Baseline Dietary Glutamic Acid Intake and the Risk of
Colorectal Cancer: The Rotterdam Study
Gilson G. Viana Veloso, MD
1
; Oscar H. Franco, MD, PhD
1
; Rikje Ruiter, MD, PhD
1,2
; Catherina E. de Keyser, MD, PhD
1
;
Albert Hofman, MD, PhD
1
; Bruno C. Stricker, MD, PhD
1,3,4
; and Jessica C. Kiefte-de Jong, RD, PhD
1,5
BACKGROUND: Animal studies have shown that glutamine supplementation may decrease colon carcinogenesis, but any relation
with glutamine or its precursors has not been studied in humans. The primary aim of this study was to assess whether dietary glu-
tamic acid intake was associated with colorectal can cer (CRC) risk in community-dwelling adults. A secondary aim was to evaluate
whether the association could be modified by the body mass index (BMI). METHODS: This study was embedded in the Rotterdam
study, which included a prospective cohort from 1990 onward that consisted of 5362 subjects who were 55 years old or older and
were free of CRC at the baseline. Glutamic acid was calculated as a percentage of the total protein intake with a validated food fre-
quency questionnaire at the baseline. Incident cases of CRC were pathology-based. RESULTS: During follow-up, 242 subjects devel-
oped CRC. Baseline dietary glutamic acid intake was significantly associated with a lower risk of developing CRC (hazard ratio [HR] per
percent increase in glutamic acid of protein, 0.78; 95% confidence interval [CI], 0.62-0.99). After stratification for BMI, the risk reduction
for CRC by dietary glutamic acid was 42% for participants with a BMI 25 kg/m
2
(HR per percent increase in glutamic acid of protein,
0.58; 95% CI, 0.40-0.85), whereas no association was found in participants with a BMI > 25 kg/m
2
(HR per percent increase in glutamic
acid of protein, 0.97; 95% CI, 0.73-1.31). CONCLUSIONS: Our data suggest that baseline dietary glutamic acid intake is associated
with a lower risk of developing CRC, but this association may be mainly present in nonoverweight subjects. Cancer 2016;122:899-907.
V
C
2015 American Cancer Society.
KEYWORDS: colorectal cancer, epidemiology, glutamic acid, glutamine.
INTRODUCTION
Colorectal cancer (CRC) is a major public health problem: globally, nearly 1.2 million new cases of CRC are diagnosed
every year, and the majority occur in Western countries.
1,2
A large body of evidence has shown that poor diet and lifestyle habits increase the risk of CRC.
3,4
For example, die-
tary intake of red and processed meat,
5,6
alcohol intake,
3
and body fatness have been linked to an increased risk of develop-
ing CRC.
7,8
In contrast, dietary protein intake has been associated with a decreased risk of CRC in some studies,
9
but results are
still inconsistent and may depend on differences in food sources rich in protein, such as meat versus fish.
9
The mechanisms
linking protein intake to CRC are largely unknown, but one of the hypotheses is that it may be explained by specific amino
acids.
One of the amino acids that may be of interest in CRC etiology is glutamic acid. Foods sources that include glutamic
acid are plant and animal protein sources such as beef, pork, poultry, milk, soy, cheese, spinach, and cabbage.
10
Glutamic
acid can form glutamine and is one of the most common amino acids in plasma.
11
Glutamine has been ascribed different
roles in the gastrointestinal tract. For instance, it stimulates crypt cell proliferation in the gut
12-14
and acts as a trophic
factor and as a protective factor for the intestinal mucosa.
12,14
Also, it serves as a precursor for nucleotide synthesis in
Corresponding author: Bruno C. Stricker, MD, PhD, Department of Epidemiology, Erasmus Medical Center, PO Box 2040, 3000 CA, Rotterdam, the Netherlands;
b.stricker@erasmusmc.nl
1
Department of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands;
2
Department of Internal Medicine, Groene Hart Hospital, Gouda, the
Netherlands;
3
Health Care Inspectorate, the Hague, the Netherlands;
4
Department of Internal Medicine, Erasmus Medical Center, Rotterdam, the Netherlands;
5
Leiden University College, the Hague, the Netherlands.
Gilson Gabriel Viana Veloso and Jessica C. Kiefte-de Jong analyzed the data and wrote the manuscript. Oscar H. Franco, Jessica C. Kiefte-de Jong, Albert Hofman,
and Bruno C. Stricker designed the study. Rikje Ruiter, Catherina E. de Keyser, and Bruno C. Stricker collected data on cancer incidence. Jessica C. Kiefte-de Jong
had primary responsibility for the final content. All authors read and approved the final manuscript.
