Nutrition, Metabolism & Cardiovascular Diseases (2017) 27, 18e31

Available online at www.sciencedirect.com

Nutrition, Metabolism & Cardiovascular Diseases

journal homepage: www.elsevier.com/locate/nmcd

REVIEW

Dietary glycemic index, glycemic load and cancer: An overview of

the literature
S. Sieri, V. Krogh*
Epidemiology and Prevention Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy

Received 6 May 2016; received in revised form 16 September 2016; accepted 30 September 2016
Available online 10 October 2016

KEYWORDS Abstract Aims: The aim of this paper is to provide an overview of the current evidence for as-
Glycemic index; sociations between dietary glycemic index (GI) and dietary glycemic load (GL), and the risk of
Glycemic load; various types of cancer, and to summarize mechanisms proposed to explain the associations
Cancer risk found.
Data synthesis: Medline was searched for cohort studies, case-control studies, and meta-analyses,
published up to February 2016, that examined associations between dietary GI/GL and cancer.
Findings from the main meta-analyses showed a weak-to-moderate association of high dietary
GI/GL with increased risk of some cancers. High dietary GI but not GL was significantly and
consistently associated with increased colorectal cancer risk in both cohort and case-control
studies. Dietary GL was directly associated with breast and endometrial cancer risk in cohort
studies. Positive associations between dietary GI or GL and cancer risk were found more
frequently in case-control studies than cohort studies. The main mechanism for these associa-
tions is thought to be chronic hyperinsulinemia. Insulin is itself a mitogen and also increases
the bioactivity of insulin-like growth factors which can promote cancer by inhibiting apoptosis
and stimulating cell proliferation.
Conclusions: The review has uncovered consistent evidence that high dietary GI is associated
with increased risk of colorectal cancer, and that high dietary GL is associated with increased risk
of breast and endometrial cancer. However the risk increases are small or moderate.
ª 2016 The Italian Society of Diabetology, the Italian Society for the Study of Atherosclerosis, the
Italian Society of Human Nutrition, and the Department of Clinical Medicine and Surgery, Feder-
ico II University. Published by Elsevier B.V. All rights reserved.

Introduction disease in general, and coronary heart disease (CHD) in

The consumption of carbohydrates has increased in recent
years in western countries following advice to consume
less fat [1], especially less saturated fat, as it was thought
that this would lower the risk of developing cardiovascular
* Corresponding author. Epidemiology and Prevention Unit,
Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, 20133
Milan, Italy. Fax: þ39 02 2390 3510.
E-mail address: vittorio.krogh@istitutotumori.mi.it (V. Krogh).
particular [2]. However it is now clear that replacing
saturated fat by sugars or refined starch does not reduce
CHD risk and may increase it [3]; by contrast replacing fat
with carbohydrates from whole fruits, vegetables, pulses,
and whole grains may decrease CHD risk [4].
In general, consumption of carbohydrate increases
glycemia and blood insulin, but to varying extents,
depending on carbohydrate type (including proportion of
digestion-resistant material that serves as dietary fiber),
amount, processing method, and presence of other nutri-
ents. These variations are captured by the glycemic index

http://dx.doi.org/10.1016/j.numecd.2016.09.014
0939-4753/ª 2016 The Italian Society of Diabetology, the Italian Society for the Study of Atherosclerosis, the Italian Society of Human Nutrition, and the Department of
Clinical Medicine and Surgery, Federico II University. Published by Elsevier B.V. All rights reserved.
Dietary glycemic index, glycemic load and cancer
(GI), a system that ranks carbohydrate foods according to
their ability to raise blood glucose levels [5]. Glycemic load
(GL), the product of a food’s GI and its available carbohy-
drate content, was introduced to incorporate the effect of
the total amount of carbohydrate consumed. Thus, dietary
GI reflects the overall ability of consumed carbohydrates to
raise blood glucose (a measure of carbohydrate quality),
dietary GL is the sum of the GLs for all carbohydrate-
containing foods consumed, usually on a daily basis, and
reflects both the quantity and the quality of consumed
carbohydrates.
Factors influencing glucose metabolism may play
important roles not only in the development cardiovascular
disease and diabetes, but also in other chronic diseases
including cancer [6]. In particular, chronically elevated
blood insulin e as a result of chronically high blood glucose
e is thought to be the underlying mechanism for GI/GL-
related increases in the risks of several cancers [7]. Insulin
acts as a growth factor and also increases the bioactivity of
insulin-like growth factors (IGFs), such as IGF-1, by
enhancing their synthesis and decreasing their binding
proteins. IGF-1 can promote tumor development by inhib-
iting apoptosis, and stimulating cell proliferation and sex-
steroid synthesis [7]. IGF-1 has also angiogenic proprieties
[8]. Other conditions associated with chronically high blood
glucose, such insulin resistance, obesity, and diabetes may
also influence cancer risk [6] and it is likely that both the
quantity and the type of carbohydrate consumed can in-
fluence the risk of these diseases.
Topic of review
We examined epidemiological studies on associations be-
tween high GI/GL diets and cancer risk, assessing all cancer
sites for which data are available, and also summarizing
the mechanisms proposed to mediate the associations
found. The English language literature up to February 2016
was examined, providing an important update to previous
reviews.
Methods
We searched MEDLINE using the search terms glycemic
index or glycemic load combined with cancer, breast, colo-
rectal, lung, kidney, prostate, endometrial, ovarian, esopha-
geal, stomach bladder, liver and pancreatic. We also
searched reference lists of retrieved studies. We identified
600 publications. Review team members screened
retrieved titles and abstracts, eliminating publications
such as editorials, single case reports, cross-sectional
studies, and articles on treatment. The full texts of appar-
ently relevant articles were obtained. Eligible articles
were: English language only, case-control studies, cohort
studies, systematic reviews, and meta-analyses, published
from 1990 to February 2016, that investigated associations
of GI or GL with cancer in general, or one or more cancer
sites, in adults only. We identified 122 such articles, but
because of a citation limit only references to meta-
19
analyses, or original studies not cited in meta-analyses,
are included. Information on mechanisms was initially
retrieved from eligible articles, but other papers (not
referenced) were subsequently examined.
Cancer in general
We found a meta-analysis [9] that evaluated 36 prospec-
tive cohort studies, involving 60 811 patients considered to
have a ‘diabetes-related cancer’ (DRC) according to a
consensus report from the American Diabetes Association
and the American Cancer Society [10]. For all DRCs com-
bined (most but not all of the 36 studies included) a
modest-to-weak association between a diet inducing a
high blood glucose response (high dietary GI) and risk of
developing a DRC was found. A high GL diet was not
associated with cancer risk. The association between GI
and DRC risk was more apparent for North American
studies than those performed elsewhere, and was also
evident in women [9].
We found one cohort study that examined associations
between GI and GL and overall cancer risk [11]. High GI
was associated with increased risk of total cancer among
men but not women, while high GL was associated with
decreased risk of total cancer in both men and women.
Subgroup analyses found that the increased all-cancer risk
associated with high GI was confined to men and women
with high body mass index; and the reduced all-cancer
risk associated with high GL was confined to men and
women with low body mass index [11].
Breast cancer
We found three meta-analyses exclusively concerned with
GI/GL and breast cancer (BC) risk [12e14]. The first, pub-
lished in 2008, examined 6 cohort studies, and found no
association between GI/GL and risk of either premeno-
pausal or postmenopausal BC. However high dietary GI
was significantly associated with BC risk when only studies
that used a food frequency questionnaire with over 100
items (considered more robust) were included [12]. The
second meta-analysis, published in 2011 [13] was per-
formed on prospective studies only (10 cohort studies) and
found that high dietary GI was associated with signifi-
cantly increased risk of BC, with no significant association
between BC risk and GL. The most recent (2016) meta-
analysis [14], which included new prospective studies,
and examined 20 973 BC cases from 12 cohort studies,
found small increases in BC risk (5% and 6% respectively)
associated with highest dietary GI and GL compared to
lowest. Associations between dietary GI and GL and BC
were not significantly influenced by menopausal status
[13,14] or geographic region [13].
We found three other meta-analyses concerned with
GI/GL and various types of cancer, including BC [9,15,16],
published in 2008 [15], 2012 [9] and 2015 [16]. One re-
ported that high GL [15] and another that high GI [9] were
20
associated with increased BC risk (Table 1). Risk factors for
BC are known to vary with cancer hormone receptor
expression [17]. However there are few data pertaining to
associations of dietary carbohydrate, GI, and GL with
receptor-defined BC. The meta-analyses discussed above
[13,14] found that estrogen receptor status did not influ-
ence the association of BC risk with dietary GI and GL. A
study from the European Prospective Investigation into
Cancer and Nutrition (EPIC), which was not included in
previous meta-analyses, reported that a high GL diet was
directly associated with increased risks of ER and of ER/
PR BC in postmenopausal women [18].
Dietary GI and GL have been proposed to play a role in
BC etiology by acting through the IGF-1 axis [6]. High
glycemic diets are in fact generally associated with high
levels of blood insulin. And insulin may influence cancer
development by altering sex hormone metabolism and
increasing IGF-1 bioactivity (partly by lowering levels of
IGF binding protein). Furthermore, the adverse effects of a
high GI/GL diet on insulin sensitivity could be exacerbated
by obesity, which is an important determinant of insulin
resistance [19]. However the mechanisms of the differing
associations of BC defined by ER and PR status with high
glycemic diet are unclear. It is noteworthy that a 2010
pooled analysis found that IGF-I was significantly associ-
ated with ERþ cancer [20], and a 2009 study on a
population-based cohort of 61 433 women found
increased risk of ERþ/PR cancer, but not ERþ/PRþ or
ER/PR cancer with high GI and GL [21].
Another possible mediator of the relation between high
dietary GI/GL and BC is adiponectin. Adiponectin is an
adipocyte-produced insulin sensitizer whose levels are
lowered by high GI and GL diets, at least in diabetic men
[22]. Furthermore lowered adiponectin levels are associ-
ated with increased risk of postmenopausal BC, irre-
spective of levels of IGF-1 and IGF-3 [23].
Colorectal cancer
A meta-analysis of studies on colorectal cancer published
in 2009 [24], assessed case-control and cohort studies;
another [25], published in 2012, considered only pro-
spective cohort studies. Neither of these analyses found
any evidence that dietary GI/GL and colorectal cancer were
associated. However a 2015 meta-analysis of 15 studies (10
cohort studies and 5 case-control studies) [16], a 2012
meta-analysis of 10 studies (6 cohort and 4 case-control
studies) [26], and a 2012 meta-analysis that considered 9
prospective cohort studies [9] reported that high GI, but
not GL, was associated with increased risk of colorectal or
colon cancer. A single meta-analysis, published in 2008
reported an association between high GL and colorectal
cancer [15] (Table 2).
As with BC, it has been proposed that a high GI diet may
increase colorectal cancer risk via effects on insulin and
IGF-1. Insulin influences short-term metabolism, while IGF
axis exerts a longer-term integrating effect on growth [27].
Insulin stimulates cell proliferation by direct activation of
insulin/IGF receptors, or via inhibition of IGF binding
S. Sieri, V. Krogh
proteins, which result in increased levels of bioavailable
IGF-1. Epidemiological studies on circulating insulin e as
reflected by plasma C-peptide and IGF-1 e have found that
inferred high insulin is associated with significantly
increased risk of developing colorectal cancer [27e30]. As
with BC, obesity e a condition associated with chronically
elevated blood insulin e could be involved in the positive
association between high GI/GL and colorectal cancer risk
by increasing insulin resistance [31].
Another possible mechanism by which a high GI diet
may influence colorectal cancer risk is via an effect on
dietary fiber. A high GI diet is typically low in resistant
starch (which escapes digestion in the small intestine),
while many low-medium GI foods (e.g. wholemeal cereals,
pulses, and pasta) have high levels of resistant starch.
Resistant starches have been associated with reduced
postprandial insulin levels and reduced risk of type 2
diabetes. Type 2 diabetes has in turn been related to
increased colon cancer risk [32]. Furthermore, resistant
starches are extensively fermented by large intestine flora
resulting, among other things, in increased butyrate levels.
Butyrate plays an important role in maintaining intestinal
homeostasis [33]; it also appears to induce the selective
apoptosis of colon cancer cells [33].
Endometrial cancer
Several meta-analyses have found that a high GL diet is
associated with increased risk of endometrial cancer
[9,16,34,35]. The risk has been reported as greater in obese
women [35]. Dietary GI was unrelated to risk of endome-
trial cancer in these studies, although a 2008 meta-
analysis [15]found that both dietary GL and GI were
significantly associated with increased risk of endometrial
cancer (Table 3).
Endometrial cancer mainly affects postmenopausal
women in industrialized countries [36]. Forty percent of
endometrial cancers have been estimated to be related to
obesity; while excessive estrogen exposure, high blood
pressure, and diabetes may also be associated with
increased risk of this cancer [36].
High carbohydrate intake may be associated with
increased risk of endometrial cancer because such a diet is
often associated with obesity which increases estrogen
levels. Estrogen is a potent mitogen for the endometrium
[37]and is strongly associated with endometrial cancer.
The increased levels of insulin and increased bioavail-
ability of IGF-1 associated with high carbohydrate intake
may also promote endometrial cancer [6].
Ovarian cancer
A high GI diet (and consequent hyperinsulinemia) has
been hypothesized to increase ovarian cancer risk by
lowering insulin growth factor binding protein, to thereby
increase IGF-1 bioavailability which in turn stimulates cell
proliferation and sex hormone synthesis [38]. However
2008 [35] and 2015 [16] meta-analyses e the latter of 5
studies (2 cohort and 3 case-control) involving 3578 cases
 Dietary glycemic index, glycemic load and cancer
Table 1 Characteristics of studies and meta-analyses on dietary glycemic index (GI) or glycemic load (GL) and risk of breast cancer.

