Vegetables, Fruit, and breast Cancer.

Abstract

OBJECTIVES:

  • To review all prospective studies which provided data about total fruits or vegetables, their subgroups, and specific types of vegetables and fruits, in relationship with breast cancer risk, disease progression and mortality/survival.
  • To define the amount of consumption found to be related with the described effects on cancer.
  • To define possible effect modification by confounders.

DATA SOURCE: The Pubmed database was searched (No start date - Sep 11, 2009) for relevant articles using the keywords "cancer, neoplasm, or neoplasms" combined with "vegetables, vegetable, fruit, or fruits" and a fair amount of other keywords. The exact search term is described in the methods.
Prospective studies published in the English language were included. Reference lists were searched for additional articles.
RESULTS: 64 articles were found which provided information about 41 different cohorts. Of these, 0 articles were excluded.
Results are described when any evidence for an association - as defined in the Methods - was found. In addition, data about total vegetables or fruits is described.

Total disease risk.

  • Total vegetables: Data was provided by a pooled analysis of 8 cohorts, and 8 additional cohorts, including a total of 18,023 cases. No evidence was found for an association between total vegetables and breast cancer risk (RR = 1.00).
  • Total fruit: Data was provided by a pooled analysis of 8 cohorts, and 9 additional cohorts, including a total of 18,425 cases. Suggestive evidence was found for a weak protective effect of total fruit against breast cancer risk (RR = 0.96). This effect may be restricted to never users of oral contraceptives. No level of consumption could be defined for this effect.
  • Rosaceae: Data was provided by a pooled analysis of 7 cohorts, including 7,217 cases. Items from the rosaceae family possibly protect against breast cancer risk (RR = 0.97).

Advanced stage/metastatic disease risk or disease progression.

  • Total vegetables: Data about disease progression was provided by 2 cohorts, including a total of 626 cases. No evidence was found for an association between total vegetables and breast cancer recurrence (RR = 0.88).
  • Total fruit: Data about disease progression was provided by 2 cohorts, including a total of 545 cases. No evidence was found for an association between total fruit and breast cancer recurrence (RR = 0.82).

Mortality risk.

  • Total vegetables: Data was provided by 9 cohorts, including a total of 1,144 cases. Protective effects against mortality were found in 4 cohorts, but these were mostly nonsignificant. Suggestive evidence was found for a protective effect of total vegetables against breast cancer mortality (RR = 0.90). No level of consumption could be defined for this effect.
  • Total fruit: Data was provided by 10 cohorts, including a total of 1,264 cases. No evidence was found for an association between total fruit and breast cancer mortality (RR = 0.90).

CONCLUSION: Suggestive evidence was found for a weak protective effect of total fruit against breast cancer risk (- 4%), which may be restricted to never users of oral contraceptives. And suggestive evidence was found for a protective effect of total vegetables against breast cancer mortality (- 10%). No evidence was found for other associations with total vegetables or fruits, though the effect size for the relation between total fruits and breast cancer mortality was identical to the effect size for total vegetables.
Items from the rosaceae family possibly protect against breast cancer risk, but the effect is weak (- 3%), and none of the specific items in this family was associated with breast cancer. No other associations were found, and no level of consumption could be defined for any of the associations.


Total vegetables and breast cancer.

Total breast cancer risk: Data about total vegetables in relation to total breast cancer risk was provided by a pooled analysis of 8 cohorts, and 8 additional cohorts, including a total of 18,023 cases.
A weak significantly increased risk was found in one cohort of very large size, including 5,815 cases (George SM [38]). No other (non)significant associations were found. The average RR = 1.00 (excluding incomplete data from Hirvonen T [29], Iwasaki M [20], Jarvinen R [11]).

Inclusion of intermediate levels of consumption:
(Non)significant effects at any level of consumption were as follows:

  • Zhang S (4) Significant protective at 154-223, and ≥ 385 g/day among premenopausal women.
  • van Gils CH (21) Significant protective at > 109-≤ 160 g/day.
  • Sonestedt E (24) Significant protective at 218 g/day.
  • George SM (38) Though the trend increased significantly, the risk increased nonsignificantly at ≥ 440 g/day.

