Journal of Internal Medicine 2001; 250: 280±290
REVIEW
Fruit and vegetable consumption in the prevention
of cancer: an update
P. TERRY1,2, J. B. TERRY3 & A. WOLK1
From the 1Department of Medical Epidemiology, The Karolinska Institute, Stockholm, Sweden; 2Department of Epidemiology and Social Medicine,
Albert Einstein College of Medicine, Bronx, NY, USA; and the 3Department of Information and Nutrition, National Food Administration, Uppsala,
Sweden
Abstract. Terry P, Terry JB, Wolk A (The Karolinska
Institute, Stockholm, Sweden; Albert Einstein Col-
lege of Medicine, Bronx, NY, USA; and National
Food Administration, Uppsala, Sweden). Fruit and
vegetable consumption in the prevention of cancer:
an update (Review). J Intern Med 2001; 250:
280±290.
Fruit and vegetables are considered to be important
means of preventing cancer through diet. However,
several recent large prospective cohort studies and
clinical trials have found no associations regarding
several major cancers, including cancers of the
Introduction
Over the past few decades the presumed bene®cial
effects of fruit and vegetables have been underscored
by the emergence of large-scale public health
campaigns, such as the well-known `Five a Day'
program, and the evolution of the scienti®c term
`antioxidants' into a household word and a multi-
billion dollar industry. The consistency with which
fruit and vegetables were previously shown to be
bene®cial to one form of cancer after another was
comforting, especially in the midst of the waxing
and waning credibility given to health claims
regarding several other dietary factors, such as oat
bran and eggs. Therefore, the results of recent large
280
colon, breast and stomach. Although these results
are disconcerting, especially for health-conscious
individuals, there is also mounting evidence that
people who avoid fruit and vegetables completely,
or consume very little, are indeed at increased risk
of these cancers. These individuals in particular
should be identi®ed, with the goal of helping them
®nd ways to include more fruit and vegetables in
their diets.
Keywords: diet, fruit, guidelines, neoplasms, primary
prevention, vegetables.
prospective cohort studies [1±3] and clinical trials
[4, 5] that found no association between fruit and
vegetable consumption and cancer at several sites
remain disconcerting. However, these surprising
negative studies are just some of the many studies
on fruit and vegetable consumption and cancer to be
published in the past few years, which together serve
to emphasize several basic ideas. Cancer is not a
single disease with a single aetiology, but is rather a
general disease category that includes many distinct
diseases with distinct aetiologies. Likewise, fruit and
vegetables are not a single exposure. Some categor-
ies of fruits or vegetables, for example, may be more
important in the aetiology of certain cancers than in
others.
ã 2001 Blackwell Science Ltd
 REVIEW: FRUIT, VEGETABLES AND CANCER
Studies published in the past few years provide the
basis for this review, which is divided into two major
categories: cancers occurring along the digestive
tract and cancers occurring at other sites. Regarding
digestive tract cancers, there is a possibility both of a
direct action by fruit and vegetable components on
digestive tract tissues and a systemic action. How-
ever, cancers at different sites along the gastrointes-
tinal tract obviously have different aetiologies. This
is demonstrated by diverging incidence rates over
the past few decades, such as the noted increase in
oesophageal adenocarcinomas in western popula-
tions, but not oesophageal squamous-cell carcino-
mas [6], and the simultaneous decrease in the
incidence of distal, but not proximal, stomach
cancers [7]. These differences in incidence rates
underscore the need for separate analyses of diges-
tive tract cancers, as well as other cancers, by site
and histological type to better understand their
speci®c aetiologies.
Cancer sites of the digestive tract considered in
this review include cancers of the mouth, pharynx
and larynx, oesophagus, gastro-oesophageal junc-
tion (gastric cardia), stomach, colon and rectum.
Cancers at other sites considered in this review
include cancers of the lung, breast, prostate, bladder,
kidney, pancreas, endometrium, and non-Hodgkins
lymphoma. Cancers at sites for which few studies of
fruit and vegetables exist, such as cancers of the
ovary, thyroid, cervix, nasal cavity, small intestine,
and skin, are not included in this review.
Cancers of the digestive tract
Cancers of the mouth, pharynx and larynx
Cancers of the mouth, pharynx and larynx together
would be the seventh most common cancer world-
wide [8]. Incidence of these cancers tends to be higher
in males (approximately 18/100 000 per year) than
in females (approximately 5/100 000 per year) [8].
Fruit and vegetable consumption has been associated
with lowered risk of oral, pharynx and larynx cancer
in at least 20 case±control studies [9±14], approxi-
mately halving the risk amongst individuals with
high (typically four to six servings per day) compared
to low consumption (typically one to two servings per
day). These earlier studies are supported by ®ve
recent case±control studies that generally found
inverse associations with fruit and vegetable
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
281
consumption and antioxidant intake [15±19].
Because of their rarity in af¯uent countries, these
cancers are often combined into a single disease
outcome, which makes separate conclusions for each
cancer dif®cult. For the same reason, prospective
cohort studies are scarce. Nonetheless, the 20-fold
variation in risk of these cancers across geographical
regions [8] suggests that environmental factors,
including diet, may be important in the respective
aetiologies of these cancers.
Cancers of the oesophagus
Oesophageal cancer is the ninth most common
cancer worldwide [8]. Incidence of oesophageal
cancer is highest amongst males in developing
countries (approximately 12/100 000 per year) [8].
Fruit and vegetable consumption has been shown to
reduce the risk of oesophageal cancers in many case±
control studies [9±12, 20], although because of the
low incidence of these cancers in af¯uent countries,
data from prospective cohort studies are lacking.
Recently, studies have begun to focus separately on
the two main histological types of oesophageal
cancer, squamous-cell and adenocarcinoma. The
incidence of oesophageal adenocarcinoma, which
historically accounted for less than 10% of all
oesophageal cancer cases, has increased dramatically
over the last decades, and they now constitute up to
50% of oesophageal cancers occurring in developed
countries [6]; meanwhile, the incidence of squa-
mous-cell carcinoma of the oesophagus has remained
stable [21]. These diverging incidence rates suggest
different aetiologies and underscore the need for
separate analyses by histological type.
