2005) Profound changes in breast cancer incidence may reflect changes into a Westernized lifestyle: a comparative population-based study in Singapore and Sweden. Int J Cancer 113:302–306PubMedCrossRefGoogle Scholar
  3. 3.
    Leung GM, Thach TQ, Lam TH et al (2002) Trends in breast cancer incidence in Hong Kong between 1973 and 1999: an age–period–cohort analysis. Br J Cancer 87:982–988PubMedCrossRefGoogle Scholar
  4. 4.
    Kolonel LN (1980) Cancer patterns of four ethnic groups in Hawaii. J Natl Cancer Inst 65:1127–1139PubMedGoogle Scholar
  5. 5.
    Ziegler RG, Hoover RN, Pike MC et al (1993) Migration patterns and breast cancer risk in Asian-American women. J Natl Cancer Inst 85:1819–1827PubMedCrossRefGoogle Scholar
  6. 6.
    World Cancer Research Fund/American Institute for Cancer Research Food, Nutrition, Physical Activity and the Prevention of Cancer: a Global Perspective (2007). AICR, WashingtonGoogle Scholar
  7. 7.
    Maskarinec G, Morimoto Y, Takata Y, Murphy SP, Stanczyk FZ (2006) Alcohol and dietary fibre intakes affect circulating sex hormones among premenopausal women. Public Health Nutr 9:875–881PubMedCrossRefGoogle Scholar
  8. 8.
    Steinmetz KA, Potter JD (1991) Vegetables, fruit, and cancer. II. Mechanisms. Cancer Causes Control 2:427–442PubMedCrossRefGoogle Scholar
  9. 9.
    Vainio H, Weiderpass E (2006) Fruit and vegetables in cancer prevention. Nutr Cancer 54:111–142PubMedCrossRefGoogle Scholar
  10. 10.
    Shibata A, Paganini-Hill A, Ross RK, Henderson BE (1992) Intake of vegetables, fruits, beta-carotene, vitamin C and vitamin supplements and cancer incidence among the elderly: a prospective study. Br J Cancer 66:673–679PubMedCrossRefGoogle Scholar
  11. 11.
    Rohan TE, Howe GR, Friedenreich CM, Jain M, Miller AB (1993) Dietary fiber, vitamins A, C, and E, and risk of breast cancer: a cohort study. Cancer Causes Control 4:29–37PubMedCrossRefGoogle Scholar
  12. 12.
    Byrne C, Ursin G, Ziegler RG (1996) A comparison of food habit and food frequency data as predictors of breast cancer in the NHANES I/NHEFS cohort. J Nutr 126:2757–2764PubMedGoogle Scholar
  13. 13.
    Verhoeven DT, Assen N, Goldbohm RA et al (1997) Vitamins C and E, retinol, beta-carotene and dietary fibre in relation to breast cancer risk: a prospective cohort study. Br J Cancer 75:149–155PubMedCrossRefGoogle Scholar
  14. 14.
    Zhang S, Hunter DJ, Forman MR et al (1999) Dietary carotenoids and vitamins A, C, and E and risk of breast cancer. J Natl Cancer Inst 91:547–556PubMedCrossRefGoogle Scholar
  15. 15.
    Key TJ, Sharp GB, Appleby PN et al (1999) Soya foods and breast cancer risk: a prospective study in Hiroshima and Nagasaki, Japan. Br J Cancer 81:1248–1256PubMedCrossRefGoogle Scholar
  16. 16.
    Olsen A, Tjonneland A, Thomsen BL et al (2003) Fruits and vegetables intake differentially affects estrogen receptor negative and positive breast cancer incidence rates. J Nutr 133:2342–2347PubMedGoogle Scholar
  17. 17.
    van Gils CH, Peeters PH, Bueno-de-Mesquita HB et al (2005) Consumption of vegetables and fruits and risk of breast cancer. JAMA 293:183–193PubMedCrossRefGoogle Scholar
  18. 18.
    Sonestedt E, Borgquist S, Ericson U et al (2008) Plant foods and oestrogen receptor alpha- and beta-defined breast cancer: observations from the Malmo Diet and Cancer cohort. Carcinogenesis 29:2203–2209PubMedCrossRefGoogle Scholar
  19. 19.
    Jayalekshmi P, Varughese SC, Kalavathi et al (2009) A nested case-control study of female breast cancer in Karunagappally cohort in Kerala, India. Asian Pac J Cancer Prev 10:241–246PubMedGoogle Scholar
  20. 20.