The authors thank Wieke Altorf-van der Kuil (Department of Human Nutrition, Wageningen University, Wageningen, the Netherlands) for her contribution to the
calculations of dietary amino acid intake.
DOI: 10.1002/cncr.29862, Received: August 13, 2015; Revised: December 2, 2015; Accepted: December 7, 2015, Published online December 30, 2015 in Wiley
Online Library (wileyonlinelibrary.com)
Cancer March 15, 2016 899
Original Article
enterocytes.
15
In addition, a previous study found that
glutamine supplementation may reduce chronic bowel
inflammation through suppression of cytokines in mouse
models with colitis-associated tumors.
16
Another potential pathway through which glutamic
acid may play a role in the etiology of CRC is body weight
control. In addition, obesity and weight gain have been
found to be associated with CRC risk,
8,17
but supplemen-
tation with a polymer of glutamic acid or glutamine has
been reported to be involved in weight control in
humans.
18,19
However, whether glutamic acid may also
have a role in CRC prevention in healthy human individ-
uals is not well established. Therefore, the primary aim
was to study the association between dietary glutamic acid
intake and CRC risk in a population-based cohort study
of adults who were 55 years old or older. A secondary aim
was to assess to what extent the association could be modi-
fied by the body mass index (BMI).
MATERIALS AND METHODS
Rotterdam Study
This study was embedded in the Rotterdam study, a
population-based, prospective cohort of subjects aged 45
years or older and living in Ommoord, a suburb in Rotter-
dam, the Netherlands.
20
The Rotterdam study comprises
3 cohorts: the first cohort included subjects who were 55
years old or older and started in 1990, the second cohort
started in 2000 and included subjects who were 50 years
old, and the third cohort started in 2007 and included
subjects who were 45 years old or older. For the current
study, the first Rotterdam study cohort (n 5 7983) that
started in 1990 and included subjects who were 55 years
old or older was used because this cohort included suffi-
cient follow-up on cancer and had dietary data available
for the baseline.
Participants were interviewed at home and were physi-
callyexaminedattheresearchcenterevery3to4years.The
medical ethics committee of the Erasmus Medical Center
(Rotterdam, the Netherlands) approved the study, and all
participants provided written informed consent.
21
Dietary Data Collection
A food frequency questionnaire (FFQ) was used to collect
dietary data from participants at the baseline. The FFQ
consisted of 170 food items, and participants were asked
to complete the questionnaire with information about the
frequency of consumption of each food and the number
of servings as described in detail previously.
22
To mini-
mize loss of information, the dietary assessment included
a 2-step protocol: a self-administered questionnaire that
included a checklist of foods that the subjects had con-
sumed at least twice per month in the previous year and
then a standardized interview based on the checklist with
a trained dietician using the computerized, validated
FFQ. For each item, the frequency was recorded as times
per day, week, or month. The number of servings per fre-
quency was expressed in natural units (eg, slice of bread),
household measures (eg, cup), or grams (eg, cooked
vegetables). Data collected from the FFQ were then ana-
lyzed to calculate macronutrient intake according to the
Dutch Food Composition Table.
23
To calculate the die-
tary amino acid content, the Dutch Food Composition
Table of 1993 was extended by the addition of data on
amino acid content from McCance and Widdowson’s
food composition table, which is based on chemically
analyzed amino acids of 150 foods,
10
as described in detail
previously.
24
For this study, the following dietary variables were
taken into account in the analyses: energy intake (kcal/
day), total amino acid and glutamic acid intake (g/d), total
monounsaturated fat intake (g/d), total polyunsaturated
fat intake (g/d), total saturated fat intake (g/d), total proc-
essed meat intake (g/d), total unprocessed meat intake (g/
d), total magnesium intake (g/d), total dietary fiber intake
(g/d), and total polysaccharide intake (g/d).
Because dietary glutamic acid can be closely corre-
lated with other amino acids, dietary data on glutamic
acid intake were analyzed first through the determination
of the glutamic acid intake percentage of the total amino
acid intake. Subsequently, this percentage was adjusted
for the total energy intake by the residual method.
25
Because dietary data from the FFQ are best suited for
ranking individuals instead of analyzing absolute values
alone,
22
the energy-adjusted glutamic acid intake was
categorized into quartiles; the quartile that included the
median intake (quartile 2) was used as a reference for
further analyses.