RRs for the association between total vegetables and breast cancer risk (g/day):


Disease progression: Data about total vegetables in relation to disease progression was provided by 2 cohorts, including a total of 626 cases.
A nonsignificant protective effect against breast cancer recurrence among postmenopausal women was found in the smallest cohort (Hebert JR [12]), but no association was found in the larger cohort (Pierce JP [35]). The average RR = 0.88

Breast cancer mortality: Results were included about a) breast cancer mortality, and b) all-cause mortality following diagnosis of breast cancer. If data about both total mortality, and breast cancer-specific mortality were available following diagnosis of breast cancer, total mortality was chosen as end point.
Data about total vegetables in relation to breast cancer death was provided by 9 cohorts, including a total of 1,144 + X cases (no amount of cases was defined for one cohort (Ewertz M [3]).
A significant protective effect was found in one cohort (McEligot AJ [31]), and nonsignificant effects were found in 3 other cohorts (Ewertz M [3], Holmes D [4], Hebert JR [12]), including a total of 547 + X cases for all 4 cohorts combined (> 48% of all cases). No other (non)significant associations were found. The average RR = 0.90 (excluding incomplete data from Goodwin PC [23], Ingram D [6], Ewertz M [3]).

Inclusion of intermediate levels of consumption:
Significant protective effects at any level of consumption were restricted to one cohort, including 96 cases at the level of consumption of ≥ 154 g/day (31; McEligot AJ). No other (non)significant associations were found.



RRs for the association between total vegetables and breast cancer mortality (g/day):


Conclusion: A significantly increased risk was found in one cohort, but no associations with breast cancer risk were found in the remaining cohorts. Furthermore, no significant associations were found with breast cancer recurrence. No evidence was found for an association between total vegetables and either breast cancer risk or breast cancer recurrence.
Protective associations against mortality were found in 4 cohorts, but these were mostly nonsignificant. The cohorts included 48% of all cases. Suggestive evidence was found for a protective effect of total vegetables against breast cancer mortality (- 10%). No level of consumption could be defined for this effect.

Prospective studies of total vegetables and breast cancer risk:
AuthorCohort nameCasesRelative Risk (RR)
38) George SM (2008)The NIH-AARP Diet and Health Study5,815RR = 1.08 (1.00-1.18; P = 0.009)
29) Hirvonen T (2006)The SU.VI.MAX Study95No significant association (P = 0.98)
24) Sonestedt E (2008)The Malmö Diet and Cancer Cohort544HR = 0.84 (0.64-1.11; P = 0.18)
22) Maynard M (2003)The Boyd Orr cohort82OR = 1.43 (0.70-2.92; P = 0.59)
21) van Gils CH (2005)The EPIC Study3,659RR = 0.98 (0.84-1.14)
20) Iwasaki M (2008)The JPHC Study144No significant association (P = 0.20)
15) Smith-Warner SA (2001)Pooled analysis of 8 cohort studies7,377RR = 0.96 (0.89-1.04; P = 0.54)
11) Jarvinen R (1997)The Finnish Mobile Clinic Health Study88No significant association (no data shown)
1) Shibata A (1992)The Leisure World Study219RR = 0.96 (0.69-1.34)
Total number of cases: 18,023Average RR = 1.00


Prospective studies of total vegetables and breast cancer disease progression:
AuthorCohort nameCasesEnd pointRelative Risk (RR)Dietary assessment following cancer diagnosis
35) Pierce JP (2007)The WHEL Trial517Breast cancer eventsHR = 0.97 (0.67-1.40)yes
12) Hebert JR (1998)No cohort name defined109Breast cancer recurrenceRR = 0.46 (P = 0.08) among postmenopausal womenYes
Total number of cases: 626Average RR = 0.88


Prospective studies of total vegetables and total/breast cancer mortality:
AuthorCohort nameCasesRelative Risk (RR)Dietary assessment following cancer diagnosis
35) Pierce JP (2007)The WHEL Trial314HR = 1.19 (0.74-1.90)Yes
32) Fink BN (2006)The Long Island Breast Cancer Study ProjectPremenopausal: 43.