Author, year (sex) Study name Study location and Age at baseline Cases (n) Cohort size/controls (n)h Comparison GI RR (95%CI) GL RR (95%CI)
follow-up (years) (years)
Prospective studies
Jonas et al., 2003 (F)1,2,3,4,5,6 CPSII USA (5) 40e87 1442 63 307 High vs. low 1.03 (0.87e1.22) 0.90 (0.78e1.08)
Cho et al., 2003 (F)1,2,3,4,5,6 NHS II USA (8) 26e46 714 90 655 High vs. low 1.05 (0.83e1.33) 1.06 (0.78e1.45)
Holmes et al., 2004 (F)1,2,3,4,5,6 NHS USA (18) 34e59 4092 88 678 High vs. low 1.02 (0.82e1.28) 0.87 (0.70e1.12)
Frazier et al., 2004 (F)2,6 NHS II USA (8) 25e42 838 47 355 High vs. low 1.47 (1.04e2.08) 1.23 (0.91e1.67)
Higginbotham et al., 2004 (F)1,2,4,5 WHS USA (6.8) >45 946 39 876 High vs. low 1.03 (0.84e1.28) 1.01 (0.76e1.35)
Nielsen et al., 2005 (F)1,5,6 DDCHS Denmark (6.6) 50e65 634 23 870 Continuous 0.94 (0.80e1.10)a 1.04 (0.90e1.19)b
Silvera et al., 2005 (F)1,2,3,4,5,6 NBSS Canada (16.6) 40e59 1461 49 613 High vs. low 0.88 (0.63e1.22) 0.95 (0.79e1.14)
Giles et al., 2006 (F)6 MCCS Australia (9.1) 40e69 324 12 273 Continuousc 0.98 (0.88e1.10) 1.19 (0.93e1.52)
Sieri et al., 2007 (F)1,2,3,4,5,6 ORDET Italy (11.5) 34e70 289 8926 High vs. low 1.57 (1.04e2.36) 2.53 (1.54e4.16)
Lajous et al., 2008 (F)3,4,5,6 MGEN France (9) 42e72 1812 62 739 High vs. low 1.14 (0.99e1.32) 1.11 (0.96e1.29)
George et al., 2009 (F)4,5 NIH-AARP USA (1995e2003) 50e71 235 183 535 High vs. low 1.05 (0.97e1.15) 0.96 (0.81e1.12)
Larsson et al., 2009 (F)3,4,5,6 SMC Sweden (17.4) 40e74 2952 61 433 High vs. low 1.08 (0.96e1.21) 1.13 (1.00e1.29)
Wen et al., 2009 (F)3,4,5,6 SWHS China (7.4) 40e70 616 74 942 High vs. low 1.03 (0.79e1.34) 1.07 (0.82e1.39)
Linos et al., 2010 (F)3 NHS II USA (7.8) 34e53 455 39 268 High vs. low 1.18 (0.88e1.58) 0.89 (0.66e1.20)
Shikany et al., 2011 (F)4,5,6 WHI USA (8) 50e79 6115 148 767 High vs. low 1.01 (0.91e1.12) 1.08 (0.92e1.29)
Romieu et al., 2012 (F)5 EPIC Europe (11.5) 34e66 11 576 334 849 High vs. low 1.05 (0.99e1.12) 1.07 (1.00e1.14)
Case-control studies
Augustin et al., 2001 (F)1,5 N/A Italy 23e74 2569 2588 High vs. low 1.36 (1.14e1.64) 1.34 (1.10e1.61)
Levi et al., 2002 (F)1,5 N/A Switzerland 35e74 331 1145 High vs. low 1.60 (1.05e2.43) N/A
Lajous et al., 2005(F)1,5 N/A Mexico 35e69 475 1391 High vs. low 0.84 (0.62e1.15) 1.62 (1.13e2.32)
McCann et al., 2007 (F)1,5 WEB USA 35e79 1166 2105 High vs. lowe 1.02 (0.68e1.53) 1.01 (0.60e1.72)
WEB USA 35e79 1166 2105 High vs. lowf 0.80 (0.61e1.03) 0.74 (0.53e1.03)
Yun et al., 2010 (F)5 N/A South Korea 30e65 362 362 High vs. low 0.44 (0.23e0.85) 0.85 (0.48e1.50)
Hu et al., 2013 (F)5 NECSS Canada 20e76 2362 2492 High vs. low 0.94 (0.78e1.12) 0.83 (0.61e1.12)
Woo et al., 2013 (F)5 N/A Korea N/A 357 357 High vs. low 2.50 (1.46e4.31) 3.27 (1.94e5.50)
Amadou et al., 2015 (F) N/A Mexico 35e69 1000 1074 High vs. low 0.90 (0.68e1.12) 1.10 (0.82e1.10)
Author, year Medline search N studies Cases (n) Study size Studies assessed Comparison GI RR (95%CI) GL RR (95%CI)
Meta-analyses
Gnagnarella et al. 20081 Up to Oct 2007 11 14 119 370 113 Cohort þ case- High vs. low 1.08 (0.96e1.22) 1.14 (1.02e1.28)
control
Mulholland et al. 20082 Up to May 2008 14 17 673 492 011 Cohort High vs. lowd 1.14 (0.95e1.38) N/A
Up to May 2008 14 17 673 492 011 Cohort High vs. lowe 1.11 (0.99e1.25) 1.03 (0.94e1.12)
Dong et al. 20113 Up to Nov 2010 10 15 839 577 538 Cohort High vs. low 1.08 (1.02e1.14) 1.03 (0.98e1.09)
Choi et al. 20124 Up to Sept 2011 11 21 434 870 572 Cohort High vs. low 1.06 (1.02e1.11) 1.04 (0.96e1.12)
Turati et al. 20155 Up to Jan 2015 19f 43 926f 1 385 962 Cohort þ case- High vs. low 1.05 (0.99e1.11) 1.07 (0.98e1.16)
18g 43 595g control
Mullie et al. 20166 Up to Dec 2011 12 20 973 773 971 Cohort High vs. low 1.05 (1.00e1.11) 1.06 (1.00e1.13)
16
Each study is referenced by the meta-analyses in which it is included. Original references can be sourced from the meta-analyses.
F: Female; M: Male; N/A: Not available.
a
Unit for GI: 10 units/d.
b
Unit for GL: 100 units/d.
c
Unit: 1 SD.
d
Post-menopause.
e
Pre-menopause.
f
For analysis on GI.