Protein intake from the FFQ was validated against
urea excretion within a representative sample (n 5 80).
The Spearman correlation coefficient between protein
intake estimated from the FFQ and urea excretion was
0.67.
22
Other macronutrients and micronutrients were
validated against multiple food records and showed a cor-
relation between 0.4 and 0.8 after adjustments for age,
sex, total energy intake, and within-person variation.
22
CRC
Cancer cases were ascertained through 4 yearly follow-up
rounds in which all available data per participant were
collected from the general practitioners. In addition, the
Original Article
900 Cancer March 15, 2016
Rotterdam study is linked to Pathan, a local pathology
and cytology laboratory in Rotterdam that provides
pathology services to the hospitals in the Rotterdam area.
Last, the study is linked to the information database for
admissions as collected by Dutch Hospital Data. Two
research physicians independently assessed the diagnoses
of CRC on the basis of pathology data and medical
records. All events were classified according to Interna-
tional Classification of Diseases, Tenth Revision.
26
In the
case of a discrepancy, consensus was sought, or an oncolo-
gist decided. Only cases confirmed by pathology were
considered in the analyses. The index date was defined as
the earliest date found in the pathology reports. The CRC
status was ascertained from the baseline until December
31, 2011. Any CRC diagnosis before the baseline mea-
surement was excluded.
Covariates
Several other medical and lifestyle variables were consid-
ered as potential confounders or mediators. For lifestyle
habits, we considered alcohol intake, smoking status,
BMI, and physical activity levels. Alcohol intake (g/d) was
analyzed continuously. The smoking status was divided
into 2 categories: never or former smokers and current
smokers. BMI (kg/m
2
) was analyzed continuously and
was also categorized on the basis of cutoffs for overweight
(25.0 kg/m
2
) and obesity (30.0 kg/m
2
). Any use of
antidiabetic medication was defined on the basis of
Anatomical Therapeutic Chemical code A010.
Physical activity was assessed only during the third
visit to the research center by means of an adapted version
of the Zutphen Physical Activity Questionnaire.
27
The
questionnaire consisted of questions on walking, cycling,
gardening, diverse sports, hobbies, and housekeeping.
The total time spent on physical activity was calculated as
the sum of minutes per week for each type of activity;
thus, the weekly duration of physical activity was
obtained.
For sociodemographic variables, the financial status
and the education level were assessed. The financial status
was divided into 2 categories: low to middle income (net
income < e1090 per month) and middle to high income
(net income e1090 per month). The education level
was divided into 2 categories as well: primary education
solely and secondary education or higher.
Population of Analysis
From the first Rotterdam study cohort (n 5 7983), sub-
jects with a history of CRC (n 5 2) as well as subjects
without dietary data (n 5 2619 or 32.8%) were excluded.
Dietary data were missing when individuals participated
in the pilot phase of the Rotterdam study cohort (between
1989 and 1990), individuals were institutionalized, or the
research dietician considered the dietary data unreliable
(eg, when subjects had difficulties with recall of their food
intake or when dementia was suspected). No difference in
CRC risk over time was observed in subjects with and
without dietary data (hazard ratio [HR], 1.23; 95% confi-
dence interval [CI], 0.92-1.67). The final population of
the analysis consisted of 5362 subjects.
Statistical Analysis
Cox regression was performed with quartiles of energy-
adjusted glutamic acid intake as the independent variable
and CRC diagnosis as the dependent variable. As the
underlying timescale, the follow-up time (in years) was
used. Participants were followed until a CRC diagnosis,
death, or the end of follow-up (December 31, 2011),
whichever occurred first.
Crude (age- and sex-adjusted) and multivariate
Cox p roportional hazards analyses were performed to
evaluate the association between glutamic acid intake
and CRC risk. Potential confounders were added
stepwise to the crude model. The selection of potential
confounders for the multivariate models was based on
previous literature as well as changes in the effect esti-
mate of more than 10%.
28
In addition, the presence of
effect modification by BMI was evaluated by the addi-
tion of a product term of glutamic acid times BMI to the
crude and multivariate models, and an analysis stratified
by BMI was performed when effect m odification was
present.
Nonlinear relations between glutamic acid intake
and CRC risk were assessed by the inclusion of a quadratic
term of glutamic acid intake.
Several sensitivity analyses were performed. To
assess potential reverse causality, we excluded CRC cases
in which the diagnosis occurred within the first 2 and 5
years of follow-up. To assess whether the association was
different for colon cancer versus rectal cancer, we reran
analyses by splitting colon cancer and rectal cancer
(including the rectum and rectosigmoid junction).