Postmenopausal: 131.
Premenopausal: HR = 1.40 (0.71-2.76; P = 0.53).

Postmenopausal: HR = 0.92 (0.57-1.48; P = 0.52)
Yes
31) McEligot AJ (2006)No cohort name defined96HR = 0.57 (0.35-0.94; P = 0.02)Yes
23) Goodwin PJ (2003)No cohort name defined52No significant effect (no data shown)Yes
22) Maynard M (2003)The Boyd Orr Cohort36OR = 0.86 (0.30-2.47; P = 0.35)No
12) Hebert JR (1998)No cohort name defined73RR = 0.31 (P = 0.08) among postmenopausal womenYes
6) Ingram D (1994)No cohort name defined21No significant associaton (P = 0.94)Yes
4) Holmes D (1999)The Nurses' Health Study378RR = 0.81 (0.59-1.11; P = 0.07)Yes
3) Ewertz M (1991)No cohort name definedNot definedA nonsignificant slightly decreased risk (no data shown)Yes
Total number of cases: 1,144 + XAverage RR = 0.90


Click here for an extended version of this table.

Total fruit and breast cancer.

Total breast cancer risk: Data about total fruit in relation to total breast cancer risk was provided by a pooled analysis of 8 cohorts, and 9 additional cohorts, including a total of 18,425 cases.
A weak significant protective effect was found in one cohort of very large size, including 5,815 cases (38). And a weak nonsignificant protective effect was found in the pooled analysis, including 7,377 cases (15). No other (non)significant associations were found. The average RR = 0.96 (excluding incomplete data from Hirvonen T [30], Iwasaki M [20], Jarvinen R [11]).
No levels of consumption were defined in the pooled analysis, but effects were restricted to above median consumption (quartiles 3-4).

Inclusion of intermediate levels of consumption:
(Non)significant effects at any level of consumption were as follows:

  • Shibata A (1) Significant protective at 192-< 296 g/day.
  • George SM (38) Though the trend was significant, the effect was nonsignificantly protective at > 608 g/day.

RRs for the association between total fruit and breast cancer risk (g/day):


Effect modification: Effect modification was found in the pooled analysis of 8 cohorts (Smith-Warner SA). Protective effects against cancer risk were restricted to never users of oral contraceptives.

Disease progression: Data about total fruit in relation to disease progression was provided by 2 cohorts, including a total of 545 cases.
No (non)significant associations were found with breast cancer recurrence. The average RR = 0.82.

Breast cancer mortality: Results were included about a) breast cancer mortality, and b) mortality following diagnosis of breast cancer. If data about both all-cause mortality, and breast cancer-specific mortality were available following diagnosis of breast cancer, total mortality was chosen as end point.
Data about total fruit in relation to breast cancer death was provided by 10 cohorts, including a total of 1,264 cases.
A significant protective effect was found in one cohort (6), and a nonsignificant protective effect was found in one other cohort (31), including a total of 117 cases for both cohorts combined. No other (non)significant associations were found. The average RR = 0.90 (excluding incomplete data from Goodwin PJ [23], Ingram D [6]).

Inclusion of intermediate levels of consumption:
Significant effects at any level of consumption were restricted to one cohort. A significantly increased risk was found at the level of consumption of 235-350 g/d (Pierce JP). Though trends were protective in 2 cohorts, no (non)significant protective effects were found in any cohort.