    Butler LM, Wu AH, Wang R, Koh WP, Yuan JM, Yu MC (2010) A vegetable-fruit-soy dietary pattern protects against breast cancer among postmenopausal Singapore Chinese women. Am J Clin Nutr 91:1013–1019PubMedCrossRefGoogle Scholar
  21. 21.
    Boggs DA, Palmer JR, Wise LA et al (2010) Fruit and vegetable intake in relation to risk of breast cancer in the Black Women’s Health Study. Am J Epidemiol 172:1268–1279PubMedCrossRefGoogle Scholar
  22. 22.
    Brasky TM, Lampe JW, Potter JD, Patterson RE, White E (2010) Specialty supplements and breast cancer risk in the VITamins And Lifestyle (VITAL) Cohort. Cancer Epidemiol Biomarkers Prev 19:1696–1708PubMedCrossRefGoogle Scholar
  23. 23.
    Lof M, Sandin S, Lagiou P, Trichopoulos D, Adami HO, Weiderpass E (2011) Fruit and vegetable intake and risk of cancer in the Swedish women’s lifestyle and health cohort. Cancer Causes Control 22:283–289PubMedCrossRefGoogle Scholar
  24. 24.
    George SM, Park Y, Leitzmann MF et al (2009) Fruit and vegetable intake and risk of cancer: a prospective cohort study. Am J Clin Nutr 89:347–353PubMedCrossRefGoogle Scholar
  25. 25.
    Gandini S, Merzenich H, Robertson C, Boyle P (2000) Meta-analysis of studies on breast cancer risk and diet: the role of fruit and vegetable consumption and the intake of associated micronutrients. Eur J Cancer 36:636–646PubMedCrossRefGoogle Scholar
  26. 26.
    Riboli E, Norat T (2003) Epidemiologic evidence of the protective effect of fruit and vegetables on cancer risk. Am J Clin Nutr 78:559S–569SPubMedGoogle Scholar
  27. 27.
    Michels KB, Mohllajee AP, Roset-Bahmanyar E, Beehler GP, Moysich KB (2007) Diet and breast cancer: a review of the prospective observational studies. Cancer 109:2712–2749PubMedCrossRefGoogle Scholar
  28. 28.
    Moher D, Liberati A, Tetzlaff J, Altman DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 339:b2535PubMedCrossRefGoogle Scholar
  29. 29.
    Key TJ, Thorogood M, Appleby PN, Burr ML (1996) Dietary habits and mortality in 11,000 vegetarians and health conscious people: results of a 17 year follow up. BMJ 313:775–779PubMedCrossRefGoogle Scholar
  30. 30.
    Appleby PN, Key TJ, Burr ML, Thorogood M (2002) Mortality and fresh fruit consumption. IARC Sci Publ 156:131–133PubMedGoogle Scholar
  31. 31.
    Sauvaget C, Nagano J, Hayashi M, Spencer E, Shimizu Y, Allen N (2003) Vegetables and fruit intake and cancer mortality in the Hiroshima/Nagasaki Life Span Study. Br J Cancer 88:689–694PubMedCrossRefGoogle Scholar
  32. 32.
    Maynard M, Gunnell D, Emmett P, Frankel S, Davey SG (2003) Fruit, vegetables, and antioxidants in childhood and risk of adult cancer: the Boyd Orr cohort. J Epidemiol Community Health 57:218–225PubMedCrossRefGoogle Scholar
  33. 33.
    Mattisson I, Wirfalt E, Johansson U, Gullberg B, Olsson H, Berglund G (2004) Intakes of plant foods, fibre and fat and risk of breast cancer–a prospective study in the Malmo Diet and Cancer cohort. Br J Cancer 90:122–127PubMedCrossRefGoogle Scholar
  34. 34.
    Mattisson I, Wirfalt E, Wallstrom P, Gullberg B, Olsson H, Berglund G (2004) High fat and alcohol intakes are risk factors of postmenopausal breast cancer: a prospective study from the Malmo diet and cancer cohort. Int J Cancer 110:589–597PubMedCrossRefGoogle Scholar
  35. 35.
    Fung TT, Hu FB, Holmes MD et al (2005) Dietary patterns and the risk of postmenopausal breast cancer. Int J Cancer 116:116–121PubMedCrossRefGoogle Scholar
  36. 36.
    Ravn-Haren G, Olsen A, Tjonneland A et al (2006) Associations between GPX1 Pro198Leu polymorphism, erythrocyte GPX activity, alcohol consumption and breast cancer risk in a prospective cohort study. Carcinogenesis 27:820–825PubMedCrossRefGoogle Scholar
  37. 37.