21
g
For analysis on GL.
h
Cohort size for prospective studies or number of controls for case-control studies.
 22
Table 2 Characteristics of studies and meta-analyses on dietary glycemic index (GI), glycemic load (GL), and risk of colorectal cancer.

Author, year (sex) Study name Study location and Age at baseline Cases (n) Cohort size/controls (n)h Comparison GI RR (95%CI) GL RR (95%CI)
follow-up (years) (years)
Prospective studies
Terry et al., 2003 (F)1,2,4 NBSS Canada (16.5) 40e59 616 49 124 High vs low N/A 1.05 (0.73e1.53)
Higginbotham et al., 2004 (F)1,2,3,4,6 WHS USA (7.9) >45 174 38 451 High vs low 1.71 (0.98e2.98) 2.85 (1.40e5.80)
WHS USA (7.9) >45 174 38 451 Continuous 1.05 (1.00e1.11)a 1.18 (1.04e1.33)b
Oh et al., 2004, (F) NHS USA (<18) 30e55 1715 34 428 High vs low 1.11 (0.94e1.32) 0.92 (0.76e1.11)
Michaud et al., 2005 (F, M)1,2,3,4,5,6 NHS USA (20) 30e55 1096 83 927 High vs low (F) 1.08 (0.87e1.34) 0.89 (0.71e1.11)
HPFS USA (14) 40e75 683 47 422 High vs low (M) 1.14 (0.88e1.48) 1.32 (0.98e1.79)
Larsson et al., 2006 (F)1,2,3,4,5,6 SMC Sweden (15.7) 40e76 870 61 433 High vs low 1.00 (0.75e1.33) 1.06 (0.81e1.39)
McCarl et al., 2006 (F)1,2,3,4,5,6 IWHS USA (15.0) 55e69 954 35 197 High vs low 1.08 (0.88e1.32) 1.09 (0.88e1.35)
Strayer et al., 2007 (F)1,2,3,4,6 BCDDP USA (8.5) 35e74 490 45 561 High vs low 0.75 (0.56e1.00) 0.91 (0.70e1.20)
Weijenberg et al., 2007 (F)1,2,3,4,5,6 NCS Netherlands 55e69 1811 120 852 High vs low 0.81 (0.61e1.08) 0.83 (0.64e1.08)
(11.3)
Howarth et al., 2008 (F,M)4 MEC USA (8) 45e75 1086 105 106 F High vs low N/A 0.75 (0.57e0.97)
MEC USA (8) 45e75 1293 85 898 M High vs low N/A 1.15 (0.89e1.48)
Kabat et al., 2008 (F)2,3,4,5 WHI USA (7.8) 50e79 1476 157 324 High vs low 1.10 (0.92e1.32) 1.11 (0.82e1.49)
George et al., 2009 (F, M)3,4,6 NIH-AARP USA 50e71 1457 183 535 High vs low (F) 1.16 (0.98e1.37) 0.87 (0.64e1.18)
NIH-AARP USA 50e71 3031 262 642 High vs low (M) 1.16 (1.04e1.30) 0.88 (0.72e1.08)
Li et al., 2011 (F)3,4,5,6 SWHS China (9.1) 40e70 475 73 061 High vs low 1.09 (0.81e1.46) 0.94 (0.71e1.24)
Sieri et al., 2014 (F,M)6 EPIC-Italy Italy (11.7) 41e70 421 47 749 High vs low 1.35 (1.03e1.78) 1.43 (0.94e2.18)
Abe et al., 2016 (F,M) JPHC Study Japan (12.5) 45e74 579 38 941 High vs low (F) 0.97 (0.73e1.30) 0.82 (0.55e1.24)
JPHC Study Japan (12.5) 45e74 889 34 560 High vs low (M) 0.92 (0.73e1.14) 0.79 (0.58e1.08)
Case-control studies
Slattery et al., 1997 (F,M)1,2,5,6 N/A USA 30e79 894 1120 High vs low (F) 1.34 (1.00e1.81) N/A
N/A USA 30e79 1099 1290 High vs low (M) 1.37 (1.04e1.82) N/A
Franceschi et al., 2001 (F,M)1,2,5,6 N/A Italy 19e74 1953 4154 High vs low 1.70 (1.40e2.00) 1.80 (1.50e2.20)
N/A Italy 19e74 1953 4154 Continuous 1.16 (1.10e1.22)c 1.20 (1.13e1.29)d
Levi et al., 2002 (F,M)1,2,5,6 N/A Switzerland 35e74 323 1145 High vs low 1.82 (1.19e2.78) N/A
Murtaugh et al., 2006 (F,M)1,2,5,6 N/A USA 30e79 2450 2821 High vs low 1.19 (0.89e1.60) N/A
Hu et al., 2013 (F,M)6 NECSS Canada 20e76 2174 5039 High vs low 1.05 (0.90e1.24) 1.28 (1.05e1.57)
Zelensky et al., 2014 (F,M) N/A USA 1093 1589 High vs low N/A 1.61 (1.25e2.07)
Author, year Medline search N studies Cases (n) Study size (n) Studies assessed Comparison GI RR (95%CI) GL RR (95%CI)
Meta-analyses
Gnagnarella et al. 20081 Up to Oct 2007 11 12 523 491 053 Cohort þ case- High vs low 1.08 (0.96e1.22) 1.14 (1.02e1.28)
control
Mulholland et al. 20092 Up to Jul 2008 10 20 330 719 066 Cohort þ case- High vs low 1.15 (0.99e1.34) 1.17 (0.98e1.39)
control
Choi et al. 20123 Up to Sept 2011 9 15 568e 1 351 019e Cohort High vs low 1.08 (1.00e1.17) 0.99 (0.90e1.09)
11 18 563f 1 591 147f

S. Sieri, V. Krogh
Dietary glycemic index, glycemic load and cancer 23

1.07 (0.99e1.15)g 1.01 (0.95e1.08)d e found no evidence for a link between GI/GL and ovarian
1.07 (0.99e1.16) 1.00 (0.91e1.10)

1.17 (1.00e1.36) 1.01 (0.84e1.21)

1.10 (0.97e1.25)
cancer (Table 3).

Prostate cancer

Two 2015 meta-analyses [16,39] e the former of 2 cohort

and 3 case-control studies, total 26 500 cases; the latter of
1.16 (1.07e1.25)

9 cohort and 18 case-control studies, 39 352 cases e found

no significant association between GI/GL and risk of
prostate cancer overall. However, associations were sig-
nificant for case-control studies and for studies carried out
in Europe [39] (Table 3).
One again a high carbohydrate diet has been proposed
to increase prostate cancer risk through hyperinsulinemia
High vs low

High vs low

High vs low
Continuous

and increased IGF-1 availability [7]. This is supported by an

analysis of prostate cancer risk in relation to circulating
IGFs in up to 10 554 prostate cancer cases and 13 618
controls. In adjusted models, high circulating IGF-I was
significantly associated with increased prostate cancer
Each study is referenced by the meta-analyses in which it is included. Original references can be sourced from the meta-analyses.

risk, which, according to the authors, was strong evidence

that IGF-I is involved in prostate cancer development [40].
Cohort þ case-

Cohort þ case-

Abdominal obesity, especially visceral fat, is associated

with increased hepatic glucose synthesis, reduced glucose
control

control

metabolism, higher free fatty acids and lower levels of

Cohort
Cohort

SHBG, which may all increase prostate cancer risk through

various pathways [41].
1 146 121e
1 405 064f

Bladder cancer
994 154
994 154
589 746

The previously cited meta-analysis [9]on risk of ‘diabetes-

related-cancers’ found no association of GI/GL with
bladder cancer. Two other epidemiological studies exam-
ined GI/GL and bladder cancer [11,42]. The cohort study
22 692e
22 014f
12 382
12 382
13 015

Cohort size for prospective studies or number of controls for case-control studies.

[11] found a positive association of dietary GL with risk,

whereas the case-control study [42] found no association
of dietary GL with risk, but did find a direct association of
GI with risk (Table 3).

Kidney cancer

A 2009 cohort study [11] and 2013 case-control study [42]

14
14
10

15

found no evidence of a relation between GI/GL and renal

Up to Oct 2011
Up to Oct 2011

Up to Jan 2015

cancer. However a 2009 case-control study [43] found that

Up to 2011

GI/GL were directly related to risk of renal cell cancer.

Cigarette smoking, obesity, and hypertension are the most
consistently established risk factors for this cancer but
account for less than half of cases [44]. Since a high car-
F: Female; M: Male; N/A: Not available.

bohydrate diet can give rise to obesity it too may predis-

pose to this cancer, however the available data do not
support this hypothesis (Table 3).
Unit for GI: 1 unit/d.

Liver cancer
Unit GL: 50 units/d.
Unit GL: 10 units/d.

Unit GI: 10 units/d.