To reduce a potential attrition bias, we performed a
multiple imputation procedure for missing covariates
(5 imputations; Tables 1 and 2).
29
Results of the Cox
regression models are presented as HRs and 95% CIs. A
P value .05 was considered to be statistically significant.
All analyses were performed with IBM SPSS Statistics
(SPSS, version 21.0; SPSS Inc, Chicago, Ill).
Dietary Glutamic Acid and Colorectal Cancer/Veloso et al
Cancer March 15, 2016 901
RESULTS
Population Characteristics
Baseline characteristics are shown in Table 3. During the
median follow-up period of 16 years (range, 21 years),
242 subjects developed CRC (104 men and 138 women);
120 (50%) had colon cancer, 117 (48%) had rectal can-
cer, and 5 (2%) had other types of CRC.
Themedianintakeofglutamicacidintakewas
16 g/d (range, 4-46 g/d), which was a median of 20%
(range, 17%-27%) of the total amino acid intake.
The main food sources of dietary glutamic acid were
dairy, fish, meat, and grains; this explained 57% of the
total variance in glutamic acid of the population.
Participants with a high intake of glutamic acid had
a lower intake of alcohol and a lower intake of other
amino acids and were less often smokers (Table 3).
Dietary Glutamic Acid Intake and CRC Risk
Associations between the baseline dietary glutamic acid
intake and the CRC risk are presented in Table 4. A higher
glutamic acid level was associated with a lower risk of CRC
(HR per percent increase in glutamic acid of protein, 0.79;
95% CI, 0.62-0.99; Table 4) after adjustments for dietary,
lifestyle, and socioeconomic confounders. Additional adjust-
ments for BMI did not alter this result (Table 4). With
respect to the median glutamic acid intake, subjects with a
high intake of glutamic acid had a 50% lower risk of CRC
(HR, 0.51; 95% CI, 0.30-0.87; Table 4). No evidence for a
nonlinear relation was found (P
quadratic term
> .05).
Subgroup and Sensitivity Analysis
The association between baseline dietary glutamic acid
and CRC risk was modified by BMI (P
interaction
5 .05).
After stratification by BMI, a significant association
between dietary glutamic acid intake and CRC was more
prominent in subjects with a normal BMI (Table 5) but
was not in subjects who were overweight with a
BMI > 25 kg/m
2
(Table 5).
HRs did not differ significantly after the exclusion
of subjects from the analyses who developed CRC after
less than 2 (n 5 25) and 5 years of follow-up (n 5 62; Ta-
ble 6). Separating the outcome by colon cancer and rectal
cancer revealed similar associations for colon and rectal
cancer (HR for colon cancer, 0.85; 95% CI, 0.74-0.97;
HR for rectal cancer, 0.70; 95% CI, 0.61-0.80 [per per-
cent increase in glutamic acid of protein after adjustments
for age, sex, alcohol intake, physical activity, smoking,
family history of cancer, dietary fiber intake, dietary mag-
nesium intake, unprocessed meat intake, dietary
TABLE 2. Comparison of Variables With Missing
Data Before and After the Imputation Procedure
Variable
Original Data
(n 5 5362),
Valid No. (%)
After Multiple
Imputation Procedure
(n 5 5362), No. (%)
Smoking status
Never or former 4080 (76.1) 4103 (76.5)
Current 1250 (23.3) 1259 (23.5)
Missing 32 (0.6)
Body mass index (kg/m
2
)
Valid No. 5328 (99.4) 5362 (100)
Missing 34 (0.6)
Family history of cancer
Yes 2218 (41.4) 2759 (51.4)
No 2074 (38.6) 2603 (48.6)
Missing 1070 (20)
Education level
Primary 2782 (51.9) 2796 (52.1)
Secondary or higher 2553 (47.6) 2566 (47.9)
Missing 27 (0.5)
Financial status
Low to middle income 130 (2.4) 145 (2.7)
Middle to high income 4716 (88) 5217 (97.3)
Missing 516 (9.6)
Physical activity (h/wk)
Below the median 1924 (35.9) 2844 (53)
Above the median 1920 (35.8) 2518 (47)
Missing 1518 (28.3)
TABLE 1. Details of the Multiple Imputation Procedure
Software SPSS 21.0 for Windows
Imputation method Custom, fully conditional specification (Markov chain Monte Carlo);
10 maximum interactions
No. of imputations 5
Variables with missing data included in the imputation procedure Smoking status, body mass index, family his tory of cancer,
education level, financial status, physical activity
Variables without missing data but added as predictors for the
imputation to improve the missing-at-random assumption
Age at start, sex, total energy intake, total glutamic acid intake,
total amino acid intake minus glutamic acid intake, total processed
meat intake, total unprocessed meat intake, total monounsaturated
fat intake, total polyunsaturated fat intake, total saturated fat intake,
total dietary fiber intake, total dietary magnesium intake,
total polysaccharide intake
Treatment of continuous variables Predictive mean matching
Treatment of categorical variables Logistic regression
Original Article
902 Cancer March 15, 2016
polysaccharide intake, financial status, smoking status,
and BMI]).