In the most recent case±control study of diet in
relation to oesophageal cancer risk, in which
histological types were analysed separately, individ-
uals with high fruit and vegetable consumption
(four to six servings per day) had 40±60% lower
risks of both histological types of oesophageal
cancer compared with individuals with low con-
sumption (one to two servings per day) [22]. Only
one other study, of smaller sample size, investigated
separately both main histological types of oesopha-
geal cancer in relation to fruit and vegetable
consumption [23]. The results of that study also
support inverse associations with both histological
types of oesophageal cancer. Thus, there is cur-
rently no clear evidence that the association
282 P . T E R R Y et al.
between fruit and vegetable consumption and
oesophageal cancers varies meaningfully according
to histological type.
Cancer of the stomach
Stomach cancer is the second most common form of
cancer amongst men and women combined, despite
its declining incidence in recent decades [8]. Migrants
from Japan (age-standardized incidence approxi-
mately 50/100 000) to the United States (age-
standardized incidence approximately 6/100 000)
[8] tend to retain the high risk that characterizes their
home country [24]. However, their children tend to
have risks closer to those observed in the US,
indicating the importance of environmental factors,
including factors related to diet [25], in the aetiology
of this cancer.
There are more studies of fruit and vegetable
consumption in relation to stomach cancer than any
other cancer site. In contrast to the majority of at
least 30 case±control studies that show inverse
associations between fruit and vegetable consump-
tion and stomach cancer risk [9±12], the results of
four prospective cohort studies are mixed. Two of
these cohort studies [26, 27] found inverse associ-
ations, whilst the two others [3, 28] found no
association with fruit and vegetable consumption.
The reasons for lack of consistency in the results of
these cohort studies are unclear, but perhaps certain
characteristics of the different study population may
modify the association with fruit and vegetables. For
example, the investigators of one of the cohort
studies [3] speculated that the relatively high level of
fruit and vegetable consumption in their study
population might have precluded their detecting
an association. Indeed, the two studies that found
inverse associations were conducted in populations
with relatively low fruit and vegetable consumption
[3, 28] (see Table 1).
In addition to the potential differences in results
according to certain characteristics of the study
populations, differences in results according to
aspects of study methodology are also important to
Table 1 Fruit and vegetables in the prevention of stomach cancer
Increasing fruit and vegetables Amongst persons with
and stomach cancer varied diets
True association? Unclear
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
consider. It has become apparent for several cancer
sites that the results of case±control studies gener-
ally are more consistent in showing bene®ts from
high fruit and vegetable consumption than the
results of prospective cohort studies [29]. Therefore,
the possibility of bias in previous case±control
studies cannot be ruled out. For example, spurious
inverse associations can result in case±control
studies when case subjects systematically recall less
fruit and vegetable consumption (compared with
control subjects) in order to explain their being
diagnosed with cancer [30]. This type of bias is
particularly worrisome in studies of fruit and veget-
able consumption, as preconceptions about the
wholesome effects of fruit and vegetables are com-
mon amongst the public. Moreover, discomfort and
other symptoms of undiagnosed stomach cancer can
lead to true lowered fruit and vegetable consump-
tion amongst cases [3], which can itself cause
patients to systematically underestimate their pre-
morbid diet.
Most randomized trials of antioxidant vitamins,
such as b-carotene and a-tocopherol, have found
no reduction in the risk of most cancers, including
cancers of the stomach [31±34]. However, it is
possible that because of the long latency period of
stomach cancer [35], observational studies with
long follow-up times show an effect that could not
be observed over the shorter follow-up periods of
the randomized trials (these trials ranged from
approximately 4 to 12 years, with an average of
close to 7 years). It is interesting to note that these
randomized trials were conducted in well-nourished
western populations. On the contrary, an interven-
tion trial of antioxidant vitamins conducted in an
under-nourished population in the Linxian Prov-
ince of China [36] observed overall reductions in
cancer rates, suggesting that inverse associations
may be stronger amongst, or even limited to,
individuals with low intake of antioxidants. In a
recent clinical trial in rural Colombia [37], supple-
mental b-carotene and vitamin C both increased
the rate of regression of gastric cancer precursor
lesions.
With speci®c fruit and With fruit and vegetable/
vegetable sub-categories antioxidant de®ciency
Unclear Likely
 REVIEW: FRUIT, VEGETABLES AND CANCER
Cancers of the gastro-oesophageal junction (gastric
cardia)
The incidence of cancers of gastric cardia, the
proximal part of the stomach adjacent to the oeso-
phagus, may be slowly rising in developed countries
[6], although the potential for misclassi®cation of
gastric cardia cancers as distal gastric cancers makes
secular trends uncertain [38]. Gastric cardia cancers
have often been studied together with cancers of the
distal stomach or, more recently, together with
oesophageal adenocarcinoma [39±41]. The few
studies that have examined gastric cardia cancer
separately generally report inverse associations with
fruit and vegetable consumption [42±45], although
the number of cardia cases in these studies tends to be
small. On the contrary, a recent large case±control
study of cardia cancer [22], with careful classi®cation
according to subsite, failed to show any association
with fruit and vegetable consumption whilst simul-
taneously ®nding a strong inverse association
between fruit and vegetable consumption and risks
of both squamous-cell and adenocarcinoma of the
oesophagus.
The results of a recent in vitro study suggest that
ingested nitrates combine with oxygen and com-
pounds contained in saliva to form carcinogenic
nitrosamines, especially at the gastric cardia [46]. As
vegetables are a major source of ingested nitrate [47],
an otherwise inverse association with total fruit and
vegetable consumption might be partly or entirely
nulli®ed by a concurrent fuelling of the nitrate±
nitrite±nitrosamine pathway [48]. Therefore, the
inclusion of cardia cancer cases in previous studies
of total stomach cancer might have led to weaker
associations than what would have been observed if
only distal stomach cancer was studied. Prospective
cohort studies of fruit and vegetable consumption
have not examined distal stomach cancer as a
separate outcome. Future studies should attempt to
study gastric cardia cancer separately, but in con-
junction with cancers of the oesophagus and distal
Table 2 Fruit and vegetables in the prevention of colorectal cancer
Increasing fruit and vegetables Amongst persons with
and colorectal cancer varied diets
True association? Unlikely/Unclear
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
283
stomach, so that the effects of dietary exposures can
be compared amongst these different sites [49].