For analysis on GL.
For analysis on GI.
Unit GI: 5 units/d.
Galeone et al. 20125

Turati et al. 20156

Aune et al. 20124

The previously-cited 2014 meta-analysis [16], assessed

1157 cases of liver cancer in four studies, and found that
dietary GL/GI were not associated with liver cancer risk.
Most of the individual studies had similar findings; how-
ever two case-control studies [45,46] found that high di-
16

f
a
b

d
c

g
h

etary GL was significantly associated with increased risk of

 24
Table 3 Characteristics of studies and meta-analyses on dietary glycemic index (GI), glycemic load (GL), and risk of reproductive system and kidney cancers.

Author, year (sex) Study name Study location and Age at baseline Cases (n) Cohort size/controls (n)f Comparison GI RR (95%CI) GL RR (95%CI)
follow-up (years) (years)
Endometrium
Prospective studies
Folsom et al., 2003 (F)1,2,3,4,5,6 IWHI USA (>15) 55e69 415 23 335 High vs. low 1.05 (0.77e1.43) 1.24 (0.90e1.72)
Silvera et al., 2005 (F)1,2,3,4,5,6 NBSS Canada (16.4) 40e59 426 49 613 High vs. low 1.47 (0.90e2.41) 1.36 (1.01e1.84)
Larsson et al., 2006 (F)1,2,3,4,5,6 SMC Sweden (15.6) 40e76 608 66 651 High vs. low 1.00 (0.77e1.30) 1.15 (0.88e1.51)
Cust et al., 2007 (F)1,2,3,4,5,6 EPIC Europe (6.4) 20e85 710 288 428 High vs. low 1.04 (0.84e1.28) 1.15 (0.94e1.41)
George et al., 2009 (F)3,4,5,6 NIH-AARP USA (1995e2003) 50e71 1.041 183 535 High vs. low 0.85 (0.70e1.04) 1.25 (0.86e1.81)
Cui et al., 2011 (F)6 NHS USA (1980e2006) 30e55 669 68 070 High vs. low N/A 1.48 (1.14e1.92)
Coleman et al. 20146 PLCO USA (9.0) 55e75 386 36 115 High vs. low 0.94 (0.70e1.26) 0.63 (0.46e0.84)
PLCO USA (9.0) 55e75 386 36 115 Continuous 0.94 (0.85e1.04) 0.81 (0.72e0.92)
Case-control studies
Augustin et al., 2003 (F)1,2,4,5,6 N/A Italy 23e74 419 753 High vs. low 1.90 (1.25e2.91) 1.10 (0.60e2.03)
N/A Italy 23e74 419 753 Continuousa 1.34 (1.07e1.69) 1.48 (1.13e1.94)
Nagle et al., 2013 (F)5,6 ANECS Australia 18e79 1290 1436 High vs. low 1.43 (1.11e1.83) 1.15 (0.90e1.48)
ANECS Australia 18e79 1290 1436 Continuous 1.13 (1.03e1.24)b 1.14 (0.93e1.40)c
Galeone et al., 2013 (F)4,6 N/A Italy 18e79 454 908 High vs. low 1.03 (0.85e1.32) 1.01 (0.64e1.61)
N/A Italy 18e79 454 908 Continuousa 1.06 (0.85e1.32) 1.04 (0.83e1.29)
Brenner et al., 2015 (F) N/A Canada 30e79 511 980 High vs. low 0.77 (0.44e1.36) 0.87 (0.52e1.46)
Xu et al., 2015 (F)6 N/A China 30e69 1199 1212 High vs. low 2.20 (1.20e4.00) 1.40 (1.00e2.00)
Ovary
Prospective studies
Silvera et al., 2007 (F)7 NBSS Canada (16.4) 40e59 264 49 613 High vs. low 1.27 (0.65e2.47) 1.72 (1.13e2.62)
George et al., 2009 (F)7 NIH-AARP USA (1995e2003) 50e71 475 183 535 High vs. low 0.90 (0.67e1.23) 0.48 (0.28e0.84)
Case-control studies
Augustin et al., 2003 (F)7 N/A Italy 23e74 1031 2411 High vs. low 1.65 (1.30e2.09) 1.65 (1.30e2.09)
Nagle et al., 2011 (F)7 ANECS Australia 18e79 1366 1414 High vs. low 1.09 (0.87e1.36) 1.24 (1.00e1.55)
Hu et al., 2013 (F)7 NECSS Canada 20e76 442 2492 High vs. low 0.84 (0.61e1.17) 1.10 (0.62e1.95)
Quin et al., 2015 (F) AACES USA 20e79 406 609 High vs. low 1.03 (0.70e1.50) 1.35 (0.93e1.97)
Prostate
Prospective studies
Shikany et al., 2011 (M)8,9 PLCO USA (9.2) 55e74 2436 38 343 High vs. low 0.95 (0.82e1.09) 0.89 (0.71e1.12)
Nimptsch et al., 2011 (M)8,9 HPFS USA (1986e2007) 40e75 5112 49 934 High vs. low 1.00 (0.91e1.10) 0.92 (0.84e1.00)
George et al., 2009 (M)8,9 NIH-AARP USA (1995e2003) 50e71 15 900 262 642 High vs. low 0.98 (0.93e1.03) 0.48 (0.28e0.84)
Case-control studies
Hu et al., 2013 (M)8,9 NECSS Canada 20e76 1799 2547 High vs. low 1.26 (1.03e1.54) 1.29 (0.92e1.81)
Augustin et al., 2004 (M)8,9 N/A Italy 46e74 1204 1352 High vs. low 1.57 (1.19e2.07) 1.41 (1.04e1.89)
N/A Italy 46e74 1204 1352 Continuousa 1.28 (1.11e1.47) 1.22 (1.05e1.41)
Vidal et al., 2015 (M)9 N/A USA 63 (SD 6.0) 156 274 High vs. low 2.16 (1.06e4.42) N/A
Bladder
Prospective studies

S. Sieri, V. Krogh
George et al., 2009 (F,M)10 NIH-AARP USA (1995e2003) 50e71 235 183 535 F High vs. low 0.91 (0.60e1.38) 0.89 (0.41e1.91)
NIH-AARP USA (1995e2003) 50e71 1246 262 642 M High vs. low 1.29 (1.07e1.54) 0.99 (0.72e1.36)
Case-control studies
Hu et al., 2013 (M) NECSS Canada 20e76 1029 5039 High vs. low 1.26 (1.01e1.58) 0.92 (0.69e1.22)
 Dietary glycemic index, glycemic load and cancer
Kidney
Prospective studies
George et al., 2009 (F,M) NIH-AARP USA (1995e2003) 50e71 322 183 535 F High vs. low 0.84 (0.59e1.21) 0.78 (0.39e1.51)
NIH-AARP USA (1995e2003) 50e71 857 262 642 M High vs. low 1.05 (0.84e1.31) 1.05 (0.72e1.55)
Case-control studies
Galeone et al., 2009 N/A Italy 24e79 767 1534 High vs. low 1.43 (1.05e1.95) 2.56 (1.78e3.70)
N/A Italy 24e79 767 1534 Continuousa 1.23 (1.05e1.44) 1.39 (1.17e1.64)
Hu et al., 2013 NECSS Canada 20e76 1345 5039 High vs. low 1.18 (0.98e1.43) 1.09 (0.86e1.38)
Meta-analyses
Author, year Medline search N studies Cases (n) Study size (n) Studies assessed Comparison GI RR (95%CI) GL RR (95%CI)
Endometrium
Mulholland et al. 20081 Up to Dec 2007 5 2569 413 558 Cohort þ case- High vs. low 1.20 (0.95e1.51) 1.38 (1.08e1.77)
control
Gnagnarella et al. 20082 Up to Oct 2007 5 2578 428 780 Cohort þ case- High vs. low 1.22 (1.01e1.49) 1.36 (1.14e1.62)
control
Choi et al. 20123 Up to Sept 2011 5 3200 611 562 Cohort High vs. low 1.00 (0.87e1.14) 1.21 (1.07e1.37)
Galeone et al. 20124 Up to Sept 2011 6 3619 612 315 Cohort þ case- High vs. low 1.09 (0.92e1.29) 1.19 (1.06e1.34)
control
Nagle et al. 20135 Up to Jul 2011 7 4909 613 751 Cohort þ case- High vs. low 1.15 (0.95e1.40) 1.21 (1.09e1.33)
control
Turati et al. 20156 Up to Jan 2015 10d 6939d 651 078d Cohort þ case- High vs. low 1.13 (0.98e1.32) 1.17 (1.00e1.37)
11e 7608e 719 148e control
Ovary
Turati et al. 20157 Up to Jan 2015 5 3578 239 465 Cohort þ case- High vs. low 1.11 (0.85e1.46) 1.19 (0.85e1.68)
control
Prostate
Turati et al. 20158 Up to Jan 2015 6d 26 656d 354 818d Cohort þ case- High vs. low 1.06 (0.96e1.18) 1.04 (0.91e1.18)
5e 26 500e 354 544e control
Wang et al. 20159 Up to April 2015 6 39 352 355 092 Cohort þ case- High vs. low 1.06 (0.96e1.18) 1.04 (0.91e1.18)
control
Bladder
Choi et al. 201210 Up to Sept 2011 1 1481 446 177 Cohort High vs. low 1.14 (0.82e1.58) 0.97 (0.73e1.31)
110
Each study is referenced by the meta-analyses in which it is included. Original references can be sourced from the meta-analyses.
F: Female; M: Male; N/A: Not available.
a
Unit: difference between 80th and 20th percentile.
b
Unit for GI: 5 units/d.
c
Unit for GL: 50 units/d.
d
For analysis on GI.
e
For analysis on GL.
f
Cohort size for prospective studies or number of controls for case-control studies.