RRs for the association between total fruit and breast cancer mortality (g/day):


Conclusion: A weak significant protective effect of total fruit against breast cancer risk was found in one cohort of very large size, and a weak nonsignificant protective effect was found in a pooled analysis of 8 cohorts. The combined amount of cases from these cohorts included 72% of all cases. Suggestive evidence was found for a weak protective effect of total fruit against breast cancer risk (- 4%). This effect may be restricted to never users of oral contraceptives. No level of consumption could be defined for this effect, but above median consumption seems required.
Protective effects against mortality were found in 2 cohorts, and these were significant once. These cohorts included 9% of all cases. No associations were found with breast cancer recurrence. No evidence was found for an association between total fruit and breast cancer recurrence or breast cancer mortality.

Prospective studies of total fruit and breast cancer risk:
AuthorCohort nameCasesRelative Risk (RR)
38) George SM (2008)The NIH-AARP Diet and Health Study5,815RR = 0.91 (0.84-1.00; P = 0.01)
30) Hirvonen T (2006)The SU.VI.MAX Study95No significant association (P = 0.73)
24) Sonestedt E (2008)The Malmö Diet and Cancer Cohort544HR = 0.93 (0.70-1.23; P = 0.58)
22) Maynard M (2003)The Boyd Orr Cohort82OR = 1.08 (0.52-2.25; P = 0.61)
21) van Gils CH (2005)The EPIC Study3,659RR = 1.09 (0.94-1.25; P = 0.11)
20) Iwasaki M (2008)The JPHC Study144No significant association (P = 0.79)
15) Smith-Warner SA (2001)Pooled analysis of 8 cohort studies7,377RR = 0.93 (0.86-1.00; P = 0.08)
13) Key TJ (1999)The Hiroshima/Nagasaki Life Span Study402RR = 0.95 (0.71-1.27; P = 0.53)
11) Jarvinen R (1997)The Finnish Mobile Clinic Health Cohort88No significant association (no data shown)
1) Shibata A (1992)The Leisure World Study219RR = 0.82 (0.60-1.12)
Total number of cases: 18,425Average RR = 0.96


Prospective studies of total fruit and breast cancer disease progression:
AuthorCohort nameCasesEnd pointRelative Risk (RR)Dietary assessment following cancer diagnosis
35) Pierce JP (2007)The WHEL Trial517Breast cancer eventsHR = 0.81 (0.57-1.16)Yes
14) Saxe GA (1999)No cohort name defined28Breast cancer recurrenceHR = 0.96 (0.63-1.45)Yes
Total number of cases: 545Average RR = 0.82


Prospective studies of total fruit and total/breast mortality:
AuthorCohort nameCasesRelative Risk (RR)Dietary assessment following cancer diagnosis
35) Pierce JP (2007)The WHEL Trial314HR = 0.76 (0.48-1.19)Yes
32) Fink BN (2006)The Long Cancer Island Breast Cancer Study ProjectPremenopausal: 43.

Posmenopausal: 132.
Premenopausal: HR = 1.10 (0.48-2.52; P = 0.28).

Postmenopausal: HR = 0.87 (0.57-1.35; P = 0.34)
Yes
31) McEligot AJ (2006)No cohort name defined96HR = 0.63 (0.38-1.05; P = 0.08)Yes
23) Goodwin PJ (2003)No cohort name defined52No significant association (no data shown)Yes
22) Maynard M (2003)The Boyd Orr Cohort36OR = 1.25 (0.40-3.92; P = 0.73)No
14) Saxe GA (1999)No cohort name defined26HR = 1.06 (0.69-1.63)Yes
13) Sauvaget C (2003)The Hiroshima/Nagasaki Life Span Study76RR = 0.91 (0.48-1.72; P = 0.70)No
8) Appleby PN (2002)The Health Food Shoppers Study90RR = 0.66 (0.38-1.14)No
6) Ingram D (1994)No cohort name defined21A significant protective effect (P = 0.01)Yes
4) Holmes MD (1999)The Nurses' Health Study378RR = 1.07 (0.77-1.49; P = 0.40)Yes
Total number of cases: 1,264Average RR = 0.90


Click here for an extended version of this table.