Cancers of the colon and rectum
Colorectal cancer is the second most common cancer
amongst women, and the third most common
amongst men [8]. Incidence of colorectal cancer
tends to be higher in developed countries, generally
30±40/100 000 per year [8]. Fruit and vegetable
consumption has been inversely associated with
colon and rectal (colorectal) cancer risks in many
case±control studies over the years [9±12, 50],
leading to a general consensus regarding the seem-
ingly causal nature of this association. However,
recent ®ndings from large prospective cohort studies
of colorectal cancer [1] and clinical trials of colorec-
tal adenomas [4] have called into question this
putative association, showing no reduction in risks
with increased fruit and vegetable consumption. This
surprising development suggests the possibility that
selective recall of fruit and vegetable consumption in
the case±control studies might have resulted in
spurious inverse associations (see Table 2).
Even more recently, the results of another large
prospective cohort study suggest that the associ-
ation between fruit and vegetable consumption
and colon cancer risk is limited to individuals with
low consumption [51]. In that study, individuals
with fewer than 1.5 servings of fruit and veget-
ables per day had a 65% higher risk of colon
cancer compared with individuals consuming
more than 2.5 servings per day. Amongst individ-
uals consuming more than 2.5 servings per day,
on the contrary, there was no association between
fruit and vegetable consumption and colon cancer
risk. Thus, studies that were conducted amongst
health-conscious and well-nourished individuals,
such as the large prospective cohort study of
nurses and other health professionals in the US [1],
might have found no association because relatively
few individuals in these populations are truly
With speci®c fruit and With fruit and vegetable/
vegetable sub-categories antioxidant de®ciency
Possibly with fruit high in Likely
vitamin C
284 P . T E R R Y et al.
de®cient in the nutrients contained in fruit and
vegetables. As was mentioned above in relation to
stomach cancer, future studies should explore the
possibility that inverse associations with colorectal
cancer are stronger amongst, or even limited to,
individuals with low fruit and vegetable consump-
tion.
Other cancer sites
The most widely studied cancers in relation to fruit
and vegetable consumption, other than cancers of
the digestive tract, include cancers of the lung,
breast, prostate, bladder, kidney, pancreas, endome-
trium, and non-Hodgkins lymphoma. Cancers that
have received very limited attention include those
occurring in the ovary, thyroid, cervix, nasal
cavity, small intestine, and skin [9±12]. These
latter are not reviewed here because of the paucity
of studies.
Cancer of the lung
Lung cancer is the most common cancer worldwide.
Incidence of lung cancer is the highest amongst men
living in developed countries, approximately 60±70/
100 000 per year. Fruit and vegetable consumption
has been inversely associated with lung cancer risk
in most previous case±control and prospective
cohort studies [9±12, 52]. However, the results of
two recent prospective cohort studies are mixed [53,
54]. One of these studies [54], using a general
population of Dutch citizens sampled from commu-
nity registers, observed 20±40% lower risks amongst
individuals consuming the recommended ®ve serv-
ings per day (expressed in grams per day) compared
with individuals with one to two servings of fruit
and vegetables per day. The other cohort study [53],
using data from health professionals who had a
higher range of fruit and vegetable consumption and
a high prevalence of multivitamin use [1], found
weak or no associations. These results, as was noted
above regarding stomach and colorectal cancers,
suggest that recommendations to increase fruit and
Table 3 Fruit and vegetables in the prevention of lung cancer
Increasing fruit and vegetables Amongst persons with
and lung cancer varied diets
True association? Unclear
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
vegetable consumption to decrease lung cancer risk
may be most bene®cial for individuals who consume
low amounts of fruit and vegetables, for example,
less than two servings per day (see Table 3).
As mentioned above, randomized trials of b-caro-
tene and a-tocopherol do not support an inverse
association between these antioxidants and lung
cancer risk [31±34]. In addition to the speculation
that trials of relatively short duration may not
adequately account for the long latency period of
this cancer, it is also possible that antioxidants
contained in fruit and vegetables other than b-caro-
tene and a-tocopherol are more important. Indeed, a
recent study of two pooled prospective cohorts [55]
found the strongest inverse associations between
antioxidants and lung cancer risk for a-carotene and
lycopene. The inverse association between a-caro-
tene and lung cancer risk was also stronger than the
association with b-carotene in another recent pros-
pective cohort study [56]. Along similar lines, the
inverse associations with folate, vitamin C, and
b-cryptoxanthin were stronger than the associations
with a-carotene, b-carotene, and lycopene in
another recent prospective cohort study from
Holland [57]. Thus, negative results from random-
ized trials of b-carotene and a-tocopherol
supplements do not preclude possible bene®ts of
longer-term exposure to these nutrients or other
potentially anticarcinogenic nutrients contained in
fruit and vegetables.
Cancer of the breast
Breast cancer is the most common cancer amongst
women worldwide [8]. Incidence of breast cancer is
higher in developed countries, where approximately
55±60/100 000 women are diagnosed with this
disease annually [8]. A meta-analysis of vegetable
consumption in relation to breast cancer risk found
consistent inverse associations with breast cancer
risk in previous epidemiological studies, mostly of
case±control design [58], whilst studies of fruit
consumption tend to show weaker associations. This
may be explained either by a true stronger associ-
With speci®c fruit and With fruit and vegetable/
vegetable sub-categories antioxidant de®ciency
Unclear Likely
 REVIEW: FRUIT, VEGETABLES AND CANCER
ation with vegetable consumption, or a greater
degree of measurement error in the assessment of
fruit consumption, and perhaps by a narrower range
of fruit consumption in the studied populations (the
latter two instances would lead to attenuated relative
risk estimates).
A recent analysis of pooled data from several
prospective cohort studies of diet and breast cancer
risk [2] revealed no associations with either fruit or
vegetable consumption. The reasons for the diver-
ging results according to study design are unknown,
yet may be the result of biased recall, which is an
inherent threat to the validity of case±control
studies [30]. A similar pattern of diverging results
according to study design was noted above for
cancers of the stomach, colorectum, and lung. In
addition to total fruit and vegetable consumption,
the pooled analysis of cohort studies did not reveal
any association with fruit or vegetable sub-categor-
ies, although data were somewhat limited in this
regard [59] (see Table 4).