DISCUSSION
This study showed that a low intake of baseline glutamic
acid was associated with a higher risk of developing CRC
among community-dwelling adults. This association was
modified by BMI and differed for those with a higher
BMI (>25 kg/m
2
) versus those with a normal BMI
(25 kg/m
2
).
This is the first population-based study evaluating
the potential role of glutamic acid in the risk of developing
CRC. Previous animal and laboratory studies have pro-
vided some evidence showing that glutamic acid may play
a role in carcinogenesis and subsequent prevention of
CRC.
The amino acid glutamic acid is closely related to
glutamine. The human body is able to produce
L-gluta-
mine itself from
L-glutamic acid.
30
Glutamic acid and glu-
tamine are considered to be the most abundant amino
TABLE 3. Baseline Characteristics
Energy-Adjusted Glutamic Acid Intake
First Quartile (Median,
19% of Protein;
n 5 1340)
Second Quartile
(Median, 20% of
Protein; n 5 1341)
Third Quartile
(Median, 21% of
Protein; n 5 1341)
Fourth Quartile
(Median, 22% of
Protein; n 5 1340)
Age at start, mean (SD), y 67 (8) 67 (8) 67 (8) 68 (8)
Sex, No. (%)
Male 596 (44) 553 (41) 534 (40) 523 (39)
Female 744 (56) 788 (59) 807 (60) 817 (61)
Alcohol intake, median (range), g/d 9 (145) 5 (99) 2 (105) 1 (97)
Smoking status, No. (%)
Never or former 946 (70.5) 1023 (76.2) 1052 (78.4) 1059 (79)
Current 387 (29) 307 (23) 281 (21) 275 (20.5)
Missing 7 (0.5) 11 (0.8) 8 (0.6) 6 (0.5)
Body mass index, mean (SD), kg/m
2
27 (3) 26 (4) 26 (3) 26 (4)
Missing, No. (%) 11 (0.2) 9 (0.2) 4 (0.1) 10 (0.2)
Family history of cancer, No. (%)
Yes 515 (38) 561 (42) 591 (44) 551 (41)
No 559 (42) 523 (39) 498 (37) 494 (37)
Missing 266 (20) 257 (19) 252 (19) 295 (22)
Education level, No. (%)
Primary 688 (51.6) 676 (50.4) 709 (53) 709 (53)
Secondary or higher 647 (48) 654 (48.8) 626 (47.6) 626 (46.6)
Missing 5 (0.4) 11 (0.8) 6 (0.4) 5 (0.4)
Financial status, No. (%)
Low to middle income 34 (2.5) 28 (2) 34 (2.5) 34 (2.5)
Middle to high income 1180 (88) 1180 (88) 1181 (88) 1175 (88)
Missing 126 (9.5) 133 (10) 126 (9.5) 131 (9.5)
Physical activity levels,
a
No. (%)
Below the median 487 (36) 477 (36) 477 (36) 483 (36)
Above the median 474 (35) 508 (38) 500 (37) 438 (33)
Missing 379 (29) 356 (26) 364 (27) 419 (31)
Use of antidiabetic medication, No. (%)
Yes 46 (3.4). 53 (4) 54 (4.0) 54 (4)
No 1293 (96.5) 1288 (96) 1286 (95.9) 1284 (95.8)
Missing 1 (0.) 1 (0.1) 2 (0.1)
Total energy intake, mean (SD), kcal/day 1965 (556) 1965 (472) 2010 (506) 1972 (467)
Total amino acid intak e minus
glutamic acid, mean (SD), g/d
86 (23) 84 (19) 84 (19) 78 (18)
Polyunsaturated fat intake, mean (SD), g/d 15 (8) 15 (7) 16 (8) 15 (7)
Monounsaturated fat intake, mean (SD), g/d 29 (12) 28 (10) 27 (10) 26 (9)
Saturated fat intake, mean (SD), g/d 32 (13) 32 (11) 32 (12) 31 (11)
Total fiber intake, mean (SD), g/d 16 (6) 17 (5) 17 (5) 16 (5)
Processed meat intake, mean (SD), servings/d 1.5 (1.2) 1.5 (1.2) 1.5 (1.2) 1.5 (1.4)
Unprocessed meat intake,
mean (SD), servings/d
1.0 (0.6) 0.8 (0.4) 0.7 (0.3) 0.5 (0.3)
Total polysaccharide intake, mean (SD), g/d 91 (30) 100 (27) 111 (30) 118 (30)
Total magnesium intake, mean (SD), g/d 300 (82) 310 (71) 315 (75) 299 (70)
Abbreviation: SD, standard deviation.