25
26
hepatocellular carcinoma in patients with chronic
hepatitis.
A supposed link between GI/GL and liver cancer has
been suggested as mediated by the association of high
dietary GI/GL with obesity and positive energy balance,
which may lead to insulin resistance. Positive risk associ-
ations have been reported between GL and risk of diabetes
[47], which is in turn risk factor for liver cancer [47], as is
obesity [48] (Table 4). Insulin resistance can increase fatty
acid turnover, leading to altered liver metabolism in turn
resulting in increased glucose output, greater synthesis of
proinflammatory cytokines, and alterations in lipoprotein
metabolism [49]. These changes could increase the risk of
hepatocellular carcinoma.
Pancreatic cancer
A 2008 meta-analysis of four studies [15], two 2012 meta-
analyses of ten cohort studies [9,50], and a 2015 meta-
analysis of 11 studies [16] found no evidence of a link
between a dietary GI/GL and risk of pancreatic cancer.
However, these findings are in contrast to those of a 2005
case-control study on men smokers, which found, after
adjusting for age, smoking, and body mass index, that
baseline fasting serum concentrations of glucose and in-
sulin, and presence of insulin resistance, were directly
associated with risk of pancreatic cancer [51]. Similarly, in
a 2007 study, on 33 293 women, 31 304 men and 2478
incident cases of various types of cancer, risk of pancreatic
cancer was significantly associated with high fasting
glucose that was unaffected by adjustment for BMI [52]
(Table 4).
Stomach cancer
A 2016 meta-analysis that analyzed six studies (2 cohort
and 4 case-control) found no significant association be-
tween dietary GI/GL and risk of gastric cancer overall [53].
However, high GL was significantly associated with gastric
cancer in males but not females [53] (Table 4). The
increased blood glucose and insulin levels resulting from
high GI/GL diets have been suggested to contribute to
gastric carcinogenesis by the activation of the IGF axis [7].
IGF-I has been linked to increased mitogenesis in gastric
cancer cell lines [54], and high IGF-I levels have been
found in patients with gastric cancer in comparison to
healthy controls [55]. Furthermore a cohort study reported
a direct relationship between fasting plasma glucose levels
and risk of stomach cancer in people seropositive for
Helicobacter pylori [56].
Esophageal cancer
A 2015 meta-analysis of four studies for a total of 1097
cases reported summary relative risks for esophageal
cancer in relation to high dietary GI and GL that were
above unity, but were not significant [16] (Table 4). As
noted, high carbohydrate intake is associated with
S. Sieri, V. Krogh
increased blood insulin which may promote cancer
development by altering sex hormone metabolism and
increasing IGF-1 bioactivity, which may in turn inhibit
apoptosis [57]. IGF-1 has been shown to stimulate growth
in human esophageal cancer cell lines [58]. By increasing
insulin resistance, hyperinsulinemia and hyperglycemia,
obesity may exacerbate the effect of the high GI diet, and
contribute to the risk of esophageal cancer.
Lung cancer
Few studies have addressed associations of GI/GL with
lung cancer. Two case-control studies [59,60] found that
high GI was significantly associated with increased lung
cancer risk; whereas none of the epidemiological studies
(3 case-control and one cohort studies) that addressed the
issue [11,42,59,60] found a significant association of GL
with lung cancer (Table 4). Animal and cell culture studies
show that IGF-1 is a potent mitogen for lung cancer and
other cancer cells [61]. However a meta-analysis of nested
case-control studies found no evidence for a significant
association of circulating IGF-1 levels and lung cancer risk
[62]. Further research is necessary to understand the
mechanisms linking GI/GL, the IGF axis, and lung cancer
risk in humans.
Discussion
The main finding of this review is that a high GI/GL diet is
associated with weakly or moderately increased risk of
only a few types of cancer. In particular high GI was
associated with increased risk of colorectal cancer in
cohort and case-control studies [9,16]; and high GI and GL
were associated with increased risk of breast cancer [14],
and high GL with increased risk of endometrial cancer [9]
in cohort studies. For cancer in general (diabetes-related
cancers) [9] a modest-to-weak association between high
dietary GI and increased risk was found, with no associa-
tion for GL.
Although several etiological hypotheses have been put
forward, chronically high blood glucose giving rise to
hyperinsulinemia, insulin resistance and enhanced bioac-
tivity of the IGF axis (and in particular of the potent mitogen
IGF-1) is the most commonly-invoked mechanism to ac-
count for associations of high carbohydrate diet with
increased cancer risk [27]. Insulin may also influence cancer
development by altering sex hormone metabolism [63].
The differing associations of dietary GI and dietary GL
with different types of cancer probably reflect that fact
that GI is a measure of glucose availability (carbohydrate
quality) and is independent of quantity, while GL is a
measure of the total glycemic effect, and is hence an in-
dicator of the insulin demand of the diet. If risks of breast
and endometrial cancer are related to the overall insulin
demand of the diet, then that risk should show a stronger
relation to GL than GI, since GL is more likely to lead to
chronically increased blood glucose levels [64]. By contrast
colorectal cancer risk seems to depend on the quality of
the carbohydrate consumed rather than the overall
Table 4 Characteristics of studies and meta-analyses on dietary glycemic index (GI), glycemic load (GL), and risk of other types of cancer.

Dietary glycemic index, glycemic load and cancer

Author, year (sex) Study name Study location and Age at baseline Cases (n) Cohort size/ Comparison GI RR (95%CI) GL RR (95%CI)
follow-up (years) (years) controls (n)h
Liver
Prospective studies
George et al., 2009 (F,M)1,2 NIH-AARP USA (1995e2003) 50e71 72 183 535 F High vs. low 0.95 (0.43e2.10) 0.18 (0.04e0.79)
NIH-AARP USA (1995e2003) 50e71 238 262 642 M High vs. low 1.62 (1.05e2.48) 0.47 (0.23e0.95)
Fedirko et al., 2013 (F,M)2 EPIC Europe (11.4) 34e66 191 477 206 High vs. low 1.09 (0.71e1.66) 1.19 (0.72e1.97)
EPIC Europe (11.4) 34e66 191 477 206 Continuous 1.04 (0.71e1.51)a 1.19 (0.64e2.21)b
Vogtmann et al., 2013 (F,M)2 SWHS/SMHS China (11.2) 40e70 139 72 966 F High vs. low 2.17 (1.08e4.35) 1.02 (0.59e1.79)
SWHS/SMHS China (11.2) 40e70 208 60 207 M High vs. low 0.89 (0.58e1.37) 1.07 (0.68e1.67)
Case-control studies
Rossi et al., 2009 (F,M)2 N/A Italy 43e84 185 412 High vs. low N/A 3.02 (1.49e6.12)
N/A Italy 43e84 185 412 Continuousc N/A 1.40 (1.15e1.69)
Lagiou et al., 2009 (F,M)2 Greece 63.9 (9.6) 333 360 High vs. low N/A 1.50 (0.90e2.50)
Continuousc N/A 1.08 (0.93e1.25)
Hu et al., 2013 (M)2 NECSS Canada 20e76 309 5039 High vs. low 0.71 (0.49e1.01) 1.17 (0.75e1.84)
Pancreas
Prospective studies
Michaud et al., 2002 (F,M)3,4,5,6 NHS USA (18) 55e69 180 88 802 High vs. low 1.16 (0.69e1.97) 1.53 (0.96e2.45)
Silvera et al., 2005 (F)3,4,5,6 NBSS Canada (16.5) 40e59 112 49 613 High vs. low 1.43 (0.56e3.65) 0.80 (0.45e1.41)
Johnson et al., 2005 (F)3,4,5,6 IWHS USA (16) 55e69 181 33 551 High vs. low 1.08 (0.74e1.58) 0.87 (0.56e1.34)
Patel et al., 2007 (F,M)3,4,5,6 USA (9) 50e74 401 124 907 High vs. low 0.80 (0.53e1.20) 1.10 (0.73e1.64)
Nothlings et al., 2007 (F,M)4,6 MCS USA (8) 45e75 434 162 150 High vs. low N/A 1.10 (0.80e1.52)
Heinen et al., 2008 (F,M)4,5 NCS Netherlands (13.3) 55e69 408 3980 High vs. low 0.87 (0.59e1.29) 0.85 (0.58e1.24)
NCS Netherlands (13.3) 55e69 408 3980 Continuous 0.98 (0.81e1.19)a 1.03 (0.77e1.39)b
George et al., 2009 (F,M)4,6 NIH-AARP USA (1995e2003) 50e71 348 183 535 F High vs. low 1.00 (0.71e1.40) 0.49 (0.26e0.94)
NIH-AARP USA (1995e2003) 50e71 601 262 642 M High vs. low 1.19 (0.92e1.55) 0.67 (0.42e1.08)
Jiao et al., 2009 (F,M)5 NIH-AARP USA (1995e2003) 50e71 1151 482 362 High vs. low 1.09 (0.90e1.32) 0.95 (0.74e1.22)
Meinhold et al., 2010 (F,M)4,5,6 PLCO USA (6.5) 55e74 266 109 175 High vs. low 1.00 (0.69e1.47) 1.41 (0.97e2.07)
Simon et al., 2010 (F)4,5 WHI USA (8) 50e79 287 139 503 High vs. low 1.13 (0.78e1.63) 0.80 (0.55e1.15)
Case-control studies
Rossi et al., 2009 (F,M)6 N/A Italy 43e84 326 652 High vs. low 1.78 (1.20e2.62) 1.26 (0.83e1.91)
Hu et al., 2013 (F,M)6 NECSS Canada 20e76 628 5039 High vs. low 1.09 (0.84e1.42) 1.00 (0.94e1.53)
Stomach
Prospective studies
Larsson et al. 20067,8 SMC Sweden (17$4) 40e74 156 61 277 High vs. low 0.77 (0.46e1.30) 0.76 (0.46e1.25)
George et al., 2009 (F,M)7,8 NIH-AARP USA (1995e2003) 50e71 127 183 535 F High vs. low 1.12 (0.64e1.97) 0.67 (0.24e1.90)
NIH-AARP USA (1995e2003) 50e71 440 262 642 M High vs. low 1.50 (1.09e2.08) 1.42 (0.81e2.49)
Case-control studies
Augustin et al. 20047,8 N/A Italy 19e79 769 2081 High vs. low 1.06 (0.82e1.38) 1.94 (1.47e2.55)
Bertuccio et al. 20097,8 N/A Italy 22e80 230 547 High vs. low 2.10 (1.20e3.60) 2.70 (1.50e4.8)
N/A Italy 22e80 230 547 Continuousc 1.40 (1.10e1.90) 1.50 (1.20e2.00)
Lazarevic et al. 20097,8 N/A Serbia 45e85 102 204 High vs. low 0.39 (0.14e1.15) 1.28 (0.44e3.76)
Hu et al. 20137,8 NECSS Canada 20e76 1182 5039 High vs. low 0.90 (0.74e1.11) 0.93 (0.67e1.29)
Esophagus
Prospective studies
George et al., 2009 (F,M)9 NIH-AARP USA (1995e2003) 50e71 76 183 535 F High vs. low 1.27 (0.60e2.67) 2.18 (0.57e8.32)
NIH-AARP USA (1995e2003) 50e71 425 262 642 M High vs. low 1.50 (1.10e2.05) 0.65 (0.38e1.11)
Case-control studies