The hypothesis remains, therefore, that certain
sub-categories of fruit and vegetables, particularly
brassica vegetables, such as cabbage, broccoli and
cauli¯ower, may lower the risk of breast cancer [59].
A recent large case±control study addressed this
hypothesis and found a 40±50% lower risk of post-
menopausal breast cancer amongst women consu-
ming one to two servings of brassica vegetables per
day compared with women consuming little or none
[60]. In that study, no associations were observed
for consumption of non-brassica vegetables or fruit.
The independent inverse association between bras-
sica vegetable consumption and breast cancer risk in
this study is supported by experimental data show-
ing that, amongst post-menopausal women, brassica
vegetable consumption signi®cantly increased the
ratio of 2-hydroxyestrone to 16-a-hydroxyestrone in
urine [61], which might be inversely associated with
breast cancer risk. In agreement with this latter
®nding and with the results of animal experiments,
indole-3-carbinol found in brassica vegetables was
shown to arrest the growth of human breast cancer
cells [62].
Table 4 Fruit and vegetables in the prevention of breast cancer
Increasing fruit and vegetables Amongst persons with
and breast cancer varied diets
True association? Unlikely/Unclear
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
285
Cancer of the prostate
Prostate cancer is the fourth most common cancer
amongst men and is more common in developed
countries, where incidence is approximately 40/
100 000 per year [8]. Case±control studies, as well
as prospective cohort studies, generally have not
found inverse association between total fruit and
vegetable consumption and prostate cancer risk
[9±12, 63±66]. However, the results of several
recent cohort studies point to an inverse association
speci®cally with lycopene [67], an antioxidant
carotenoid found mostly in tomatoes. On the cellular
level, it was shown that carotenoids protect lym-
phocytes from NO2 radical damage, and lycopene
was at least twice as effective as b-carotene in that
regard [67]. Proposed biological mechanisms
remain speculative, however. Further examination
of the association between lycopene (and tomato)
intake and prostate cancer risk is warranted (see
Table 5).
Cancer of the bladder
Bladder cancer is the tenth most common cancer
worldwide [8]. This disease is most common
amongst men living in developed countries, who
are diagnosed at a rate of approximately 17/
100 000 per year [8]. Fruit and vegetable con-
sumption has been inversely associated with bladder
cancer risk in a large majority of at least eight case±
control studies [9±12, 68]. The results of three
recent prospective cohort studies [69±71] tend to
support these ®ndings, although the association is
sometimes weak and statistically nonsigni®cant. For
example, one of these studies [69] found an inverse
association only with brassica vegetables. Further
studies are therefore necessary before conclusions
regarding this association can be made.
Cancer of the kidney
Kidney cancer is most common amongst men living
in developed countries, who are diagnosed at a rate
With speci®c fruit and With fruit and vegetable/
vegetable sub-categories antioxidant de®ciency
Possibly with brassica vegetables Unclear
286 P . T E R R Y et al.
Table 5 Fruit and vegetables in the prevention of prostate cancer
Increasing fruit and vegetables Amongst persons with
and prostate cancer varied diets
True association? Unclear
of approximately 9/100 000 per year [8]. Renal-cell
cancers comprise 80% of all kidney cancers and
studies of fruit and vegetable consumption have
been limited to this form of the disease [72]. At least
six case±control studies of this disease found weak to
moderate inverse associations with fruit and veget-
able consumption [9±12, 72], with an average 20%
reduced risk amongst individuals meeting or
exceeding the recommended ®ve servings per day
compared with individuals consuming low amounts
(typically one to two servings per day). In some of
these studies, particular sub-categories of fruit and
vegetables, such as orange and dark green veget-
ables [73] and brassica vegetables [74], were shown
to be most important. One prospective cohort study
[75], albeit with a low number of cases, also found
an inverse association between fruit and vegetable
consumption and renal-cell cancer. However, as
strong inverse associations have not generally been
found, and data from prospective cohort studies are
lacking, conclusions regarding causality cannot yet
be made. The association between fruit and veget-
able consumption and kidney cancer warrants
further investigation.
Cancer of the pancreas
Pancreatic cancer is also the most common amongst
men living in developed countries, who are diag-
nosed at a rate of approximately 8/100 000 per
year [8]. Fruit and vegetable consumption was
inversely associated with pancreatic cancer in a
large majority of at least nine case±control studies
[9±12, 76], although the results of three prospective
cohort studies are less clear. These latter studies
found either no association with fruit and vegetable
consumption [77], a nonsigni®cant inverse associ-
ation in an elderly population [78], or an inverse
association mainly limited to legumes and dried fruit
amongst Seventh Day Adventists [79]. Given these
mixed results and the paucity of prospective cohort
studies, the association between fruit and vegetable
consumption and pancreatic cancer risk warrants
further investigation.
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
With speci®c fruit and With fruit and vegetable/
vegetable sub-categories antioxidant de®ciency
Possibly with tomatoes (lycopene) Unclear
Cancer of the endometrium
Endometrial cancer is the seventh most common
cancer amongst women [8]. Endometrial cancer
occurs three times more frequently in developed
countries, where incidence is approximately 11/
100 000 per year [8]. Fruit and vegetable con-
sumption has shown mixed results in relation to
endometrial cancer risk in ®ve early case±control
studies [9±12, 80]. Two recent case±control studies
[81, 82] found inverse associations with the con-
sumption of vegetables, but not fruit. One can
speculate that brassica vegetables may help to
explain the inverse association found for total
vegetable consumption in these two studies because
of their bene®cial effect on oestrogen metabolism
[61]. However, the earlier case±control studies do
not support a distinction between the respective
associations with fruit and vegetable consumption
[80], nor does a review of brassica vegetables in
relation to various cancer sites support a consistent
association between this sub-category of vegetables
and endometrial cancer risk [83]. Considering the
mixed results of case±control studies and the lack of
prospective cohort studies, further investigation of
the association between fruit and vegetable con-
sumption and endometrial cancer risk is warranted.
Non-Hodgkins lymphoma
Non-Hodgkins lymphoma is most common amongst
men living in developed countries, who are diag-
nosed at a rate of approximately 10/100 000 per
year [8]. Fruit and vegetable consumption has
recently been shown to lower the risk of non-
Hodgkins lymphoma in a large prospective cohort
study of female American nurses [84]. Consumption
of three or more servings per day of vegetables,
compared to one or fewer servings per day, was
associated with a 40% decreased risk of this cancer.