a
Measured during the third visit at the research center (not baseline)
Dietary Glutamic Acid and Colorectal Cancer/Veloso et al
Cancer March 15, 2016 903
acids in the body.
31
Glutamine has been shown to play a
role in protecting cells from inflammation and oxidative
stress,
32
and it is preferred as fuel for many cells, including
enterocytes
33
and colonocytes.
34
For the intestinal mucosa, the role of glutamine as a
trophic factor is well described.
35
Glutamine repairs the
epithelial layer by preserving mucosal integrity and main-
tains bowel barrier functions by reducing permeability
36
and bacterial translocation.
37
Moreover, glutamine has
been shown to have several anti-inflammatory activities.
32
In addition, the relation between inflammation and carci-
nogenesis of the colon has been hypothesized previously
by studies showing neoplastic transformation in subjects
with inflammatory bowel disease.
38
If we combine these latter mechanisms, it can be
hypothesized that glutamic acid may have a protective
role through glutamine against the development of CRC.
Previous mouse models have shown that glutamine pre-
vents the progression of colitis-associated CRC.
16
Also, it
has been recently demonstrated that glutaminase, an
enzyme that converts glutamine to glutamate, suppresses
proliferation and induces apoptosis in cell lines of colo-
rectal adenomas.
39
In our study, we evaluated the interaction between
glutamic acid intake and BMI with respect to CRC risk
because BMI is an established risk factor for CRC.
8
We
demonstrated that the protective effect of glutamic acid
against CRC risk may exist only for participants with a
BMI 25 kg/m
2
. It may be speculated that being over-
weight dilutes any potentially protective effect of glutamic
acid on CRC risk. Another explanation might be that the
association between glutamic acid and CRC is mediated
by body weight control. In addition, it has been shown
that both glutamic acid and glutamine may play a role in
weight reduction.
18,19
It can be argued that the association between dietary
glutamic acid and CRC risk may be explained by an
increased demand for glutamine due to early malignant
processes in the colon. In addition, tumor progression is
associated with an increased demand for glutamine from
tumor cells.
40
As a result, depression of the activity of
other immune cells may occur because of decreased gluta-
thione concentrations.
41
To account for this potential
reverse causality, we excluded CRC cases at the baseline.
Also, the exclusion of CRC cases after 2 and 5 years of
follow-up did not show different effect estimates; this sug-
gests that the influence of reverse causality in the relation
between glutamic acid and CRC may be limited.
The main strengths of our study are the population-
based setting, which strengthens the generalizability of
our results, and the prospective study design, which mini-
mizes the recall bias potentially associated with CRC.
Moreover, this is the first study on glutamic acid and
CRC in a healthy population. Most studies have been per-
formed with animals, cell lines, or CRC patients, all of
which may have limited generalizability.
To appreciate the findings of this study, limitations
must be taken into consideration. Although our FFQ did
have fairly good agreement with protein intake assessed
from 24-hour urine testing (r 5 0.67), we were not able to
validate the FFQ for glutamic acid specifically. Hence,
our FFQ could still be prone to measurement error.
42
By
adjusting for the total energy intake, we aimed to reduce
the magnitude of the systematic measurement error.