27
(continued on next page)
Table 4 (continued )

28
Author, year (sex) Study name Study location and Age at baseline Cases (n) Cohort size/ Comparison GI RR (95%CI) GL RR (95%CI)
follow-up (years) (years) controls (n)h
Augustin et al. 20039 N/A Italy <77 304 743 High vs. low 1.5 (0.90e2.6) 1.3 (0.5e2.9)
N/A Italy <77 304 743 Continuous 1.1 (1.0e1.4)d 1.2 (1.0e1.5)e
Mulholland et al., 2009 FINBAR Ireland 85 220 260 High vs. low 1.52 (0.84e2.76) 1.14 (0.55e2.33)
FINBAR Ireland 85 220 260 Continuous 1.42 (1.07e1.89)d 0.93 (0.63e1.39)b
Eslamian et al. 20139 N/A Kurdistan 47 96 High vs. low 2.95 (1.68e3.35) 2.95 (1.68e3.35)
Lahmann et al. 20149 N/A Australia 18e79 245 1507 High vs. low 0.73 (0.46e1.14) 0.52 (0.33e0.82)
N/A Australia 18e79 245 1507 Continuous 0.89 (0.66e1.21)d 0.70 (0.50e0.97)b
Lung
Prospective studies
George et al., 2009 (F,M) NIH-AARP USA (1995e2003) 50e71 2228 183 535 F High vs. low 1.12 (0.98e1.27) 0.81 (0.64e1.03)
NIH-AARP USA (1995e2003) 50e71 3769 262 642 M High vs. low 1.08 (0.98e1.20) 0.93 (0.78e1.11)
Case-control studies
Hu et al., 2013 NECSS Canada 20e76 3341 5039 High vs. low 1.04 (0.89e1.23) 0.98 (0.80e1.21)
De stefani et al., 2016 N/A Uruguay 30e89 463 465 High vs. low 2.77 (1.28e5.97) N/A
Melkonian et al., 2016 N/A Texas Mean 60.7 1905 2413 High vs. low 1.49 (1.21e1.83) 1.16 (0.94e1.42)
Meta-analyses
Author, year Medline search N studies Cases (n) Study size (n) Studies assessed Comparison GI RR (95%CI) GL RR (95%CI)
Liver
Choi et al. 20121 Up to Sept 2011 1 310 446 177 Cohort High vs. low 1.38 (0.86e2.23) 0.37 (0.16e0.84)
Turati et al. 20152 Up to Jan 2015 4f 1157f 1 061 595f Cohort þ case- High vs. low 1.11 (0.80e1.53) 1.14 (0.78e1.67)
6g 1675g 1 062 367g control
Pancreas
Gnagnarella et al. 20083 Up to Oct 2007 4 874 296 873 Cohort þ case- High vs. low 1.11 (0.86e1.43) 1.14 (1.02e1.28)
control
Choi et al. 20124 Up to Sept 2011 8f 2183f 1 157 858f Cohort High vs. low 1.05 (0.93-1-19) 0.95 (0.79e1.13)
9g 2617g 1 320 008g
Aune et al. 20125 Up to Sept 2011 8 2550f 1 031 893f Cohort High vs. low 1.04 (0.93e1.17) 1.01 (0.88e1.15)
2986g 1 194 043g
Continuous 1.02 (0.93e1.11)a 1.03 (0.93e1.14)b
Turati et al. 20156 Up to Jan 2015 10f 3680f 1 001 399f Cohort þ case- High vs. low 1.10 (0.99, 1.22) 1.01 (0.85, 1.19)
11g 4114g 1 163 549g control
Stomach
Ye et al. 20157 Up to March 2015 6 3600 515 325 Cohort þ case- High vs. low 1.17 (0.80e1.69) 1.06 (0.90e1.26)
control
Turati et al. 20158 Up to Jan 2015 6 3600 515 325 Cohort þ case- High vs. low 1.17 (0.83e1.66) 1.10 (0.85e1.42)
control
Esophagus
Turati et al. 20159 Up to Jan 2015 4 1097 448 523 Cohort þ case- High vs. low 1.46 (0.90e2.38) 1.25 (0.45e3.48)
control
19
Each study is referenced by the meta-analyses in which it is included. Original references can be sourced from the meta-analyses.
F: Female; M: Male; N/A: Not available.
a