In comparison, the association with fruit consump-
tion was slightly weaker and was statistically
nonsigni®cant. An earlier case±control study [85]
found citrus fruit and dark green vegetables
 REVIEW: FRUIT, VEGETABLES AND CANCER
decreased the risk in men, but not women. The
association between fruit and vegetable consump-
tion and non-Hodgkins lymphoma warrants further
investigation.
Antioxidative mechanisms
Fruit and vegetables are rich in antioxidants that
are thought to neutralize the harmful effects of
DNA-damaging free radicals such as those pro-
duced by metabolism and by external factors such
as pollution and smoking [86]. The apparent lack
of ef®cacy of supplemental b-carotene and a-toco-
pherol in reducing the risk of cancer was therefore
disconcerting [5]. As discussed above, however,
these trials were conducted in well-nourished
populations and were of limited duration with
respect to the long latency period of these cancers.
Fruit and vegetables are also the primary source of
many other substances with hypothesized antioxi-
dative (and other antiocarcinogenic) properties,
such as carotenoids, dithiolthiones, glucosinolates,
indoles, isothiocyanates, ¯avonoids, phenols, prote-
ase inhibitors, plant sterols, allium compounds and
limonene [47].
When associations between antioxidant intake and
cancer are stronger in the presence of a third factor,
1 Conclusion
one that is presumed to be a marker for increased free
radical exposure (such as smoking), this may be seen
as support for the neutralizing role antioxidants are
thought to play in the pathway between fruit and
vegetable consumption and disease risk. From this
perspective, one can speculate that antioxidant
mechanisms will be more important in the presence
of higher levels of oxidizing free radicals. Indeed,
stronger associations with fruit and vegetable con-
sumption (and antioxidant vitamins) have been seen
amongst smokers when such analyses are performed,
for example with oesophageal squamous-cell carci-
noma [87] and stomach cancer [42, 88]. However,
the b-carotene/a-tocopherol trials do not support a
bene®t to smokers.
Folic acid
Folic acid (folate), found in several fruit and veget-
ables, has attracted increasing attention in recent
years, particularly regarding its potential role in the
prevention of colorectal cancer [89±91]. Folate is
crucial for normal DNA synthesis and repair, and
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
287
de®ciency of this nutrient is hypothesized to lead to
cancer through consequent faults in these processes
[91]. Recent prospective cohort studies have shown
inverse associations between folate intake and
colorectal cancer risk [90], as well as a greater risk
of alcohol-induced cancer in the colorectum [92]
and breast [93] amongst individuals with low folate
intake.
Folate de®ciency (erythrocyte folate <140 ng mL)1
or plasma folate <3 ng mL)1) [93] is considered to be
one of the most common nutrient de®ciencies in
developed countries, estimated to affect approxi-
mately 10% of the US population [91]. Supplemental
sources of folate, as well as folate obtained directly
from fruit and vegetable consumption, appear to
mitigate the risk of colorectal cancer [91]. The results
of ongoing clinical trials of folate supplementation, as
yet unpublished, may provide valuable evidence
regarding this hypothesis. Meanwhile, because of
the limited evidence [91], some investigators suggest
that population-wide interventions with folate sup-
plements to prevent cancer would not necessarily be
prudent at this time. Recommendations to increase
fruit and vegetable consumption, however, do not
have need of any such caveat.
Fruit and vegetables are amongst the most widely
studied dietary risk factors for cancer, yet many
questions remain. Differences in the observed asso-
ciations by cancer site, histological type, and
speci®c sub-categories of fruit and vegetables are
just being realized and underlying biological mech-
anisms are coming to light, but still remain
unclear. Fruit and vegetable consumption is not a
single exposure, nor is cancer a single disease. It is
therefore important for the public to realize that a
null ®nding from a single study of fruit and
vegetable consumption and a speci®c form of
cancer, viewed in the scienti®c literature or in the
popular media, does not necessarily mean that the
advice to increase fruit and vegetable consumption
is no longer prudent.
Considering the results of recent dietary surveys,
most people in Europe [95] and in the US [96] do not
meet the minimum recommended level [11] of
approximately ®ve servings of fruit and vegetable
consumption per day. This suggests that public
health efforts to increase fruit and vegetable con-
288 P . T E R R Y et al.
sumption, such as the noted Five a Day program
[97], have the potential to reduce cancer incidence.
However, weak or null ®ndings from recent pros-
pective cohort studies of several different cancer sites
suggest that the inverse associations with fruit and
vegetable consumption found in case±control stud-
ies may have been overstated [29]. On the other
hand, data from populations and individuals with
low fruit and vegetable consumption more consis-
tently show inverse associations, even amongst
prospective cohort studies [3, 26, 51].
In this context, it is important to note that
individuals with the lowest fruit and vegetable
consumption may be less exposed to, or less in¯u-
enced by, dietary messages aimed at the general
public [97]. In such a case, interventions such as the
Five a Day program [98] may succeed in raising the
average fruit and vegetable consumption of a pop-
ulation without necessarily modifying the dietary
habits of individuals with the lowest consumption.
From both clinical and public health perspectives,
individuals who are most likely to bene®t from
increased fruit and vegetable consumption should be
identi®ed, with the goal of helping them ®nd ways to
include more fruit and vegetables in their diets.
References
1 Michels KB, Edward G, Joshipura KJ et al. Prospective study of
fruit and vegetable consumption and incidence of colon and
rectal cancers. J Natl Cancer Inst 2000; 92: 1740±52.
2 Smith-Warner SA, Spiegelman D, Yaun SS et al. Intake of
fruits and vegetables and risk of breast cancer: a pooled
analysis of cohort studies. Jama 2001; 285: 769±76.
3 Botterweck AA, van den Brandt PA, Goldbohm RA. A
prospective cohort study on vegetable and fruit consumption
and stomach cancer risk in The Netherlands. Am J Epidemiol
1998; 148: 842±53.
4 Schatzkin A, Lanza E, Corle D et al. Lack of effect of a low-fat,
high-®ber diet on the recurrence of colorectal adenomas. Polyp
Prevention Trial Study Group. N Engl J Med 2000; 342:
1149±55.
