See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/8421875

Resistant starch and colorectal neoplasia

Article  in  Journal of AOAC International · May 2004

Source: PubMed

CITATIONS READS
34 44

2 authors:

Graeme Young Richard Le Leu

Flinders University Royal Adelaide Hospital
482 PUBLICATIONS   11,928 CITATIONS    84 PUBLICATIONS   2,258 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Personal and Family History of Type 2 Diabetes and Colorectal Cancer Risk in Young Adults View project

Resistant starch and protection against colorectal cancer View project

All content following this page was uploaded by Richard Le Leu on 19 November 2015.

The user has requested enhancement of the downloaded file.

 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 775
SPECIAL GUEST EDITOR SECTION
Resistant Starch and Colorectal Neoplasia
GRAEME P. YOUNG and RICHARD K. LE LEU
Flinders University of South Australia, Department of Medicine, Bedford Park, Adelaide, South Australia, Australia 5042
There are several approaches to examining the
relationship between resistant starch (RS) and
development of colorectal cancer (CRC). These
include examination of epidemiological
relationships, objective testing of effects of RS
given to humans on biological events of relevance
to CRC, and studies in animal models where
protection and mechanisms of protection can be
directly tested. Nine epidemiological studies have
examined the relationship between starch and CRC
and/or adenomas. Most show a significant
protective effect. However, epidemiological tools
for measuring consumption of RS are poorly
developed and so a benefit for RS can only be
inferred. On balance, the magnitude of protection
by starch appears to be in the order of 25–50%.
Human intervention studies have examined the
effect of various types and amounts of RS
consumption on colonic biology. To generalize
from these studies, RS softens stools and
increases stool bulk, decreases pH, increases
short-chain fatty acids (SCFAs) including butyrate,
reduces products of protein fermentation, and
decreases bile salts in fecal water. Such changes
seem to be achieved within about 4 weeks of
commencing consumption. The greatest effects
are seen with the highest doses where increased
fecal starch recovery is observed. A modest
number of animal studies have been undertaken.
Those examining effects of RS on colonic biology
and biomarkers for CRC confirm and extend the
results in humans. RS modifies the lumenal
environment, largely through altered fermentation
of polysaccharides and proteins. RS also affects
epithelial biology in that it increases apoptotic
deletion of genetically damaged cells. More work is
needed to define what types and combinations of
RS, perhaps with probiotics, exert the greatest
effects on colonic environment and epithelial
biology, and then to test these in the cancer
models for their protective effect. A few studies
have examined effect of RS on cancer as an end
point in several rodent models, but the results are
Guest edited as a special report on "Novel Dietary Fibers: The
Importance of Carbohydrates in the Diet" by Barry McCleary and Ian
Brown.
Corresponding author's e-mail: graeme.young@flinders.edu.au.
not clear cut. In conclusion, consumption of RS
dramatically affects the colonic lumenal
environment and facilitates apoptotic deletion of
genetically damaged cells in the colon, several of
which are considered to be biomarkers associated
with risk for CRC. These effects can be interpreted
as reflecting improved colonic health, which might
be of benefit in protection against CRC. Direct
evidence for protection is still not available.
C
olorectal cancer (CRC) is not a simple biological
event. It is a process that progresses over time, is due
largely to acquired but occasionally inherited genetic
abnormalities (1), and is potentially able to be regulated at
various stages in its development. Certainly, it can be
regulated by a range of dietary factors that putatively act by a
large range of biological actions (2).
Addressing the relationship between resistant starch (RS)
and CRC development is further complicated by the range of
dietary forms of RS. Measuring RS consumption is difficult.
Furthermore, the mechanisms by which it might protect
against CRC are diverse.
This paper addresses the evidence from 3 perspectives:
epidemiological, animal intervention studies, and human
intervention studies. Some studies involve end points that are
directly informative (e.g., cancer as an end point); others are
indirectly informative (i.e., address effects of RS on possible
mechanisms of prevention).
Epidemiologic Studies
Of the many studies that have examined the relationship
between diet and CRC, the association with starch,
carbohydrate, or polysaccharide intake has been studied in a
minority of these (Table 1). Most of the studies were
conceived and/or conducted before we understood the nature
of RS. One is, therefore, left to infer what the role of RS might
be, as none has directly addressed it at an epidemiological
level.
Cassidy et al. (3) examined the relationship between starch
intake and CRC in an international correlative study (across
12 populations) where they included nonstarch
polysaccharides (NSP), fat, and protein in the analysis. Strong
inverse associations were found between starch intake and
both colon and rectal cancer (r = –0.7 and r = –0.47,
respectively). The effect remained statistically significant
776 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004
after adjusting for fat and protein intake. NSP itself was not
significantly associated with CRC incidence; a significant
correlation appeared when NSP was combined with starch.
Only one cohort study addresses the relationship between
CRC and starch consumption (4). The Health Professionals
Follow-Up Study of 170 U.S. males found that the highest
levels of starch intake were associated with a lower risk of
colorectal adenomas [relative risk (RR) 0.47; 95% confidence
interval (CI), 0.28–0.77]. Cases were self-reported adenomas
and, although efforts were made to minimize the chance of
adenomas in controls, some may have had them, hence, the
apparent benefit might be even greater.
In several case-control studies examining the relationship
between starch and CRC, quantitative data were accessible.
Case-control studies may suffer from recall bias, failure of
blinding of measurements, and inappropriate ascertainment of
control groups. As a general rule, such evidence is not
considered to be of the same quality as cohort studies. Two
such studies have not shown a significant benefit. One related
CRC to starch intake (5) and found an odds ratio (OR) of 0.98;
a second case relating CRC to polysaccharides (6) found an
OR of 0.83. One (7) found a significant reduction in risk for
polysaccharides with an energy-adjusted OR of 0.55 for
polysaccharides. All were large studies.
The relationship between high-carbohydrate intake and
adenomas has been examined in 4 case-control studies. Three
indicated that high carbohydrates appear to be
protective (8–10), although the first showed it for men but not
women and the latter showed this for women but not men. The
study by Sandler et al. (10) involved 236 subjects with
adenomas and 409 adenoma-free controls, and potential
biases were carefully managed. Comparing women in the
lowest and highest quintiles of carbohydrate consumption,
there was a 60% reduction (OR 0.39, 0.19–0.8). Men showed
the same trend (OR 0.48), although the difference was not
significant. Neugut et al. (11) found a nonsignificant trend to
protection in men and women. Although sugar contributed to
total carbohydrate intake in some of these studies, refined
sugars appeared to be associated with an increased risk (2),
implying that complex polysaccharides mediate protection.
Epidemiological studies are essentially correlative and so
suggest associations rather than demonstrate cause-effect
relationships. The observations generate hypotheses that need
to be tested and by several approaches: by examining
mechanistic questions, i.e., how might starch protect, or by
prospectively testing interventions. Of course, investigating
cancer as an end point in humans requires extensive resources,
and to date no study has been published on a planned dietary
intervention with starch or RS and its effect on CRC
occurrence in humans. Currently the Concerted Action Polyp
Prevention (CAPP) studies are in progress, which test the
efficacy of RS on CRC prevention. The CAPP-1 study is
using a randomized double blind factorial design controlled
trial of aspirin (600 mg daily) and RS [30 g as raw potato
starch-Hylon VII (1 + 1, w/w)/day] in suppressing colorectal
adenoma formation in young subjects with Familial
Adenomatous Polyposis (FAP). The CAPP-2 study is a
randomized double-blind factorial design controlled trial of
aspirin (600 mg daily) and RS [30 g as Novelose
260-Novelose 330 (1 + 1, w/w)/day] in gene carriers or
affected members of families with HNPCC (hereditary
nonpolyposis CRC). The intervention is for 2–4 years, with
outcome measures of colorectal adenomas and cancers.
Biopsies of the macroscopically normal rectal mucosa are also
taken for investigation of putative biomarkers of CRC risk,
including cell proliferation, apoptosis, and crypt fission.
Results will not be available for some years.
Although not all of the epidemiological studies show
significant protection, none shows an adverse effect and the
balance of the evidence favors protection by starch. Such
findings raise questions as to how starch might exert a
protective action.
Mechanism of Protection by Starch
The action of NSP in protection appears multifactorial,
including dilution of carcinogens, binding of carcinogens,
alteration of bile salt excretion profiles, and hastening of gut