S. Sieri, V. Krogh
Unit for GI: 5 units/d.
b
Unit GL: 50 units/d.
c
Unit: difference between 80th and 20th percentile.
d
Unit for GI: 10 units/d.
e
Unit GL: 100 units/d.
f
For analysis on GI.
g
For analysis on GL.
h
Cohort size for prospective studies or number of controls for case-control studies.
Dietary glycemic index, glycemic load and cancer
quantity. Thus, although the Italian EPIC study found that
high dietary GI was significantly associated with increased
risk of colorectal cancer, when carbohydrate intake was
divided into that from high GI (mainly white bread, sugar/
honey and jam, pizza and refined rice) and that from low-
medium GI foods (mainly pasta and fruit), only high car-
bohydrate intake from high GI foods was significantly
associated with increased colorectal cancer risk, while
high carbohydrate intake from low GI foods was signifi-
cantly associated with decreased cancer risk [65].
The overall finding that associations of GI/GL with
cancer risk are only weak-to-modest is in line with the fact
that dietary GI/GL is one of several factors that contribute
to hyperinsulinemia. Others are genetic predisposition,
adiposity, physical activity [66], non-carbohydrate foods
that influence insulin secretion [67], and coffee con-
sumption that influences insulin resistance but not insulin
secretion [68].
The plasma insulin response is not, in general, propor-
tional to the blood glucose response Protein and fat intake
may also exert strong influences on insulin secretion [69].
For some cancers significant associations of high dietary
GI/GL intake with cancer risk were found only for those
who were overweight [16]. BMI strongly influences insulin
resistance and hyperinsulinemia [31] and overweight
persons may in general have a greater insulin response to
dietary intake than those who are lean [70].
Most of the reviewed studies did not do separate sub-
site analyses. Subsites for some cancers may have differing
sensitivities to the effects of a high carbohydrate diet, that
would not have been revealed by the aggregate analyses.
For example, colon cancer cells are more susceptible to the
growth-promoting effects of insulin in vitro than rectal
cancer cells [71] It is noteworthy also that colon and
rectum derive from different segments of embryonic in-
testinal tract and differ in function [72].
For esophageal cancer, the main histologic types
adenocarcinoma, gastro-esophageal junction adenocarci-
noma, and squamous cell carcinoma, are distinct entities
with distinct risk factors [73]. For gastric cancer also, the
various anatomic locations and histologic types appear to
be etiologically heterogeneous [68].
Errors in measuring carbohydrate consumption, and
between-study variation in methods of estimating carbo-
hydrate consumption, could also weaken hypothesized
associations between dietary carbohydrate and cancer risk.
Carbohydrate consumption is estimated by food frequency
questionnaires (FFQs). Various FFQs are in use, varying
from one with only 29-items [74]to one with 200 items
[75]. The more comprehensive instruments may provide
more accurate estimates of consumption [76]. Further-
more, while the FFQs used by most studies were designed
to provide estimates of total carbohydrate and energy
intake, they were not specifically designed to provide ac-
curate estimates of dietary GI and GL.
The GI of a mixed meal should be estimated from the
GIs of its carbohydrate constituents, but it is not clear that
different studies and FFQs calculate average dietary GI in
the same way. To complicate matters it is known that
29
glucose and insulin responses to a food item are influenced
by other items or macronutrients eaten with it, the cook-
ing procedure, and even the chewing time. Such factors are
not easily assessed by an FFQ.
Once consumption estimates has been obtained from
an FFQ, it is necessary to use food composition tables to
estimate GI. However the GI of a given food may vary
between countries since food variety, processing and
cooking vary. Local composition tables that measure GIs of
local foods are likely to be more accurate than those pre-
sent in international food composition tables.
Finally, positive associations between dietary GI/GL and
cancer risk were more frequently reported by case-control
than cohort studies. A possible reason for this is that case-
control studies are more prone to problems of recall and
selection bias than cohort studies. Furthermore most case-
control were carried out in Europe and most cohort
studies were carried out in North America. The differences
in results between European and North American may also
reflect differences in dietary habits between the two
continents. People in Europe consume higher levels of
carbohydrates and also a greater variety of carbohydrates
[77] than for people in North America [78], who tend to
consume proportionately more fats. This apparently more
homogenous and fat-rich diet of North Americans
compared to Europeans, would explain why North Amer-
ican studies often fail to find associations between carbo-
hydrate and cancer risk.
To conclude, this review has uncovered consistent evi-
dence that only the risks of colorectal cancer, breast can-
cer, and endometrial cancer are increased by a diet
inducing high blood glucose levels. The risk increases are
in all cases small-to-moderate. The risks of developing
other cancers do not appear to be influenced by dietary GI
or GL. However low GI diets have several other health
advantages and tend to be rich in dietary fiber, micro-
nutrients and phytochemicals (e.g. plant sterols) that may
be important elements in a healthy diet.
Funding
This research did not receive any specific grant from
funding agencies in the public, commercial, or not-for-
profit sectors.
Acknowledgement
The authors thank Don ward for help with the English.
References
[1] Austin GL, Ogden LG, Hill JO. Trends in carbohydrate, fat, and
protein intakes and association with energy intake in normal-
weight, overweight, and obese individuals: 1971-2006. Am J
Clin Nutr 2011 Apr;93(4):836e43.
[2] Jakobsen MU, O’Reilly EJ, Heitmann BL, Pereira MA, Balter K,
Fraser GE, et al. Major types of dietary fat and risk of coronary
heart disease: a pooled analysis of 11 cohort studies. Am J Clin
Nutr 2009 May;89(5):1425e32.
30
[3] DiNicolantonio JJ, Lucan SC, O’Keefe JH. The evidence for satu-
rated fat and for sugar related to coronary heart disease. Prog
Cardiovasc Dis 2015 Nov 14;58(5):464e72.
[4] Li Y, Hruby A, Bernstein AM, Ley SH, Wang DD, Chiuve SE, et al.
Saturated fats compared with unsaturated fats and sources of
carbohydrates in relation to risk of coronary heart disease: a
prospective cohort study. J Am Coll Cardiol 2015 Oct 6;66(14):
1538e48.
[5] Jenkins DJ, Wolever TM, Taylor RH, Barker H, Fielden H,
Baldwin JM, et al. Glycemic index of foods: a physiological basis
for carbohydrate exchange. Am J Clin Nutr 1981 Mar;34(3):
362e6.
[6] Brand-Miller JC. Glycemic load and chronic disease. Nutr Rev
2003 May;61(5 Pt 2):S49e55.
[7] Biddinger SB, Ludwig DS. The insulin-like growth factor axis: a
potential link between glycemic index and cancer. Am J Clin Nutr
2005 Aug;82(2):277e8.
[8] Samani AA, Yakar S, LeRoith D, Brodt P. The role of the IGF system
in cancer growth and metastasis: overview and recent insights.
Endocr Rev 2007 Feb;28(1):20e47.
[9] Choi Y, Giovannucci E, Lee JE. Glycaemic index and glycaemic
load in relation to risk of diabetes-related cancers: a meta-
analysis. Br J Nutr 2012 Dec 14;108(11):1934e47.
[10] Giovannucci E, Harlan DM, Archer MC, Bergenstal RM,
Gapstur SM, Habel LA, et al. Diabetes and cancer: a consensus
report. CA Cancer J Clin 2010 Jul;60(4):207e21.
[11] George SM, Mayne ST, Leitzmann MF, Park Y, Schatzkin A,
Flood A, et al. Dietary glycemic index, glycemic load, and risk of
cancer: a prospective cohort study. Am J Epidemiol 2009 Feb 15;
169(4):462e72.
[12] Mulholland HG, Murray LJ, Cardwell CR, Cantwell MM. Dietary
glycaemic index, glycaemic load and breast cancer risk: a systematic
review and meta-analysis. Br J Cancer 2008 Oct 7;99(7):1170e5.
[13] Dong JY, Qin LQ. Dietary glycemic index, glycemic load, and risk
of breast cancer: meta-analysis of prospective cohort studies.
Breast Cancer Res Treat 2011 Apr;126(2):287e94.
[14] Mullie P, Koechlin A, Boniol M, Autier P, Boyle P. Relation between
breast cancer and high glycemic index or glycemic load: a meta-
analysis of prospective cohort studies. Crit Rev Food Sci Nutr
2016;56(1):152e9.
[15] Gnagnarella P, Gandini S, La VC, Maisonneuve P. Glycemic index,
glycemic load, and cancer risk: a meta-analysis. Am J Clin Nutr
2008 Jun;87(6):1793e801.
[16] Turati F, Galeone C, Gandini S, Augustin LS, Jenkins DJ, Pelucchi C,
et al. High glycemic index and glycemic load are associated with
moderately increased cancer risk. Mol Nutr Food Res 2015 Jul;
59(7):1384e94.
[17] Althuis MD, Fergenbaum JH, Garcia-Closas M, Brinton LA,
Madigan MP, Sherman ME. Etiology of hormone receptor-defined
breast cancer: a systematic review of the literature. Cancer Epi-
demiol Biomarkers Prev 2004 Oct;13(10):1558e68.
[18] Romieu I, Ferrari P, Rinaldi S, Slimani N, Jenab M, Olsen A, et al.
Dietary glycemic index and glycemic load and breast cancer risk
in the European Prospective Investigation into cancer and
Nutrition (EPIC). Am J Clin Nutr 2012 Aug;96(2):345e55.
[19] Liu S, Manson JE, Stampfer MJ, Holmes MD, Hu FB, Hankinson SE,
et al. Dietary glycemic load assessed by food-frequency ques-
tionnaire in relation to plasma high-density-lipoprotein choles-
terol and fasting plasma triacylglycerols in postmenopausal
women. Am J Clin Nutr 2001 Mar;73(3):560e6.
[20] Key TJ, Appleby PN, Reeves GK, Roddam AW. Insulin-like growth
factor 1 (IGF1), IGF binding protein 3 (IGFBP3), and breast cancer
risk: pooled individual data analysis of 17 prospective studies.
Lancet Oncol 2010 Jun;11(6):530e42.
[21] Larsson SC, Bergkvist L, Wolk A. Glycemic load, glycemic index
and breast cancer risk in a prospective cohort of Swedish women.
Int J Cancer 2009 Jul 1;125(1):153e7.
[22] Qi L, Rimm E, Liu S, Rifai N, Hu FB. Dietary glycemic index, gly-
cemic load, cereal fiber, and plasma adiponectin concentration in
diabetic men. Diabetes Care 2005 May;28(5):1022e8.