5 Omenn GS. Chemoprevention of lung cancer: the rise and
demise of b-carotene. Annu Rev Public Health 1998; 19:
73±99.
6 Blot WJ, Devesa SS, Kneller RW, Fraumeni JF. Rising incidence
of adenocarcinoma of the esophagus and gastric cardia. Jama
1991; 265: 1287±9.
7 Howson CP, Hiyama T, Wynder EL. The decline in gastric
cancer: epidemiology of an unplanned triumph. Epidemiol Rev
1986; 8: 1±27.
8 Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide
incidence of 25 major cancers in 1990. Int J Cancer 1999; 80:
827±41.
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
9 Steinmetz KA, Potter JD. Vegetables, fruit, and cancer. I.
Epidemiology. Cancer Causes Control 1991; 2: 325±57.
10 Steinmetz KA, Potter JD. Vegetables, fruit, and cancer
prevention: a review. J Am Diet Assoc 1996; 96: 1027±39.
11 World Cancer Research Fund. Food, Nutrition and the Preven-
tion of Cancer: a Global Perspective. Washington, DC: World
Cancer Research Fund/American Institute for Cancer
Research, 1997.
12 Block G, Patterson B, Subar A. Fruit, vegetables, and cancer
prevention: a review of the epidemiological evidence. Nutr
Cancer 1992; 18: 1±29.
13 Marshall JR, Boyle P. Nutrition and oral cancer. Cancer Causes
Control 1996; 7: 101±11.
14 Riboli E, Kaaks R, EsteÂve J. Nutrition and laryngeal cancer.
Cancer Causes Control 1996; 7: 147±56.
15 Riboli E. Nutrition and cancer of the respiratory and digestive
tract: results from observational and chemoprevention studies.
Eur J Cancer Prev 1996; 5 (Suppl. 2): 9±17.
16 Nomura AM, Ziegler RG, Stemmermann GN, Chyou PH, Craft
NE. Serum micronutrients and upper aerodigestive tract cancer.
Cancer Epidemiol Biomarkers Prev 1997; 6: 407±12.
17 Negri E, Franceschi S, Bosetti C et al. Selected micronutrients
and oral and pharyngeal cancer. Int J Cancer 2000; 86:
122±7.
18 Bosetti C, Negri E, Franceschi S et al. Risk factors for oral and
pharyngeal cancer in women: a study from Italy and
Switzerland. Br J Cancer 2000; 82: 204±7.
19 Esteve J, Riboli E, Pequignot G et al. Diet and cancers of the
larynx and hypopharynx: the IARC multi-center study in
southwestern Europe. Cancer Causes Control 1996; 7: 240±52.
20 Cheng KK, Day NE. Nutrition and esophageal cancer. Cancer
Causes Control 1996; 7: 33±40.
21 Thomas RM, Sobin LH. Gastrointestinal cancer. Cancer 1995;
75: 154±70.
22 Terry P, Lagergren J, Wolk A, Nyren O. Fruit and vegetable
consumption in the prevention of esophageal and cardia
2 cancers. Eur J Cancer Prev 2001 (in press).
23 Tzonou A, Lipworth L, Garidou A et al. Diet and risk of
esophageal cancer by histologic type in a low-risk population.
Int J Cancer 1996; 68: 300±4.
24 Haenszel W, Kurihara M. Studies of Japanese migrants. I.
Mortality from cancer and other diseases among Japanese in
the United States. J Natl Cancer Inst 1968; 40: 43±68.
25 Kono S, Hirohata T. Nutrition and stomach cancer. Cancer
Causes Control 1996; 7: 41±55.
26 Terry P, Nyren O, Yuen J. Protective effect of fruits and
vegetables on stomach cancer in a cohort of Swedish twins. Int
J Cancer 1998; 76: 35±7.
27 Hertog MG, Bueno de Mesquita HB, Fehily AM, Sweetnam PM,
Elwood PC, Kromhout D. Fruit and vegetable consumption
and cancer mortality in the Caerphilly study. Cancer Epidemiol
Biomarkers Prev 1996; 5: 673±7.
28 Kneller RW, McLaughlin JK, Bjelke E et al. A cohort study of
stomach cancer in a high-risk American population. Cancer
1991; 68: 672±8.
29 Willett WC. Diet and cancer: one view at the start of the
millennium. Cancer Epidemiol Biomarkers Prev 2001; 10: 3±8.
30 Rothman KJ, Greenland S. Modern Epidemiology. Philadelphia:
Lippincott, 1998.
31 Hennekens CH, Buring JE, Manson JE et al. Lack of effect of
long-term supplementation with b-carotene on the incidence
 REVIEW: FRUIT, VEGETABLES AND CANCER
of malignant neoplasms and cardiovascular disease. N Engl J
Med 1996; 334: 1145±9.
32 Varis K, Taylor PR, Sipponen P et al. Gastric cancer and
premalignant lesions in atrophic gastritis: a controlled trial on
the effect of supplementation with a-tocopherol and b-caro-
tene. The Helsinki Gastritis Study Group. Scand J Gastroenterol
1998; 33: 294±300.
33 Omenn GS, Goodman GE, Thornquist MD et al. Effects of
a combination of b-carotene and vitamin A on lung
cancer and cardiovascular disease. N Engl J Med 1996;
334: 1150±5.
34 Albanes D, Malila N, Taylor PR et al. Effects of supplemental
a-tocopherol and b-carotene on colorectal cancer: results from
a controlled trial (Finland). Cancer Causes Control 2000; 11:
197±205.
35 Fujita S, Hatori T. Cell proliferation, differentiation, and
migration in the gastric mucosa: a study of the back-
ground of carcinogenesis. In: Farber T, Kawachi T,
3 Nagayo H, Sugano T, Weisburger, eds. Pathophysiology of
Carcinogenesis in Digestive Organs. Tokyo: University of
4 Tokyo Press, 1977.
36 Blot WJ, Li JY, Taylor PR et al. Nutrition intervention trials in
Linxian, China: supplementation with speci®c vitamin/min-
eral combinations, cancer incidence, and disease- speci®c
mortality in the general population. J Natl Cancer Inst 1993;
85: 1483±92.