transit (12, 13). Also much evidence indicates that NSPs
protect by the results and/or products of fermentation (3).
Undigested starch, i.e., RS, might act in the same way.
Butyrate, one of the products of polysaccharide
fermentation, was first suggested as a protective factor in
1981 (14) and was shown a few years after to suppress
proliferation of cancer cells in vitro (15) and subsequently to
induce apoptosis in cancer cells by inhibition of histone
deacetylase (16). Colonic production of butyrate by
fermentation is associated with tumor mass in an animal
model (17). Thus, measuring degree of fermentation and
production of butyrate in relation to consumption of RS is of
interest in exploring the relationship between RS and CRC.
Cassidy et al. (3) attempted to estimate the relationship
between CRC and RS by assuming that 5% of all consumed
starch was resistant. This estimate represents a significant
amount of fermentable substrate reaching the colon, as starch
intake averages 8–10 times higher than NSP intake (3). Not
surprisingly, they found a significant relationship between
CRC and RS based on this assumption. However, this
approach does not take into account the between-country
variation in dietary sources of starch, and hence a likely great
variation in the proportion of dietary starch that was resistant
to digestion. The actual amount of RS in the diet is difficult to
determine because published tables record only retrograded
amylose (18). Furthermore, precooking history, such as
growth conditions and nature of storage and different modes
of preparation of foods , has greatly differing effects on the RS
component. In other words, RS composition in diet varies
significantly and is not easily defined in the usual types of
food tables.
As a consequence, epidemiologic studies addressing the
relationship between CRC and RS (estimated from food
sources and food preparation types) have not been reported. It
is possible to address the question indirectly at an
epidemiologic level, however, by studying the associations
Table 1. Epidemiological studies in humans addressing relationship between starch intake and colorectal cancer or colorectal adenoma
Author Study type
Giovannucci et al. 1992 (4) Cohort; males in U.S. health
Professionals Follow-Up Study
Tuyns et al. 1987 (7) Case-control; Belgium
Haenszel et al. 1980 (5) Case-control; Japan
Macquart-Moulin et al. 1986 (6) Case-control; France
Sandler et al. 1993 (10) Case-control; U.S.
Macquart-Moulin et al. 1987 (9) Case-control; France
Hoff et al. 1986 (8) Case-control; Norway
Neugut et al. 1993 (11) Case-control; U.S.
Cassidy et al. 1994 (3) International correlational
in 12 populations
a
Result for top quartile or quintile of consumption; outcome measure given is as stated by the authors.
End point
Adenoma n = 170;
controls n = 7284
Colorectal cancer n = 818;
population controls n = 2851
Colorectal cancer n = 588; hospitalized
controls n = 1176
Colorectal cancer n = 399;
controls undergoing
rehabilitation n = 399
Adenoma n = 236; colonoscoped
controls n = 409
Adenoma n = 252;
hospitalized controls n = 238
Adenoma n = 23 >5 mm;
colonoscoped controls n = 77
Adenoma (recurrent) n = 286;
colonoscoped controls n = 480
Cancer incidence
by population
Variable Resulta
Carbohydrates OR 0.47 (0.28–0.77)
Polysaccharides adjusted OR 0.55 colon (p = 0.017);
for energy intake OR 0.53 rectum (p = 0.012)
Starch RR 0.98 (not significant)
Polysaccharides RR 0.83 (not significant)
Carbohydrates Women OR 0.39 (0.19–0.8);
men OR 0.48 (0.19–1.2)
Carbohydrates RR 0.33 (p = 0.008)
Carbohydrates Women 14 g less in cases (p = 0.27);
men 25 g less in cases (p = 0.01)
Carbohydrates Women OR 0.3 (0.1–1.3);
men OR 0.6 (0.2–1.7)
Starch consumption Colon r = –0.76, p < 0.001;
by population rectum r = –0.47, p < 0.05
YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 777
778 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004
between CRC and effects of RS on colonic function, biology,
or lumenal conditions. Where such effects serve as the
putative mechanisms by which RS protects, they may be
informative.
Walker et al. (19) examined the relationships between fiber
and fat intakes and fecal pH in several African populations.
They found the lowest pH in populations with the lowest
incidence of CRC. This study appears to be the only obvious
reassessment of fermentation-dependent effects in the colonic
lumenal environment at the population level. Although it does
not directly address the effect of RS on the colonic mucosa or
lumenal environment, it suggests that fermentation-dependent
effects are important.
To pursue additional evidence for protection by RS in
humans, we need to better understand how RS consumption
influences colonic epithelial biology and the colonic lumenal
environment. From a clear understanding of how RS affects
the colon, more informative epidemiological and intervention
studies need to be conducted in humans.
Human Intervention Studies with RS
To date numerous intervention studies have examined the
effect of RS on human colonic function (Table 2). These
intervention studies have evaluated not only different types of
RS but also combinations of types of RS and a range of
amounts of RS. Duration of feeding has also varied between
studies. Some of the studies have mimicked human diets in
that they have used a range of food-based sources, whereas
others are rather artificial, having used single manufactured
forms. Despite these differences, results from these studies
enable one to make a general statement on the effect of RS on
physiological processes in the colon. Insofar as these serve as
biomarkers for CRC risk, they also provide preliminary
evidence for impact on CRC development and the
mechanisms by which they might protect.
Effect of RS on Fecal Characteristics
Of the 7 studies that have examined the effect of RS on
fecal characteristics, 6 showed a significant increase in output
and/or bulk (20, 21, 23, 25–27; Table 2). This effect is
dose-dependent, and increasing doses are associated with
increasing recoverable starch in stools (21). Most reported a
stool softening effect. One study (28) showed no effect. That
study was conducted in subjects with recently removed
colonic adenomas, and the RS was administered in capsule
form. They also failed to observe any changes in fecal
fermentation, which suggests that the capsules did not
effectively release the RS.
RS appears to influence fermentation of other substrates
and may increase bacterial biomass. The fecal excretion of
NSPs was also increased with RS intake (21, 23), and this
would contribute to the increased fecal output associated with
RS consumption.
Fecal bacterial enzymes are possibly relevant to risk for
CRC because evidence shows that they might activate dietary
procarcinogens (31). In the only study that examined these
bacterial enzymes, consumption of a high RS diet
significantly decreased the activity of several such bacterial
enzymes in the feces (26). On balance, RS has a mild to
moderate effect on fecal bulk (increasing it), and it softens the
stools.
Effect of RS on Starch Fermentation Events
RS is subject to colonic fermentation, and RS consumption
alters many fermentation-dependent events (Table 2). Of the
studies that measured fecal pH, RS was shown to either
decrease pH (21, 22, 24) or have no effect (20, 25, 28). Of the
studies showing an effect, control diets were designed to be
low in RS and fermentable substrate, and the pH values
achieved were relatively low. Of those not showing an effect,
one used encapsulated RS, while the others included
fermentable substrates in the control diets that might have
masked the effect. On balance, RS can be expected to acidify
lumenal contents.
Most studies that examined the relationship between RS
intake and SCFA production/concentration have shown an
increase in SCFAs in feces or lumenal content, particularly
butyrate (20, 21, 23, 24, 27). Others have shown no
effect (25, 26, 28). The reasons given for lack of apparent
effect on pH also apply to these latter studies. Very short
periods of feeding and type or source of RS might also
contribute to the difference. Overall, consumption of RS can
be expected to increase lumenal concentrations of SCFAs
including butyrate; however, duration of consumption, dose,
and food source need more exploration. Phillips et al. (21)
observed highest levels of SCFAs, including butyrate, in those
consuming the most RS and who, in parallel, had the highest
levels of detectable starch in the feces. It would thus seem that
the greatest changes in feces are achieved by RS consumption
that increases fecal starch; but apart from increasing dose, we
do not understand the strategies needed to achieve this.
Effect of RS on Products of Protein Fermentation
RS has been shown to influence levels of by-products of
protein fermentation in the human colon. In the presence of
RS, lower concentrations of both ammonia and phenols in the
feces have been observed (22). This finding suggests either a
change in conditions that decrease likelihood of fermentation
of undigested/unabsorbed proteins, or else a change in
bacterial flora to those with lesser capacity to ferment protein.
Available studies are limited and provide no insight. As
protein fermentation products seem likely to be carcinogenic,
and if the observations by Birkett et al. (22) are born out, this
might be an additional mechanism by which RS might be
protective.
Effect of RS on Bile Acids and Cytotoxicity
Epidemiological studies have shown a positive correlation
between fecal bile acid concentration and CRC risk (32).
However, the situation is complex, as secondary and other
bacterially derived bile salts might be the most toxic. Of the
human intervention studies that have examined the
relationship between RS consumption and bile acids in feces