    Sonestedt E, Gullberg B, Wirfalt E (2007) Both food habit change in the past and obesity status may influence the association between dietary factors and postmenopausal breast cancer. Public Health Nutr 10:769–779PubMedCrossRefGoogle Scholar
  38. 38.
    Trichopoulou A, Bamia C, Lagiou P, Trichopoulos D (2010) Conformity to traditional Mediterranean diet and breast cancer risk in the Greek EPIC (European Prospective Investigation into Cancer and Nutrition) cohort. Am J Clin Nutr 92:620–625PubMedCrossRefGoogle Scholar
  39. 39.
    Frazier AL, Li L, Cho E, Willett WC, Colditz GA (2004) Adolescent diet and risk of breast cancer. Cancer Causes Control 15:73–82PubMedCrossRefGoogle Scholar
  40. 40.
    DerSimonian R, Laird N (1986) Meta-analysis in clinical trials. Control Clin Trials 7:177–188PubMedCrossRefGoogle Scholar
  41. 41.
    Greenland S, Longnecker MP (1992) Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 135:1301–1309PubMedGoogle Scholar
  42. 42.
    Bandera EV, Kushi LH, Moore DF, Gifkins DM, McCullough ML (2007) Fruits and vegetables and endometrial cancer risk: a systematic literature review and meta-analysis. Nutr Cancer 58:6–21PubMedCrossRefGoogle Scholar
  43. 43.
    Royston P (2000) A strategy for modelling the effect of a continuous covariate in medicine and epidemiology. Stat Med 19:1831–1847PubMedCrossRefGoogle Scholar
  44. 44.
    Higgins JP, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis. Stat Med 21:1539–1558PubMedCrossRefGoogle Scholar
  45. 45.
    Egger M, Davey Smith G, Schneider M, Minder C (1997) Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629–634Google Scholar
  46. 46.
    Smith-Warner SA, Spiegelman D, Yaun SS et al (2001) Intake of fruits and vegetables and risk of breast cancer: a pooled analysis of cohort studies. JAMA 285:769–776PubMedCrossRefGoogle Scholar
  47. 47.
    Goldin BR, Woods MN, Spiegelman DL et al (1994) The effect of dietary fat and fiber on serum estrogen concentrations in premenopausal women under controlled dietary conditions. Cancer 74:1125–1131PubMedCrossRefGoogle Scholar
  48. 48.
    Toniolo P, Van Kappel AL, Akhmedkhanov A et al (2001) Serum carotenoids and breast cancer. Am J Epidemiol 153:1142–1147PubMedCrossRefGoogle Scholar
  49. 49.
    Sato R, Helzlsouer KJ, Alberg AJ, Hoffman SC, Norkus EP, Comstock GW (2002) Prospective study of carotenoids, tocopherols, and retinoid concentrations and the risk of breast cancer. Cancer Epidemiol Biomarkers Prev 11:451–457PubMedGoogle Scholar
  50. 50.
    Tamimi RM, Hankinson SE, Campos H et al (2005) Plasma carotenoids, retinol, and tocopherols and risk of breast cancer. Am J Epidemiol 161:153–160PubMedCrossRefGoogle Scholar
  51. 51.
    Fowke JH, Chung FL, Jin F et al (2003) Urinary isothiocyanate levels, brassica, and human breast cancer. Cancer Res 63:3980–3986PubMedGoogle Scholar
  52. 52.
    He K, Hu FB, Colditz GA, Manson JE, Willett WC, Liu S (2004) Changes in intake of fruits and vegetables in relation to risk of obesity and weight gain among middle-aged women. Int J Obes Relat Metab Disord 28:1569–1574PubMedCrossRefGoogle Scholar
  53. 53.
    Aune D, Lau R, Chan DS et al (2011) Nonlinear reduction in risk for colorectal cancer by fruit and vegetable intake based on meta-analysis of prospective studies. Gastroenterology 141:106–118PubMedCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC. 2012

Authors and Affiliations

  • D. Aune
    • 1
  • D. S. M. Chan
    • 1
  • A. R. Vieira
    • 1
  • D. A. Navarro Rosenblatt
    • 1
  • R. Vieira
    • 1
  • D. C. Greenwood
    • 2
  • T. Norat
    • 1
  1. 1.Department of Epidemiology and Biostatistics, School of Public HealthImperial College LondonLondonUK
  2. 2.Biostatistics Unit, Centre for Epidemiology and BiostatisticsUniversity of LeedsLeedsUK

Personalised recommendations