37 Correa P, Fontham ET, Bravo JC et al. Chemoprevention of
gastric dysplasia: randomized trial of antioxidant supplements
and anti-helicobacter pylori therapy. J Natl Cancer Inst 2000;
92: 1881±8.
38 Ekstrom AM, Signorello LB, Hansson LE, BergstroÈm R,
Lindgren A, NyreÂn O. Evaluating gastric cancer misclassi®ca-
tion: a potential explanation for the rise in cardia cancer
incidence. J Natl Cancer Inst 1999; 91: 786±90.
39 Kabat GC, Ng SK, Wynder EL. Tobacco, alcohol intake, and
diet in relation to adenocarcinoma of the esophagus and
gastric cardia. Cancer Causes Control 1993; 4: 123±32.
40 Zhang ZF, Kurtz RC, Yu GP et al. Adenocarcinomas of the
esophagus and gastric cardia: the role of diet. Nutr Cancer
1997; 27: 298±309.
41 Brown LM, Swanson CA, Gridley G et al. Adenocarcinoma of
the esophagus: role of obesity and diet. J Natl Cancer Inst 1995;
87: 104±9.
42 Ekstrom AM, Sera®ni M, Nyren O, Hansson LE, Ye W, Wolk A.
Dietary antioxidant intake and the risk of cardia cancer and
noncardia cancer of the intestinal and diffuse types: a
population-based case±control study in Sweden. Int J Cancer
2000; 87: 133±40.
43 Ji BT, Chow WH, Yang G et al. Dietary habits and stomach
cancer in Shanghai, China. Int J Cancer 1998; 76: 659±64.
44 Palli D, Bianchi S, Decarli A et al. A case±control study of
cancers of the gastric cardia in Italy. Br J Cancer 1992; 65:
263±6.
45 Wu-Williams AH, Yu MC, Mack TM. Life-style, workplace, and
stomach cancer by subsite in young men of Los Angeles
County. Cancer Res 1990; 50: 2569±76.
46 Grant J, Moriya A, Williams C et al. Reaction of salivary nitrite
with gastric acid and vitamin C relevant to proximal gastric
cancer and Barrett's oesophagus. Gastroenterology 1999; 116:
A469.
47 Steinmetz KA, Potter JD. Vegetables, fruit, and cancer II.
Mechanisms. Cancer Causes Control 1991; 2: 427±42.
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
289
48 Terry P, Lagergren J, Ye W, Wolk A, Nyren O. Inverse
association between intake of cereal ®ber and risk of gastric
cardia cancer. Gastroenterology 2001; 120: 387±91.
49 Kurtz RC, Zhang ZF. Gastric cardia cancer and dietary ®ber.
Gastroenterology 2001; 120: 568±70.
50 Potter JD. Nutrition and colorectal cancer. Cancer Causes
Control 1996; 7: 127±46.
51 Terry P, Giovannucci E, Michels KB et al. Fruit, vegetables,
dietary ®ber, and the risk of colorectal cancer. J Natl Cancer
Inst 2001; 93: 120±6.
52 Ziegler RG, Mayne ST, Swanson CA. Nutrition and lung
cancer. Cancer Causes Control 1996; 7: 157±77.
53 Feskanich D, Ziegler RG, Michaud DS et al. Prospective study
of fruit and vegetable consumption and risk of lung cancer
among men and women. J Natl Cancer Inst 2000; 92:
1812±23.
54 Voorrips LE, Goldbohm RA, Verhoeven DT et al. Vegetable and
fruit consumption and lung cancer risk in The Netherlands
cohort study on diet and cancer. Cancer Causes Control 2000;
11: 101±15.
55 Michaud DS, Feskanich D, Rimm EB et al. Intake of speci®c
carotenoids and risk of lung cancer in 2 prospective US
cohorts. Am J Clin Nutr 2000; 72: 990±7.
56 Knekt P, Jarvinen R, Teppo L, Aromaa A, Seppanen R. Role of
various carotenoids in lung cancer prevention. J Natl Cancer
Inst 1999; 91: 182±4.
57 Voorrips LE, Goldbohm RA, Brants HA et al. A prospective
cohort study on antioxidant and folate intake and male lung
cancer risk. Cancer Epidemiol Biomarkers Prev 2000; 9:
357±65.
58 Gandini S, Merzenich H, Robertson C, Boyle P. Meta-analysis
of studies on breast cancer risk and diet: the role of fruit and
vegetable consumption and the intake of associated micronu-
trients. Eur J Cancer 2000; 36: 636±46.
59 Slattery ML. Does an apple a day keep breast cancer away?
Jama 2001; 285: 799±801.
60 Terry P, Wolk A, Persson I, Magnusson C. Brassica vegetables
and breast cancer risk. JAMA 2001.
61 Fowke JH, Longcope C, Hebert JR. Brassica vegetable con-
sumption shifts estrogen metabolism in healthy postmeno-
pausal women. Cancer Epidemiol Biomarkers Prev 2000; 9:
773±9.
62 Cover CM, Hsieh SJ, Tran SH et al. Indole-3-carbinol inhibits
the expression of cyclin-dependent kinase-6 and induces a G1
cell cycle arrest of human breast cancer cells independent of
estrogen receptor signalling. J Biol Chem 1998; 273:
3838±47.
63 Kolonel LN. Nutrition and prostate cancer. Cancer Causes
Control 1996; 7: 83±44.
64 Giles G, Ireland P. Diet, nutrition and prostate cancer. Int J
5 Cancer 1997; (Suppl): 13±7.
65 Nixon DW. Prostate cancer and nutrition. J S C Med Assoc
2000; 96: 85±6.
66 Chan JM, Giovannucci E, Andersson SO, Yuen J, Adami HO,
Wolk A. Dairy products, calcium, phosphorous, vitamin D,
and risk of prostate cancer (Sweden). Cancer Causes Control
1998; 9: 559±66.
67 Giovannucci E. Tomatoes, tomato-based products, lycopene,
and cancer: review of the epidemiologic literature. J Natl
Cancer Inst 1999; 91: 317±31.
68 La Vecchia C, Negri E. Nutrition and bladder cancer. Cancer
Causes Control 1996; 7: 95±100.
290 P . T E R R Y et al.
69 Michaud DS, Spiegelman D, Clinton SK, Rimm EB, Willett WC,
Giovannucci EL. Fruit and vegetable intake and incidence of
bladder cancer in a male prospective cohort. J Natl Cancer Inst
1999; 91: 605±13.