Table 2. Human intervention studies with resistant starch (RS)
Author
van Munster et al. 1994 (20)
Phillips et al. 1995 (21)
Birkett et al. 1996 (22)
Cummings et al. 1996 (23)
Noakes et al. 1996 (24)
Heijnen et al. 1998 (25)
YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 779
Intervention End point Outcome
14 healthy subjects fed 45 g native Fecal output ­ Fecal output
amylomaize starch for 4 weeks; ref. diet:
standardized low RS,
20 g natural fiber
pH No effect on pH
SCFA ­ Fecal SCFA output,
­ fecal butyrate excretion
Cell proliferation ¯ Cell proliferation
Bile acids ¯ Secondary bile acids in feces and fecal
water
Cytotoxicity ¯ Fecal water cytotoxicity
11 healthy subjects fed Hi-maize, cooked or ­ Fecal output (dose-dependent)
Fecal output
uncooked green banana flour for 3 weeks in
cross-over study with reference diet. RS
intake varied up to 45 g per day according
to energy intake; ref. diet: low-RS
pH ¯ pH
SCFA ­ Butyrate and acetate (dose-dependent)
Excretion of starch ­ Excretion of starch
(dose-dependent)
11 healthy subjects fed Hi-maize and green pH ¯ Fecal pH
banana flour for 3 weeks (39 g/day); ref.
diet: low-RS
Ammonia ¯ Fecal ammonia
Phenols ¯ Fecal phenols
12 healthy subjects fed controlled diets for Fecal output ­ Fecal weight
15-day periods as RS2 (potato and banana)
and RS3 (maize and wheat; 17–30 g
RS/day); ref. diet: RS-
free diet (wheat starch)
SCFA ­ Fecal total SCFA
NSP breakdown ¯ NSP breakdown and ­ fecal NSP
23 hypertyriglyceridemic subjects fed high- pH ¯ fecal pH
amylose maize starch
(17–25 g RS/day) for 4 weeks;
ref. diet: oat bran
SCFA ­ SCFA concn., ­ fecal butyrate
Bile acids ¯ Secondary bile acids in fecal water
24 healthy men fed uncooked Hylon VII Fecal output ­ Fecal output, no effect on dry weight
(RS2), retrograded high-amylose cornstarch
(RS2) as supplements for 1-week periods
(32 g RS per day calculated by in vitro
method); ref. diet: glucose supplement to
self-selected diet
pH No effect
SCFA No effect
Bile acids No effect in fecal water
Cytotoxicity No effect
780 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004