TABLE 4. Dietary Glutamic Acid Intake and Colorectal Cancer Risk
Hazard Ratio (95% Confidence Interval)
Crude
a
Multivariate 1
b
Multivariate 2
c
Energy-adjusted glutamic acid
(continuously per % of protein)
0.87 (0.76-0.99)
d
0.79 (0.62-0.99)
d
0.78 (0.62-0.99)
d
Energy-adjusted glutamic acid (categorical)
Quartile 1 0.84 (0.60-1.18) 0.85 (0.54-1.33) 0.84 (0.53-1.30)
Quartile 2 Reference Reference Reference
Quartile 3 0.77 (0.54-1.08) 0.81 (0.52-1.26) 0.80 (0.52-1.24)
Quartile 4 0.67 (0.47-0.96)
d
0.52 (0.31-0.88)
d
0.51 (0.30-0.87)
d
a
The crude model was adjusted for age (continuously) and sex (male and female).
b
This model included the crude model plus additional adjustments for the following: alcohol intake (continuously), physical activity (continuously), family history
of cancer (yes vs no), dietary fiber intake (continuously), dietary magnesium intake (continuously), unprocessed meat intake (continuous ly), dietary polysaccha-
ride intake (continuously), smoking status (current vs former/no), and financial status (low-middle vs middle-high). Additional adjustments for education level,
saturated fat intake, monounsaturated fat, polyunsaturated fat, processed meat intake, and use of antidiabetic medication did not change the results by more
than 10%.
c
This model included multivariate model 1 plus an additional adjustment for the body mass index (continuously).
d
P <.05.
Original Article
904 Cancer March 15, 2016
However, a random measurement error still may have led
to an underestimation of dietary intake. Although it has
been reported that diet-disease associations are biased to-
ward the null when a random error of dietary intake is
present,
42
residual confounding by other dietary variables
due to measurement error may still have been present and
may have led to an overestimation of our associations. We
had only baseline dietary measurements available for our
study. However, in studies of diet and cancer, the dietary
assessment should address long-term intake because a
long latency period may be involved. We used baseline di-
etary data under the assumption that dietary habits remain
similar over time. Indeed, the latter has been confirmed
by another study using 5 annual repeated FFQs similar to
the one used in our study; it showed that the ranking of
subjects according to dietary intake remained fairly similar
over time.
43
This suggests that our single FFQ measure-
ment may be relevant for reflecting recent dietary habits
not only at the baseline but also over a longer period.
43
Nonetheless, repeated measurement of diet could reduce
the magnitude of any random error, and this could lead to
stronger or weaker associations than 1 dietary measure-
ment at the baseline or the most recent dietary intake.
44
Dietary data were not collected for those institution-
alized and for those with recall difficulties. As a result,
subjects without dietary data reflect a more diseased and
vulnerable study group. Although the CRC risk was not
different for subjects with dietary data and subjects with-
out dietary data and this decision was made a priori to
maintain the quality of the dietary data assessment, this
selection may have affected the generalizability of the
study results.
This study was observational in design, so conclu-
sions regarding the causality of the observed association
should be made with caution. In addition, although we
had data on the family history of cancer, we did not have
specific data on the family history of CRC or polyposis.
Individuals with a family history of CRC and polyposis
are at increased risk for developing CRC.
45
Although we
adjusted for any self-reported family history of cancer,
residual confounding by a family history of CRC and pol-
yposis may still be present. Also, we did not have any data
on chronic inflammatory bowel disease at the baseline,
and this has been found to be associated with a higher risk
of CRC
46
; this may have influenced our results.
At last, physical activity levels were assessed only
during the third visit to the research center and not at the
TABLE 6. Sensitivity Analysis
Hazard Ratio (95% Confidence Interval)
Crude
a
Multivariate 1
b
Multivariate 2
c
>2 y of follow-up:
glutamic acid %
of protein
(continuously)
0.85 (0.74-0.98)
d
0.77 (0.61-0.99)
d
0.77 (0.61-0.98)
d
>5 y of follow-up:
glutamic acid %
of protein
(continuously)
0.86 (0.73-1.00) 0.76 (0.63-0.93)
d
0.76 (0.59-0.97)
d
Cases with less than 2 and 5 years of follow-up were excluded.
a
The crude model was adjusted for age (continuously) and sex (male and
female).
b
This model included the crude model plus additional adjustments for the
following: alcohol intake (continuously), physical activity (continuously), fam-
ily history of cancer (yes vs no), dietary fiber intake (continuously), dietary
magnesium intake (continuously), unprocessed meat intake (continuously),
dietary polysaccharide intake (continuously), smoking status (current vs for-
mer/no), and financial status (low-middle vs middle-high). Additional adjust-
ments for education level, saturated fat intake, monounsaturated fat,
polyunsaturated fat, processed meat intake, and use of antidiabetic medi-
cation did not change the results by more than 10%.<zaq;12>
c
This model included multivariate model 1 plus an additional adjustment
for the body mass index (continuously).
d
P <.05.
