Vegetables, fruit, and cancer

[Guest guest]

Hi All,

A new article has arrived on " Vegetables, fruit, and cancer " .

Link LB, Potter JD.

Raw versus cooked vegetables and cancer risk.

Cancer Epidemiol Biomarkers Prev. 2004 Sep;13(9):1422-35.

PMID: 15342442

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve & db=pubmed & dopt=Abstra\

ct & list_uids=15342442 & query_hl=8

http://tinyurl.com/8x74v

See the below for the current article, with is not in Medline yet and seems to

be

likely to not have an abstract when Medline does present the paper.

There appears to be confusion.

In keeping with the policy for posting in this forum, the below is highly

abbreviated, but is pdf-available.

Vegetables, fruit, and cancer

Lancet. 2005 Aug 6-12;366(9485):527-530

D Potter

A role for plant foods in the maintenance of health has been known for several

thousand years. Plants and plant extracts also provide the bulk of the

pharmacopoeia. In 1991, Steinmetz and Potter1 and 2 summarised the available

(largely case-control) data on cancer, vegetables, and fruit, concluding that

the

evidence was consistent with higher consumption being associated with a lower

risk

of many epithelial cancers. The 1997 World Cancer Research Fund report,3

reviewing

the available literature to the end of 1996, concluded that vegetables and fruit

were probably or convincingly associated with a lower risk of cancers of mouth,

oesophagus, lung, stomach, colorectum (vegetables only), larynx, pancreas,

breast,

and bladder. In October 2004, Hsin-Chia Hung and colleagues,4 using data from

the

Harvard Nurses Health and Health Professionals follow-up studies,5 concluded

that

vegetables and fruit were associated with a lower risk of cardiovascular

disease,

but that the relation with cancer, overall, was null. Clearly there is an

inconsistency here. Is there an explanation?

First, is Hung and colleagues' report4 consistent with earlier findings from the

Harvard investigators? Previously, Walter Willett's group at Harvard had

reported

lower risks in association with high intakes of vegetables and fruit for

premenopausal, but not postmenopausal, breast cancer (and, in 1993, a reduced

risk

for all breast cancer with vegetables alone6), female lung cancer,7 male bladder

cancer but only for cruciferae,8 non-Hodgkin lymphoma,9 and prostate cancer (for

fruit only).10 Only for colon cancer have they previously reported no

association

with vegetables and fruit.11

.... For lung cancer, cohort studies again suggest, if any association, a lower

risk

for women but not men;12, 19, 20, 21, 22, 23, 24 and 25 by far the largest study

(n

about 500 000) shows a reduced risk associated with fruit consumption (0·60,

0·46–0·78), but not vegetables.25

.... So there is something of a paradox. The literature up to 1997 was sufficient

to

conclude that vegetables and fruit were associated with a lower risk of various

cancers.3 The earlier Harvard findings were included in the World Cancer

Research

Fund review, but even as a separate body of literature, this group has, as noted

above, independently reported lower risks for several cancers. Against the

background of all findings, there are several possible conclusions about the

results

of Hung and colleagues.4

The first possibility is that there really is an association between lower risk

of

cardiovascular disease and intake of vegetables and fruit, but none between such

intake and cancer, and earlier papers, including those of the Harvard group, are

somehow in error. ... “high fruit and vegetable intakes were associated with

lower

risks of lung cancer in women” (0·79, 0·59–1·06);7 but for bladder cancer,

“consumption of vegetables other than cruciferae may not confer appreciable

benefit”

(0·72, 0·47–1·09).8 So has the Harvard group never been sure which way to jump

about

cancer and vegetables and fruit?

.... First, Hung and colleagues implicitly use the discrepancy in findings for

cardiovascular disease and cancer to argue that null findings cannot be solely a

measurement-error problem; measurement error would show itself with all

outcomes.

There is some merit to this argument, although the Harvard group earlier used

exactly this argument to show that their finding that fat was unrelated to

breast

cancer28 was not due to measurement error because there was still an association

between fat and colon cancer.29 Subsequently, in the Harvard studies, the

association between fat and colon cancer has gone away30 and 31 and an

association

between saturated fat and breast cancer has emerged,32 though not in all

settings.33

Second, cardiovascular disease is a small cluster of diseases, whereas cancer is

a

much larger and more heterogeneous collection of outcomes. ...

Third, the median intake of vegetables and fruit in the Harvard cohorts is

greater

than five servings a day for both men and women, much higher than the US norm.

....

Finally, whilst it might be useful to collapse across all cancers to make a

public-health point, any given individual's risk will be elevated for a much

smaller

subset of cancers; grouping sites thus obscures not only important biology, but

also

useful information for individuals.

We still have an unclear picture of the relation between diet and chronic

disease,

the role of specific foods, specific meal patterns, and changes over time ...

Al Pater, PhD; email: old542000@...

__________________________________________________