Table 2. (continued)
Author Intervention End point Outcome

Hylla et al. 1998 (26) 12 healthy subjects fed highly resistant Fecal ouput ­ Fecal output and dry weight
amylomaize starch (Hylon VII,
55 g RS/day) for 4 weeks; ref. diet:
cornstarch (low-RS)
SCFA No effect
Bacterial enzymes ¯ b-glucosidase activity
Bile acids ¯ Fecal concn. of total and secondary bile
acids
Sterols ¯ Concn. fecal total neutral sterols
Jenkins et al. 1998 (27) 24 healthy subjects fed diets for 2 weeks. Fecal bulk ­ Fecal bulk
RS derived from RS2 and RS3 sources (RS
equiv. to 30 g fiber); ref. diet: low-fiber
SCFA ­ Fecal butyrate
Grubben et al. 2001 (28) 23 patients with recently removed colonic Fecal output No effect
adenomas fed supplement of 45 g native pH No effect
amylomaize starch (as a capsule containing SCFA No effect
28 g RS) for 4 weeks; ref. diet: 45 g Cell proliferation No effect
maltodextrin supplement
Bile acids ¯ Primary and secondary bile acids in fecal
water
van Gorkom et al. 2002 (29) 111 sporadic adenoma patients fed 30 g Cell proliferation No effect on cell proliferation
highly resistant amylomaize starch Hylon VII
(19 g RS); ref. diet: controlled placebo
Wacker et al. 2002 (30) 12 subjects (healthy) fed highly resistant DNA adduct levels ­ DNA adduct levels
amylomaize starch (Hylon VII,
50–60 g RS/day) enriched to starchy foods
for 4 weeks, rectosigmoidal biopsy
taken at end of 4 weeks;
ref. diet: cornstarch (low-RS)
Cell proliferation No effect on cell proliferation