70 Nagano J, Kono S, Preston DL et al. Bladder-cancer incidence
in relation to vegetable and fruit consumption: a prospective
study of atomic-bomb survivors. Int J Cancer 2000; 86:
132±8.
71 Steinmaus CM, Nunez S, Smith AH. Diet and bladder cancer: a
meta-analysis of six dietary variables. Am J Epidemiol 2000;
151: 693±702.
72 Wolk A, Lindblad P, Adami HO. Nutrition and renal cell
cancer. Cancer Causes Control 1996; 7: 5±18.
73 Wolk A, Gridley G, Niwa S et al. International renal cell cancer
study. VII. Role diet. Int J Cancer 1996; 65: 67±73.
74 Yuan JM, Gago-Dominguez M, Castelao JE, Hankin JH, Ross
RK, Yu MC. Cruciferous vegetables in relation to renal cell
carcinoma. Int J Cancer 1998; 77: 211±6.
75 Fraser GE, Phillips RL, Beeson WL. Hypertension, antihyper-
tensive medication and risk of renal carcinoma in California
Seventh-Day Adventists. Int J Epidemiol 1990; 19: 832±8.
76 Howe GR, Burch JD. Nutrition and pancreatic cancer. Cancer
Causes Control 1996; 7: 69±82.
77 Zheng W, McLaughlin JK, Gridley G et al. A cohort study of
smoking, alcohol consumption, and dietary factors for pan-
creatic cancer (United States). Cancer Causes Control 1993; 4:
477±82.
78 Shibata A, Mack TM, Paganini-Hill A, Ross RK, Henderson BE.
A prospective study of pancreatic cancer in the elderly. Int J
Cancer 1994; 58: 46±9.
79 Mills PK, Beeson WL, Abbey DE, Fraser GE, Phillips RL. Dietary
habits and past medical history as related to fatal pancreas
cancer risk among Adventists. Cancer 1988; 61: 2578±85.
80 Hill H, Austin H. Nutrition and endometrial cancer. Cancer
Causes Control 1996; 7: 19±32.
81 Jain MG, Howe GR, Rohan TE. Nutritional factors and
endometrial cancer in Ontario, Canada. Cancer Control 2000;
7: 288±96.
82 McCann SE, Freudenheim JL, Marshall JR, Brasure JR,
Swanson MK, Graham S. Diet in the epidemiology of
endometrial cancer in western New York (United States).
Cancer Causes Control 2000; 11: 965±74.
83 Verhoeven DT, Goldbohm RA, van Poppel G, Verhagen H, van
den Brandt PA. Epidemiological studies on brassica vegetables
and cancer risk. Cancer Epidemiol Biomarkers Prev 1996; 5:
733±48.
84 Zhang SM, Hunter DJ, Rosner BA et al. Intakes of fruits,
vegetables, and related nutrients and the risk of non-
Hodgkin's lymphoma among women. Cancer Epidemiol Bio-
markers Prev 2000; 9: 477±85.
85 Ward MH, Zahm SH, Weisenburger DD et al. Dietary factors
and non-Hodgkin's lymphoma in Nebraska (United States).
Cancer Causes Control 1994; 5: 422±32.
ã 2001 Blackwell Science Ltd Journal of Internal Medicine 250: 280±290
86 Shklar G. Mechanisms of cancer inhibition by anti-oxidant
nutrients. Oral Oncol 1998; 34: 24±9.
87 Terry P, Lagergren J, Ye W, Nyren O, Wolk A. Antioxidants
and cancers of the esophagus and gastric cardia. Int J Cancer
2000; 87: 750±4.
88 Hansson LE, Nyren O, Bergstrom R et al. Diet and risk of
gastric cancer. A population-based case±control study in
Sweden. Int J Cancer 1993; 55: 181±9.
89 Willett WC. Goals for nutrition in the year 2000. CA Cancer J
Clin 1999; 49: 331±52.
90 Giovannucci E, Stampfer MJ, Colditz GA et al. Multivitamin
use, folate, and colon cancer in women in the Nurses' health
study. Ann Intern Med 1998; 129: 517±24.
91 Duthie SJ. Folic acid de®ciency and cancer: mechanisms of
DNA instability. Br Med Bull 1999; 55: 578±92.
92 Giovannucci E, Stampfer MJ, Colditz GA et al. Folate, methi-
onine, and alcohol intake and risk of colorectal adenoma.
J Natl Cancer Inst 1993; 85: 875±84.
93 Zhang S, Hunter DJ, Hankinson SE et al. A prospective study of
folate intake and the risk of breast cancer. Jama 1999; 281:
1632±7.
94 Blount BC, Mack MM, Wehr CM et al. Folate de®ciency causes
uracil misincorporation into human DNA and chromosome
breakage: implications for cancer and neuronal damage. Proc
Natl Acad Sci USA 1997; 94: 3290±5.
95 Naska A, Vasdekis VG, Trichopoulou A et al. Fruit and veget-
able availability among ten European countries: how does it
compare with the `®ve-a-day' recommendation? DAFNE I and
II projects of the European Commission. Br J Nutr 2000; 84:
549±56.
96 Subar AF, Heimendinger J, Patterson BH, Krebs-Smith SM,
Pivonka E, Kessler R. Fruit and vegetable intake in the United
States: the baseline survey of the Five a Day for better health
program. Am J Health Promot 1995; 9: 352±60.
97 Cox DN, Anderson AS, Reynolds J, McKellar S, Lean ME,
Mela DJ. Take Five, a nutrition education intervention to
increase fruit and vegetable intakes: impact on consumer
choice and nutrient intakes. Br J Nutr 1998; 80: 123±31.
98 Trudeau E, Kristal AR, Li S, Patterson RE. Demographic and
psychosocial predictors of fruit and vegetable intakes differ:
implications for dietary interventions. J Am Diet Assoc 1998;
98: 1412±7.
Received 20 April 2001; revision received 17 June 2001;
accepted 7 June 2001.
Correspondence: Paul Terry, Department of Medical Epidemiology,
Karolinska Institute, Box 281, 171 77 Stockholm, Sweden (fax:
+46-8-314957; e-mail: paul.terry@mep.ki.se).