Skip navigation

Update on fruit and vegetable intake and breast cancer risk

 

Research has suggested that increased fruit and vegetable consumption is associated with a reduced risk of breast cancer. However, the majority of this research is retrospective; based on case-control studies, which are susceptible to recall and selection bias. This paper examines the association between fruit and vegetable consumption and breast cancer by pooling results from prospective cohort studies.

Bottom line

This paper concludes that fruit and vegetable consumption during adulthood is not associated with breast cancer risk.

Reference

SA Smith-Warner et al. Intake of fruits and vegetables and risk of breast cancer. Journal of the American Medical Association 2001 285: 769-776.

Study selection

Studies were from the Pooling Project of Prospective Studies of Diet and Cancer. To be included in this analysis, studies had to have (i) at least 200 cases of breast cancer; (ii) assessed usual dietary intake; and (iii) validated the questionnaire used to assess diet (or a closely related instrument). From the studies that met these criteria, participants who reported a history of cancer at baseline were excluded.

Eight studies were identified with a total of 351,825 participants and 7,377 cases of breast cancer. Participant numbers ranged from 14,006 to 157,863 and cases of breast cancer ranged from 160 to 2,661. Participants were between 28 and 90 years old and were followed for between five and ten years. Diet was assessed using food frequency questionnaires with the number of fruit and vegetable questions ranging from nine to 54. Participants’ dietary intake was divided into quartiles.

The following were examined: fruit (without juice); fruit juice; total fruit (fruit and fruit juice); total vegetables (vegetables and vegetable juice); total fruit and vegetables; 17 individual fruit and vegetables for which intake had been assessed in at least five studies; and eight fruit and vegetable groups based on botanical taxonomy (e.g. Cruciferae [e.g. broccoli and cabbage]). The effects of these groups were analysed according to an intake increment of 100 g a day and (with the exception of botanical groups and individual foods) as quartiles.

Results

Median total fruit and vegetable intake ranged from 110 g/day to 331 g/day for quartile one between studies and from 462 g/day to 1,007 g/day for quartile four.

The results were adjusted for several variables (a total of 14), e.g. age at menarche, menopausal status at follow-up, age at first-born child, family history of breast cancer, smoking status, body mass index and alcohol intake.

Intakes of fruit, fruit juice, total fruit, total vegetables and total fruit and vegetables were not associated with breast cancer.

For 100 g/day intake increment, the relative risks and 95% confidence intervals were as follows: total fruit 0.99 (0.98 to 1.00); fruit 0.99 (0.97 to 1.00); fruit juice 0.99 (0.97 to 1.02); total vegetables 1.00 (0.97 to 1.02); and total fruit and vegetables 0.99 (0.98 to 1.00).

Compared with the lowest quartiles of intake, the relative risks and 95% confidence intervals for the highest quartiles were as follows: total fruit 0.93 (0.86 to 1.00); fruit 0.93 (0.84 to 1.02); fruit juice 0.93 (0.86 to 1.00); total vegetables 0.96 (0.89 to 1.04); and total fruit and vegetables 0.93 (0.86 to 1.00).

To examine whether there was an effect of very high fruit and vegetable consumption, total fruit, total vegetable and total fruit and vegetable intakes were categorised into deciles. No associations were found, e.g. the relative risk for the uppermost decile of intake versus the lowermost was 0.96 for total fruit and vegetables (95% confidence interval 0.83 to 1.10).

None of the botanical groups or the individual fruit and vegetables were associated with breast cancer risk.

Comment

Evidence of an association between breast cancer risk and fruit and vegetable consumption has been based mainly on case-control studies. This paper, pooling results from cohort studies found no association. This is a classic example of how differences in research quality can lead to conflicting conclusions; and how important it is to continue to update evidence with better quality research when it becomes available.

On balance, cohort studies produce better evidence. In cohort studies, participants are asked about their diets and then followed for a number of years to see if they develop the disease. In contrast, participants in case-control studies have already developed the disease and are asked to remember their diets from previous years. As well as the possibility of recall bias, bias can also occur in the selection of participants, e.g. patients willing to participate in a study may differ from those who are not. Typically case-control studies have fewer participants than cohort studies (hundreds or less versus several thousand). The US Nurses’ Health Study for example, with over 150,000 participants, contributed to this analysis. Larger sample sizes reduce the possibility that chance has influenced the observed results.