and fecal water, most have shown a reduced bile acid
concentration (20, 24, 26, 28). The concentration of soluble
bile acids in fecal water is thought to be more relevant to
colonic mucosal damage than that of total bile acids in feces,
as the soluble bile acids are those available for contact with the
colonic mucosa (33). Three studies found that bile acid
concentration in the fecal water was decreased by
RS (20, 24, 28) while one study found RS had no effect (25).
However, in the latter study (25), the RS diets were fed for a
period of only 1 week, which might not have been long
enough for any significant changes to be developed. In a study
by Hylla et al. (26), the concentrations of total neutral sterols
and 4-cholesten-3-one were shown to decrease with RS
intake; it is not clear whether these compounds play a role in
carcinogenesis, but 4-cholesten-3-one may act as a
co-carcinogen (34).
The cytotoxicity of the fecal water to colon cancer cell
lines has been examined in several studies in which RS was
fed to subjects. This probably reflects bile salt concentrations
in part. One study showed that RS decreased the cyototoxicity
of fecal water (20), whereas the other study, which used
encapsulated RS, showed no effect (25). On balance, RS
decreases bile acid concentration in fecal water and this would
be expected to be associated with reduced risk for CRC.
Effect of RS on Epithelial Biology
Epithelial proliferation is of interest in terms of
maintenance of epithelial mass but also a possible relationship
with CRC. Of the 4 studies that have examined the effect of
RS intake on colonic epithelial proliferation, 3 have shown no
effect (28–30) whereas one study showed a small
decrease (20). It is difficult to know exactly what this means in
biological terms. The control diets used in these studies were
not sufficient to induce epithelial atrophy of the degree seen in
animal models as some fermentable substrate was included.
Therefore, the lack of effect on proliferation does not exclude
a trophic effect as would be expected from a
butyrate-generating substrate.
From another perspective, the lack of a hyperproliferative
effect might be interpreted as beneficial because
 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 781
hyperproliferation is thought to be a risk factor for cancer
development (35). Several studies have shown an abnormally
high proliferation rate of the colonic epithelium in subjects
with adenomas or cancer in the colon (36–38). This
hyperproliferation might reflect a field change in the colon
associated with cancer development and so have a quite
different cause from a trophic effect of a fermentable
substrate.
Without carefully designed control diets, the effect of RS
on epithelial proliferation in humans cannot be determined.
Even if it were shown to be trophic, an effect to be expected
because of its impact on fermentation, the biological
relevance of this seems more likely to be beneficial than
harmful; such a hyperproliferative effect should not be
confused with the hyperproliferation seen in subjects with
coexistent colorectal neoplasia.
RS and Epithelial DNA
DNA adducts in the colonic mucosa are, in part, a result of
oxidative stress (39) and may enhance tumorigenesis. Higher
levels of DNA adducts have been detected in nontumoral
colonic mucosa of patients with adenocarcinoma compared
with tissues from healthy controls (40). In a single study (30),
higher levels of DNA adducts were found in the colonic
mucosa of healthy volunteers fed a high-RS diet when
compared with controls fed a low-RS diet. The implications
are unclear for CRC development.
Summary of Effects of RS in Humans
To generalize from these studies, RS softens stools and
increases stool bulk, decreases pH; increases SCFAs,
including butyrate; reduces products of protein fermentation;
and decreases bile salts in fecal water. Such changes seem to
be within approximately within 4 weeks of commencing
consumption. The greatest effects are seen with the highest
doses where increased fecal starch recovery is observed.
Although the effect on epithelial biology is unclear, these
effects can be interpreted as reflecting improved colonic
health. The implication is that RS will protect against CRC,
but these studies do not demonstrate a direct effect.
Animal Studies
The effect of RS has been tested in several different
experimental models. Certain studies allow direct evaluation
of RS on colorectal neoplasia, such as carcinogen-induced
CRC or preneoplastic aberrant crypt foci, whereas several
have used mouse strains with cancer-related genes that are
“knocked out." These studies were conducted in a controlled
environment and allowed examination of biomarkers for CRC
risk as well as some of the mechanisms whereby RS might act.
Other studies have simply examined the mechanisms by
which RS might act.
RS Effect on Colorectal Neoplasia
Table 3 summarizes the effect of RS on colorectal
neoplasia in several experimental animal models. Two studies
support a protective role of RS (41, 44). Both of these studies
used aberrant crypt foci (ACF), which are microscopically
identified precursor lesions of CRC. In one study (44),
retrograded high-amylose cornstarch, added at a level of
25 g/kg, diet significantly inhibited total ACF when fed
during the promotion phase of CRC. In the other protective
study (41), raw potato starch (67 g/kg diet), when introduced
after initiation, was shown to reduce total ACF and the larger
ACF, when compared to a basic sucrose or cornstarch diet.
Three animal studies have shown that RS has no apparent
effect on colorectal neoplasia (43, 45, 46). In one study (46),
retrograded high-amylose cornstarch (25 g/kg diet) had no
effect on ACF when fed for 4 weeks prior to and during
initiation with dimethylhydrazine (DMH). Another study (43)
found no effect of RS (3 or 10% by weight of high-amylose
maize starch hydrolyzed with pancreatin) on tumor incidence,
numbers, and size. Pierre et al. (45) found that retrograded
high-amylose cornstarch (18.8 g/kg diet) had no effect on
tumor incidence and numbers in Min mice. Min mice are
heterozygous for a non-sense mutation in the Apc gene, and
provide a model for both familial adenomatous polyposis and
sporadic colon cancer. However, in Min mice most of the
tumors arise in the small intestine, and therefore this model is
unable to explore effects of fermentable substrates in the colon
on tumors arising in the colon.
Two animal studies published to date have shown an
apparent promotional effect with RS (42, 47). Young et
al. (42) found that raw potato starch (14.4 g/kg diet) increased
the density of ACF and enhanced the size and number of
tumors when compared with a control diet containing no
added NSP or RS, although it should be noted that RS did not
increase the proportion of rats with tumors. Williamson et
al. (47) found that a mixture (1 + 1) of raw potato starch and
high-amylose starch (25 g/kg diet), when fed to Apc1638N
mice, resulted in significantly more small intestinal tumors;
however, no colorectal tumors were reported in that study.
Several explanations could clarify the varying results
among the reported studies: the different carcinogen protocols
(dose and duration), the different types of RS, and the
different feeding regimens used, thereby altering lumenal
conditions. The conditions in the colonic lumen, especially
butyrate and pH, have a major influence on colonic
oncogenesis (17, 48, 49). On balance, one cannot reach a firm
conclusion at this point, but an increase in incidence of
animals with CRC has not been demonstrated.
RS and Biomarkers for Colorectal Cancer
Table 4 summarizes the experimental animal studies that
have examined the effect of RS on epithelial and lumenal
biomarkers for CRC risk.
RS has been shown to have a significant effect on SCFA
production, including generating higher butyrate levels and/or
greater molar proportion of butyrate (43, 44, 50, 52–54, 56).
Increased SCFA production, in particular butyrate, may be
a mechanism for the protection against CRC (17, 49).
Butyrate, which is produced by anaerobic fermentation of RS
or NSP and other substrates in the colonic lumen, has been
782 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004

Table 3. Summary of animal studies examining the relationship between resistant starch (RS) and colorectal
neoplasiaa
Study Model RS type/reference diet End point Outcome

Thorup et al. 1995 (41)
Young et al. 1996 (42)
Sakamoto et al. 1996 (43)
Cassand et al. 1997 (44)
Wistar rats Raw potato starch (RS2) Aberrant crypt foci ¯ Total aberrant crypt foci
Azoxymethane (67 g/100 g diet) ¯ Larger aberrant crypt foci
ref. diet: NSP- and RS-free
Sprague-Dawley rats Raw potato starch (RS2) Tumor incidence No effect
Dimethylhydrazine (14.4 g/100 g diet) Tumor size ­
ref. diet: NSP- and RS-free Tumor multiplicity ­
Aberrant crypt foci ­ Density of aberrant crypt foci
Sprague-Dawley rats High amylose maize starch Tumor incidence No effect
hydrolyzed with pancreatin (RS3)
(3 or 10 g/100 g)
Dimethylhydrazine ref. diet: NSP- and RS-free Tumor size No effect
Tumor multiplicity No effect
Sprague-Dawley rats Retrograded high-amylose Aberrant crypt foci ¯ Total aberrant crypt foci

Pierre et al. 1997 (45)
Maziere et al. 1998 (46)
Williamson et al. 1999 (47)
Min mouse Retrograded high-amylose Tumor incidence No effect
cornstarch (RS3) Tumor multiplicity No effect
(18.8 g/100 g diet)
ref. diet: 2% cellulose/RS-free
Sprague-Dawley rats Retrograded high-amylose Aberrant crypt foci No effect
Dimethylhydrazine cornstarch (RS3); (25 g/100 g diet)
ref. diet: 2% cellulose/RS-free
Min mouse 1:1 Raw potato starch and high- Tumor incidence ­ Small intestinal tumors
amylose maize starch
(25 g/100 g diet)
ref. diet: NSP- and RS-free

a
RS = Resistant starch (type shown in parentheses); NSP = nonstarch polysaccharide; Ref. Diet – reference or control diet.