[23] Ziemke F, Mantzoros CS. Adiponectin in insulin resistance: les-
sons from translational research. Am J Clin Nutr 2010 Jan;91(1):
258Se61S.
[24] Mulholland HG, Murray LJ, Cardwell CR, Cantwell MM. Glycemic
index, glycemic load, and risk of digestive tract neoplasms: a
S. Sieri, V. Krogh
systematic review and meta-analysis. Am J Clin Nutr 2009 Feb;
89(2):568e76.
[25] Aune D, Chan DS, Lau R, Vieira R, Greenwood DC, Kampman E,
et al. Carbohydrates, glycemic index, glycemic load, and colo-
rectal cancer risk: a systematic review and meta-analysis of
cohort studies. Cancer Causes Control 2012 Apr;23(4):521e35.
[26] Galeone C, Pelucchi C, La VC. Added sugar, glycemic index and
load in colon cancer risk. Curr Opin Clin Nutr Metab Care 2012
Jul;15(4):368e73.
[27] Giovannucci E. Insulin, insulin-like growth factors and colon
cancer: a review of the evidence. J Nutr 2001 Nov;131(11 Suppl.):
3109Se20S.
[28] Ma J, Giovannucci E, Pollak M, Leavitt A, Tao Y, Gaziano JM, et al.
A prospective study of plasma C-peptide and colorectal cancer
risk in men. J Natl Cancer Inst 2004 Apr 7;96(7):546e53.
[29] Schoen RE, Tangen CM, Kuller LH, Burke GL, Cushman M, Tracy RP,
et al. Increased blood glucose and insulin, body size, and incident
colorectal cancer. J Natl Cancer Inst 1999 Jul 7;91(13):1147e54.
[30] Kaaks R. Nutrition, energy balance and colon cancer risk: the role
of insulin and insulin-like growth factor-I. IARC Sci Publ 2002;
156:289e93.
[31] Giovannucci E, Michaud D. The role of obesity and related
metabolic disturbances in cancers of the colon, prostate, and
pancreas. Gastroenterology 2007 May;132(6):2208e25.
[32] Kendall CW, Emam A, Augustin LS, Jenkins DJ. Resistant starches
and health. J AOAC Int 2004 May;87(3):769e74.
[33] Leonel AJ, varez-Leite JI. Butyrate: implications for intestinal
function. Curr Opin Clin Nutr Metab Care 2012 Sep;15(5):474e9.
[34] Galeone C, Augustin LS, Filomeno M, Malerba S, Zucchetto A,
Pelucchi C, et al. Dietary glycemic index, glycemic load, and the
risk of endometrial cancer: a case-control study and meta-anal-
ysis. Eur J Cancer Prev 2013 Jan;22(1):38e45.
[35] Mulholland HG, Murray LJ, Cardwell CR, Cantwell MM. Dietary
glycaemic index, glycaemic load and endometrial and ovarian
cancer risk: a systematic review and meta-analysis. Br J Cancer
2008 Aug 5;99(3):434e41.
[36] World Health Organization. World cancer report 2014. 2014.
Report No.: 5.12.
[37] Kaaks R, Lukanova A, Kurzer MS. Obesity, endogenous hormones,
and endometrial cancer risk: a synthetic review. Cancer Epi-
demiol Biomarkers Prev 2002 Dec;11(12):1531e43.
[38] Lukanova A, Kaaks R. Endogenous hormones and ovarian cancer:
epidemiology and current hypotheses. Cancer Epidemiol Bio-
markers Prev 2005 Jan;14(1):98e107.
[39] Wang RJ, Tang JE, Chen Y, Gao JG. Dietary fiber, whole grains,
carbohydrate, glycemic index, and glycemic load in relation to
risk of prostate cancer. Onco Targets Ther 2015;8:2415e26.
[40] Travis RC, Appleby PN, Martin RM, Holly JM, Albanes D, Black A,
et al. A meta-analysis of individual participant data reveals an
association between circulating levels of IGF-I and prostate
cancer risk. Cancer Res 2016 Feb 26;76(8):2288e300.
[41] Garaulet M, Perez-Llamas F, Zamora S, Tebar FJ. Interrelationship
between serum lipid profile, serum hormones and other com-
ponents of the metabolic syndrome. J Physiol Biochem 2002 Sep;
58(3):151e60.
[42] Hu J, La VC, Augustin LS, Negri E, de GM, Morrison H, et al.
Glycemic index, glycemic load and cancer risk. Ann Oncol 2013
Jan;24(1):245e51.
[43] Galeone C, Pelucchi C, Maso LD, Negri E, Talamini R, Montella M,
et al. Glycemic index, glycemic load and renal cell carcinoma risk.
Ann Oncol 2009 Nov;20(11):1881e5.
[44] Lipworth L, Tarone RE, Lund L, McLaughlin JK. Epidemiologic
characteristics and risk factors for renal cell cancer. Clin Epi-
demiol 2009;1:33e43.
[45] Lagiou P, Rossi M, Tzonou A, Georgila C, Trichopoulos D, La VC.
Glycemic load in relation to hepatocellular carcinoma among
patients with chronic hepatitis infection. Ann Oncol 2009 Oct;
20(10):1741e5.
[46] Rossi M, Lipworth L, Maso LD, Talamini R, Montella M, Polesel J, et al.
Dietary glycemic load and hepatocellular carcinoma with or without
chronic hepatitis infection. Ann Oncol 2009 Oct;20(10):1736e40.
[47] El-Serag HB, Hampel H, Javadi F. The association between dia-
betes and hepatocellular carcinoma: a systematic review of
epidemiologic evidence. Clin Gastroenterol Hepatol 2006 Mar;
4(3):369e80.
Dietary glycemic index, glycemic load and cancer
[48] Larsson SC, Wolk A. Overweight, obesity and risk of liver cancer:
a meta-analysis of cohort studies. Br J Cancer 2007 Oct 8;97(7):
1005e8.
[49] Cornier MA, Dabelea D, Hernandez TL, Lindstrom RC, Steig AJ,
Stob NR, et al. The metabolic syndrome. Endocr Rev 2008 Dec;
29(7):777e822.
[50] Aune D, Chan DS, Vieira AR, Navarro Rosenblatt DA, Vieira R,
Greenwood DC, et al. Dietary fructose, carbohydrates, glycemic
indices and pancreatic cancer risk: a systematic review and meta-
analysis of cohort studies. Ann Oncol 2012 Oct;23(10):2536e46.
[51] Stolzenberg-Solomon RZ, Graubard BI, Chari S, Limburg P,
Taylor PR, Virtamo J, et al. Insulin, glucose, insulin resistance, and
pancreatic cancer in male smokers. J Am Med Assoc 2005 Dec 14;
294(22):2872e8.
[52] Stattin P, Bjor O, Ferrari P, Lukanova A, Lenner P, Lindahl B, et al.
Prospective study of hyperglycemia and cancer risk. Diabetes
Care 2007 Mar;30(3):561e7.
[53] Ye Y, Wu Y, Xu J, Ding K, Shan X, Xia D. Association between
dietary carbohydrate intake, glycemic index and glycemic load,
and risk of gastric cancer. Eur J Nutr 2016 Feb 12. http:
//dx.doi.org/10.1007/s00394-016-1166-4.
[54] Yi HK, Hwang PH, Yang DH, Kang CW, Lee DY. Expression of the
insulin-like growth factors (IGFs) and the IGF-binding proteins
(IGFBPs) in human gastric cancer cells. Eur J Cancer 2001 Nov;
37(17):2257e63.
[55] Franciosi CM, Piacentini MG, Conti M, Romano F, Musco F,
Caprotti R, et al. IGF-1 and IGF-1BP3 in gastric adenocarcinoma.
Preliminary study. Hepatogastroenterology 2003 Jan;50(49):
297e300.
[56] Yamagata H, Kiyohara Y, Nakamura S, Kubo M, Tanizaki Y,
Matsumoto T, et al. Impact of fasting plasma glucose levels on
gastric cancer incidence in a general Japanese population: the
Hisayama study. Diabetes Care 2005 Apr;28(4):789e94.
[57] Juan HC, Tsai HT, Chang PH, Huang CY, Hu CP, Wong FH. Insulin-
like growth factor 1 mediates 5-fluorouracil chemoresistance in
esophageal carcinoma cells through increasing survivin stability.
Apoptosis 2011 Feb;16(2):174e83.
[58] Khandwala HM, McCutcheon IE, Flyvbjerg A, Friend KE. The ef-
fects of insulin-like growth factors on tumorigenesis and
neoplastic growth. Endocr Rev 2000 Jun;21(3):215e44.
[59] De Stefani E, Pellegrini H, Mendilaharsu M, Ronco A, Carzoglio JC.
Dietary sugar and lung cancer: a case-control study in Uruguay.
Nutr Cancer 1998;31(2):132e7.
[60] Melkonian SC, Daniel CR, Ye Y, Pierzynski JA, Roth JA, Wu X.
Glycemic index, glycemic load, and lung cancer risk in non-
hispanic whites. Cancer Epidemiol Biomarkers Prev 2016 Mar;
25(3):532e9.
[61] DiGiovanni J, Kiguchi K, Frijhoff A, Wilker E, Bol DK, Beltran L,
et al. Deregulated expression of insulin-like growth factor 1 in
prostate epithelium leads to neoplasia in transgenic mice. Proc
Natl Acad Sci U S A 2000 Mar 28;97(7):3455e60.
31
[62] Chen B, Liu S, Xu W, Wang X, Zhao W, Wu J. IGF-I and IGFBP-3
and the risk of lung cancer: a meta-analysis based on nested
case-control studies. J Exp Clin Cancer Res 2009;28:89.
[63] Weroha SJ, Haluska P. The insulin-like growth factor system in
cancer. Endocrinol Metab Clin North Am 2012 Jun;41(2):335e50.
vi.
[64] Brand-Miller J, McMillan-Price J, Steinbeck K, Caterson I. Dietary
glycemic index: health implications. J Am Coll Nutr 2009 Aug;
28(Suppl.):446Se9S.
[65] Sieri S, Krogh V, Agnoli C, Ricceri F, Palli D, Masala G, et al. Dietary
glycemic index and glycemic load and risk of colorectal cancer:
results from the EPIC-Italy study. Int J Cancer 2015 Jun 15;
136(12):2923e31.
[66] Harriss DJ, Atkinson G, Batterham A, George K, Cable NT, Reilly T,
et al. Lifestyle factors and colorectal cancer risk (2): a systematic
review and meta-analysis of associations with leisure-time
physical activity. Colorectal Dis 2009 Sep;11(7):689e701.
[67] World Cancer Research Fund. Food, nutrition, and the prevention
of cancer: a global perspective. Washington, DC. 2007.
[68] Yu X, Bao Z, Zou J, Dong J. Coffee consumption and risk of can-
cers: a meta-analysis of cohort studies. BMC Cancer 2011;11:96.
[69] Pi-Sunyer FX. Glycemic index and disease. Am J Clin Nutr 2002
Jul;76(1):290Se8S.
[70] Wilcox G. Insulin and insulin resistance. Clin Biochem Rev 2005
May;26(2):19e39.
[71] Hjartaker A, Aagnes B, Robsahm TE, Langseth H, Bray F, Larsen IK.
Subsite-specific dietary risk factors for colorectal cancer: a re-
view of cohort studies. J Oncol 2013;2013:703854.
[72] Wei EK, Giovannucci E, Wu K, Rosner B, Fuchs CS, Willett WC,
et al. Comparison of risk factors for colon and rectal cancer. Int J
Cancer 2004 Jan 20;108(3):433e42.
[73] Domper Arnal MJ, Ferrandez AA, Lanas AA. Esophageal cancer: risk
factors, screening and endoscopic treatment in Western and Eastern
countries. World J Gastroenterol 2015 Jul 14;21(26):7933e43.
[74] Augustin LS, Gallus S, Negri E, La VC. Glycemic index, glycemic
load and risk of gastric cancer. Ann Oncol 2004 Apr;15(4):581e4.
[75] Nothlings U, Murphy SP, Wilkens LR, Henderson BE, Kolonel LN.
Dietary glycemic load, added sugars, and carbohydrates as risk
factors for pancreatic cancer: the Multiethnic Cohort Study. Am J
Clin Nutr 2007 Nov;86(5):1495e501.
[76] Freedman LS, Potischman N, Kipnis V, Midthune D, Schatzkin A,
Thompson FE, et al. A comparison of two dietary instruments for
evaluating the fat-breast cancer relationship. Int J Epidemiol
2006 Aug;35(4):1011e21.
[77] Wirfalt E, McTaggart A, Pala V, Gullberg B, Frasca G, Panico S,
et al. Food sources of carbohydrates in a European cohort of
adults. Public Health Nutr 2002 Dec;5(6B):1197e215.
[78] O’Neil CE, Keast DR, Fulgoni VL, Nicklas TA. Food sources of en-
ergy and nutrients among adults in the US: NHANES 2003-2006.
Nutrients 2012 Dec;4(12):2097e120.