shown to inhibit cell proliferation, induce differentiation, and
enhance apoptosis in CRC cells in vitro (15, 16, 49, 57). In a
study by Le Leu et al. (56), higher doses of RS were associated
with an enhanced apoptotic response in the colon in response
to genetic damage caused by a chemical carcinogen. As the
acute apoptotic response to genotoxic carcinogens acts to
remove genetically damaged cells that might otherwise form
mutated clones that progress to malignancy, RS might protect
against the progression of mutated clones. Also observed with
this study was a significant negative correlation between the
acute apoptotic response to carcinogen and fecal butyrate
levels (56).
Expansion of mucosal cell proliferation has been
suggested as a risk factor for CRC (35). This issue is discussed
in more detail above. RS has been shown to have no obvious
effect on cell proliferation in the absence of any carcinogen
(52, 56), although another study has shown a slight increase in
cell proliferation in preneoplastic epithelium compared with a
no-fiber diet that caused mucosal atrophy (42).
RS consumption has been shown in some studies to
increase fecal output (42–44, 46, 51, 53) and decrease transit
time (53). Increased fecal bulk or reduced transit time may act
by diluting potential toxins and carcinogens and reducing
their contact time with the colonic epithelium (22, 58).
Consistently, RS has been shown to lower pH levels either
in cecal or fecal contents (44, 46, 50, 56). Lowered pH is
considered to be protective for CRC (59). A secondary effect
of lowered pH may be related to RS ability to increase the
excretion of fecal bile acids (51). Lowered pH may inhibit the
bacterial transformation of primary to secondary bile
acids (60). Secondary bile acids are cytotoxic to colonic cells
and are suggested to be colon tumor promoters (61).
RS appears to be beneficial to colonic bacteria (50, 52, 54)
and to possess prebiotic properties, thereby exerting
beneficial impacts within the colon. RS has also been shown
to lower the activity of certain bacterial enzymes (i.e.,
b-glucuronidase) in the lumen (46, 52). Bacterial enzymes
play an important role in the generation of toxic and
 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 783

Table 4. Summary of animal studies examining the relationship between resistant starch (RS) and biomarkers for
colorectal cancera
Study Model RS type/reference diet End point Outcome

Sakamoto et al. 1996 (43) Sprague-Dawley rats High-amylose maize starch Fecal output ­ Fecal output
Dimethylhydrazine hydrolyzed with pancreatin SCFA ­ Total SCFA, 3% RS­
(3 or 10% RS) butyrate
ref. diet: NSP/RS-free
Young et al. 1996 (42) Sprague-Dawley rats Raw potato starch (RS2 Fecal output ­ Fecal output
Dimethylhydrazine 14.4 g/kg diet) Cell proliferation ­ Cell proliferation
ref. diet: NSP-RS-free
Cassand et al. 1997 (44) Sprague-Dawley rats Retrograded high-amylose Fecal output ­ Fecal output
Dimethylhydrazine cornstarch (RS3 Cecal pH ¯ pH
25 g/kg diet)
ref. diet: 2% cellulose/RS- SCFA ­ Cecal/fecal SCFA,
free ­ butyrate

Kleesson et al. 1997 (50)
Wistar rats Raw potato starch (RS1), SCFA Microbial populations ­ Total SCFA, ­ butyrate
Retrograded potato for RS2
starch (RS2); 10 g/100 g
diet)
ref. diet: 3% cellulose/RS- ­ Desirable bacteria
free

Ebihara et al. 1998 (51) Wistar rats Chemically modified starch Fecal output ­ Fecal output
(CMS)

ref. diet: potato starch Cecal SCFA ¯ Butyrate with CMS

Bile acids
­ Fecal bile acids with CMS
Maziere et al. 1998 (46) Sprague-Dawley rats Retrograded high-amylose Fecal output ­ Fecal output
cornstarch (RS3;
25 g/kg diet)
Dimethylhydrazine ref. diet: 2% cellulose/RS- pH ¯ pH
free
Bacterial enzymes ¯ Cecal b-glucuronidase
Silvi et al. 1999 (52) Human flora-associated Retrograded amylose starch Cecal SCFA ­ Proportion of butyrate
(15 g/100 g diet)
Fisher rats ref. diet: 2.1% cellulose/RS- Microbial populations Desirable bacteria in cecum
free
Bacterial enzymes ¯ Cecal b-glucuronidase
Ammonia ¯ Cecal ammonia
Cell proliferation No effect on cell proliferation
Potato starch, high-amylose
Ferguson et al. 2000 (53) Wistar rats maize starch, and a-amylase Fecal output All RS ­ fecal output
treated Hi-maize
(35 g/100 g diet)
ref. diet: NSP- and RS-free
SCFA ­ Total rectal SCFA with
potato starch, ­ proportion
of butyrate

Transit time ¯ Transit for potato starch

and a-amylase treated
Hi-maize
784 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004

Table 4. (continued)

Study Model RS type/reference diet End point Outcome

Wang et al. 2002 (54) Balb/c mice Amylomaize starch and SCFA ­ Butyrate in feces
modified amylomaize starch;
(40 g/100 g diet)
ref. diet: 10% wheat Microbial populations ­ Desirable bacteria in colon
bran/RS-free
Ferguson et al. 2003 (55) Wistar rats Potato starch and high- Fecal output ­ Fecal output
amylose maize starch
(35 g/100 g diet)
ref. diet: NSP- and RS-free Excretion of food ­ Carcinogen bioavailability
carcinogen, IQ

Le Leu et al. 2003 (56) Sprague-Dawley rats High-amylose maize starch pH ¯ pH

a
RS = Resistant starch (type shown in parentheses); NSP = nonstarch polysaccharide; ref. diet = reference or control diet.

carcinogenic metabolites from dietary and endogenous
dietary substances and therefore can have implications for
colonic health. A negative finding for RS was observed in a
study by Ferguson et al. (55), who showed that RS can
increase the bioavailability of the food carcinogen,
2-amino-3-methylimidazo[4,5-f]quinoline (IQ). They implied
that this would more likely increase the initiatory events of
CRC; however, they did not show that this was actually the
case and no other studies testing this possibility have been
reported.
Overall, the effect of RS on biomarkers associated with
CRC risk suggests a protective action. RS modifies lumenal
environment, probably largely through altered fermentation.
RS also affects epithelial biology in that it appears to be
proapoptotic in vivo. More work is needed to define what
types and combinations of RS, perhaps with probiotics, exert
the greatest effects on colonic environment and epithelial
biology, and then to test these in the cancer models for their
protective effect.
Conclusions
Most epidemiological studies show a significant protective
effect of RS. However, epidemiological tools for measuring
consumption of RS are poorly developed; hence a benefit for
RS can only be inferred. On balance, the magnitude of
protection by starch appears to be in the order of 25–50%.
Human intervention studies have examined the effect of
various types and amounts of RS consumption on colonic
biology. To generalize from these studies, RS softens stools
and increases stool bulk, decreases pH, increases SCFAs
including butyrate, reduces products of protein fermentation,
and decreases bile salts in fecal water. The greatest effects are
seen with the highest doses where increased fecal starch
recovery is observed.
A modest number of animal studies have been undertaken.
Those examining effects of RS on colonic biology and
biomarkers for CRC confirm and extend the results in
humans. RS modifies the lumenal environment, largely
through altered fermentation of polysaccharides and proteins.
RS also affects epithelial biology in that it increases apoptotic
deletion of genetically damaged cells. More work is needed to
define what types and combinations of RS, perhaps with
probiotics, exert the greatest effects on colonic environment
and epithelial biology, and then to test these in the cancer
models for their protective effect. A few studies have
examined effect of RS on cancer as an end point in several
rodent models, but the results are not clear cut.
RS appears of great promise as part of a dietary approach to
reducing risk for CRC, but studies examining the relationship
more directly are required.
Acknowledgments
We acknowledge the ever-thoughtful stream of ideas
provided by Ian Brown and the grants from The Cancer
Council South Australia, NHMRC, Australia. Industry
funding for research has also been provided by Penford, Inc.,
Denver, CO, and National Starch and Chemical Co.,
Bridgewater, NJ.
References
(1) Young, G.P., Rozen, P., & Levin, B. (2002) in Colorectal
Cancer in Clinical Practice: Prevention and Early Detection,
Isis Publications, London, UK, pp 23–38
(2) World Cancer Research Fund (1997) Cancers: Colon,
Rectum, in Food, Nutrition and the Prevention of Cancer: A
 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004 785
Global Perspective, American Institute for Cancer Research,
Washington, DC, pp 216–251
(3) Cassidy, A., Bingham, S.A., & Cummings, J.H. (1994) Br. J.
Cancer 69, 937–942
(4) Giovannucci, E., Stampfer, M.J., Colditz, G., Rimm, E.B., &
Willett, W.C. (1992) J. Nat. Cancer Inst. 84, 91–98
(5) Haenszel, W., Locke, F.B., & Segi, M. (1980) J. Nat. Cancer
Inst. 64, 17–22
(6) Macquart-Moulin, G., Riboli, E., Cornee, J., Charnay, B.,
Berthezene, P., & Day, N. (1986) Int. J. Cancer 38, 183–191
(7) Tuyns, A.J., Haelterman, M., & Kaaks, R. (1987) Nutr.
Cancer 10, 181–196
(8) Hoff, G., Moen, I.E., Trygg, K., Frolich, W., Sauar, J., Vatn,
M., Gjone, E., & Larsen, S. (1986) Scand. J. Gastroenterol.
21, 199–204
(9) Macquart-Moulin, G., Riboli, E., Cornee, J., Kaaks, R., &
Berthezene, P. (1987) Int. J. Cancer 40, 179–188
(10) Sandler, R.S., Lyles, C.M., Peipins, L.A., McAuliffe, C.A.,
Woosley, J.T., & Kupper, L.L. (1993) J. Nat. Cancer Inst. 85,
884–891
(11) Neugut, A.I., Garbowski, G.C., Lee, W.C., Murray, T.,
Nieves, J.W., Forde, K.A., Treat, M.R., Waye, J.D., &
Fenoglio Preiser, C. (1993) Ann. Int. Med. 118, 91–95
(12) Kritchevsky, D. (1995) Eur. J. Cancer Prev. 4, 345–352
(13) Stephen, A.M., & Cumming, J.H. (1980) Nature 284,
283–284
(14) Cummings, J.H. (1981) Proc. Nutr. Soc. 40, 7–14
(15) Whitehead, R.H., Young, G.P., & Bhathal, P.S. (1987) Gut 27,
1457–1463
(16) Medina, V., Edmonds, B., Young, G.P., James, R., Appleton,
S., & Zalewski, P.D. (1997) Cancer Res. 57, 3697–3707
(17) McIntyre, A., Gibson, P.R., & Young, G.P. (1993) Gut 34,
386–391
(18) Holland, B., Unwin, I.D., & Buss, D.H. (1988) Cereals and
Cereal Products, Third supplement to The Composition of
Foods, 4th Ed., Royal Society of Chemistry, Letchworth,
Herts, UK
(19) Walker, A.R.P., Walker, B.F., & Walker, A.J. (1986) Br. J.
Cancer 53, 489–495
(20) van Munster, I.P., Tangerman, A., & Nagengast, F.M. (1994)
Dig. Dis. Sci. 39, 834–842
(21) Phillips, J., Muir, J.G., Birkett, A., Lu, Z.X., Jones, G.P.,
O’Dea, K., & Young, G.P. (1995) Am. J. Clin. Nutr. 62,
121–130
(22) Birkett, A., Muir, J., Phillips, J., Jones, G., & O’Dea, K.
(1996) Am. J. Clin. Nutr. 63, 766–772
(23) Cummings, J.H., Beatty, E.R., Kingman, S.M., Bingham,
S.A., & Englyst, H.N. (1996) Br. J. Nutr. 75, 733–747
(24) Noakes, M., Clifton, P.M., Nestel, P.J., Le Leu, R., &
McIntosh, G. (1996) Am. J. Clin. Nutr. 64, 944–951
(25) Heijnen, M.L., van Amelsvoort, J.M., Deurenberg, P., &
Beynen, A.C. (1998) Am. J. Clin. Nutr. 67, 322–331
(26) Hylla, S., Gostner, A., Dusel, G., Anger, H., Bartram, H.P.,
Christl, S.U., Kasper, H., & Scheppach, W. (1998) Am. J.
Clin. Nutr. 67, 136–142
(27) Jenkins, D.J., Vuksan, V., Kendall, C.W., Wursch, P.,
Jeffcoat, R., Waring, S., Mehling, C.C., Vidgen, E., Augustin,
L.S., & Wong, E. (1998) J. Am. Coll. Nutr. 17, 609–616
(28) Grubben, M.J., van den Braak, C.C., Essenberg, M., Olthof,
M., Tangerman, A., Katan, M.B., & Nagengast, F.M. (2001)
Dig. Dis. Sci. 46, 750–756
(29) van Gorkom, B.A., Karrenbeld, A., van der Sluis, T., Zwart,
N., van der Meer, R., de Vries, E.G., & Kleibeuker, J.H.
(2002) Nutr. Cancer 43, 31–38
(30) Wacker, M., Wanek, P., Eder, E., Hylla, S., Gostner, A., &
Scheppach, W. (2002) Cancer Epidemiol. Biomarkers Prev.
11, 915–920
(31) Mallett, A.K., Rowland, I.R., & Fathing, M.S. (1987) Scand.
J. Gastrenterol. Suppl. 129, 251–257
(32) Hill, M.J. (1991) Eur. J. Cancer Prev. Suppl. 1, 69–72
(33) Rafter, J.J., Eng, V.W., Furrer, R., Medline, A., & Bruce,
W.R. (1986) Gut 27, 1320–1329
(34) Suzuki, K., Bruce, W.R., Baptista, J., Furrer, R., Vaughan,
D.J., & Krepinsky, J.J. (1986) Cancer Lett. 33, 307–316
(35) Preston-Martin, S., Pike, M.C., Ross, R.K., Jones, P.A., &
Henderson, B.R. (1990) Cancer Res. 50, 7415–7421
(36) Terpstra, O.T., Van Blankenstein, M., Dees, J., & Eilers,
G.A.M. (1987) Gastroenterology 92, 704–708
(37) Ponz de Leon, M., Roncucci, I., Di Donato, P., Tassi, L.,
Smerieri, O., Amorico, M.G., Malagoli, G., De Maria, D.,
Antonioli, A., & Chahin, N.J. (1988) Cancer Res. 48,
4121–4126
(38) Welberg, J.W., de Vries, E.G., Hardonk, M.J., Mulder, N.H.,
Harms, G., Grond, J., Zwart, N., Koudstaal, J., de Ley, L., &
Kleibeuker, J.H. (1990) J. Clin. Pathol. 43, 453–456
(39) Chung, F.L., Nath, R.G., Nagao, M., Nishikawa, A., Zhou,
G.D., & Randerath, K. (1999) Mutat., Res. 424, 71–81
(40) Pfohl-Leszkowicz, A., Grosse, Y., Carriere, V., Cugnenc,
P.H., Berger, A., Carnot, F., Beaune, P., & de Waziers, I.
(1995) Cancer Res. 55, 5611–5616
(41) Thorup, I., Meyer, O., & Kristiansen, E. (1995) Anticancer
Res. 15, 2101–2105
(42) Young, G.P., McIntyre, A., Albert, V., Folino, M., Muir, J.G.,
& Gibson, P.R. (1996) Gastroenterology 110, 508–514
(43) Sakamoto, J., Nakaji, S., Sugawara, K., Iwane, S., &
Munakata, A. (1996) Gastroenterology 110, 116–120
(44) Cassand, P., Maziere, S., Champ, M., Meflah, K., Bornet, F.,
& Narbonne, J. (1997) Nutr. Cancer 27, 53–59
(45) Pierre, F., Perrin, P., Champ, M., Bornet, F., Meflah, K., &
Menanteau, J. (1997) Cancer Res. 57, 225–228
(46) Maziere, S., Meflah, K., Tavan, E., Champ, M., Narbonne,
J.F., & Cassand, P. (1998) Nutr. Cancer 31, 168–177
(47) Williamson, S.L., Kartheuser, A., Coaker, J., Kooshkghazi,
M.D., Fodde, R., Burn, J., & Mathers, J.C. (1999)
Carcinogenesis 20, 805–810
(48) Cummings, J.H. (1997) The Large Intestine in Nutrition and
Disease, Institut Danone, Brussels, Belgium, pp 61–63
(49) Young, G.P., & Gibson, P.R. (1995) in Physiological and
Clinical Aspects of Short-Chain Fatty Acids, J.H.
Cumminigs, J.L. Rombeau, & T. Sakata (Eds), Cambridge
University Press, Cambridge, UK, pp 319–335
(50) Kleessen, B., Stoof, G., Proll, J., Schmiedl, D., Noack, J., &
Blaut, M. (1997) J. Anim. Sci. 75, 2453–2462
(51) Ebihara, K., Shiraishi, R., & Okuma, K. (1998) J. Nutr. 128,
848–854
(52) Silvi, S., Rumney, C.J., Cresci, I., & Rowland, I.R. (1999) J.
Appl. Microbiol. 86, 521–530
786 YOUNG & LE LEU: JOURNAL OF AOAC INTERNATIONAL VOL. 87, NO. 3, 2004

(53) Ferguson, L.R., Tasman-Jones, C., Englyst, H., & Harris, P.J.
(2000) Nutr. Cancer 36, 230–237
(54) Wang, X., Brown, I.L., Khaled, D., Mahoney, M.C., Evans,
A.J., & Conway, P.L. (2002) J. Appl. Microbiol. 93, 390–397
(55) Ferguson, L.R., Zhu, S., & Kestell, P. (2003) Food Chem.
Toxicol. 41, 785–792
(56) Le Leu, R.K., Brown, I.L., Hu, Y., & Young, G.P. (2003)
Carcinogenesis 24, 1347–1352
(57) Heerdt, B.G., Houston, M.A., & Augenlicht, L.H. (1997) Cell
Growth Differ. 8, 523–532
(58) Cummings, J.H., Bingham, S.A., Heaton, K.W., & Eastwood,
M.A. (1992) Gastroenterology 103, 1783–1789
(59) Thorton, J.R. (1981) Lancet 16, 1081–1083
(60) Nagengast, F.M., Hectors, M.P., Buys, W.A., & van
Tongeren, J.H. (1988) Eur. J. Clin. Invest. 18, 56–61
(61) Summerton, J., Goeting, N., Trotter, G.A., & Taylor, I. (1985)
Digestion 31, 77–81

View publication stats