TABLE 5. Dietary Glutamic Acid Intake and Colo-
rectal Cancer Risk Stratified by the Weight Status
Hazard Ratio (95% Confidence
Interval)
Crude
a
Multivariate
b
BMI < 25 kg/m
2
Energy-adjusted glutamic
acid (continuously
per % of protein)
0.69 (0.55-0.86)
11
0.58 (0.40-0.84)
11
Energy-adjusted glutamic
acid (categorical)
1.11 (0.65-1.91) 1.16 (0.55-2.47)
Quartile 1 Reference Reference
Quartile 2 0.64 (0.36-1.16) 0.73 (0.35-1.54)
Quartile 3 0.50 (0.27-0.92)
11
0.38 (0.15-0.93)
11
Quartile 4
BMI 25 kg/m
2
Energy-adjusted glutamic acid
(continuously per % of protein)
1.01 (0.85-1.20) 0.97 (0.73-1.31)
Energy-adjusted glutamic
acid (categorical)
0.71 (0.46-1.10) 0.68 (0.38-1.19)
Quartile 1 Reference Reference
Quartile 2 0.85 (0.56-1.31) 0.84 (0.49-1.45)
Quartile 3 0.80 (0.51-1.25) 0.60 (0.31-1.16)
Quartile 4
a
The crude model was adjusted for age (continuously) and sex (male and
female).
b
This model included the crude model plus additional adjustments for the
following: alcohol intake (continuously), physical activity (continuously), fam-
ily history of ca ncer (yes vs no), dietary fiber intake (continuously), dietary
magnesium intake (continuously), unprocessed meat intake (continuously),
dietary polysaccharide intake (continuously), smoking status (current vs
former/no), and financial status (low-middle vs middle-high). Additional
adjustments for education level, saturated fat intake, monounsaturated
fat, polyunsaturated fat, processed meat intake, and use of antidiabetic
medication did not change the results by more than 10%.
c
P <.05.
Dietary Glutamic Acid and Colorectal Cancer/Veloso et al
Cancer March 15, 2016 905
baseline. Hence, residual confounding by physical activity
levels may still be partly present because it has been sug-
gested that physical activity may also play a role in CRC
risk.
3
At last, although we did not find any indication for
reverse causality by excluding 2 and 5 years of follow-up,
there is evidence that CRC may develop over a 10-year
interval,
47
and this still may have influenced our results.
We were not able to study the association between
glutamic acid intake and right-sided colon cancer versus
left-sided colon cancer, although it has been suggested
that diet as well as BMI may have different effects on left-
sided lesions versus right-sided lesions.
48
This suggests
different etiologies for left-sided and right-sided colon
cancer.
In conclusion, this is the first study showing in a
population-based setting that a high baseline intake of
dietary glutamic acid might be associated with a lower risk
for CRC. This relation may be modified by BMI and
needs further replication in other population-based
studies.
FUNDING SUPPORT
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
Gilson G. Viana Veloso, Jessica C. Kiefte-de Jong, and Oscar H.
Franco work for ErasmusAGE, a center for aging research across the
course of life that is funded by Nestle Nutrition (Nestec, Ltd),
Metagenics, Inc, and AXA. These funding sources had no role in
the design or conduct of the study; in the collection, management,
analysis, or interpretation of the data; or in the preparation, review,
or approval of this article.
REFERENCES
1. Ferlay J, Soerjomatara m I, Ervik M, et al. GLOBOCAN 2012 v1.0,
Cancer Incidence and Mortality Worldwide. Lyon, France: International
Agency for Research on Cancer; 2013. IARC CancerBase; No. 11
2. Boyle P, Leon ME. Epidemiology of colorectal cancer. Br Med Bull.
2002;64:1-25.
3. World Cancer Research Fund, American Institute for Cancer
Research. Colorectal Cancer 2011 Report: Food, Nutrition, Physical
Activity, and the Prevention of Colorectal Cancer. http://www.wcrf.
org/sites/default/files/Colorectal-Cancer-2011-Report.pdf. Accessed
August 12, 2015.
4. Huxley RR, Ansary-Moghaddam A, Clifton P, Czernichow S, Parr
CL, Woodward M. The impact of dietary and lifestyle risk factors
on risk of colorectal cancer: a quantitative overview of the epidemio-
logical evidence. Int J Cancer. 2009;125:171-180.
5. Corpet DE. Red meat and colon cancer: should we become vegeta-
rians, or can we make meat safer? Meat Sci. 2011;89:310-316.
6. Kim E, Coelho D, Blachier F. Review of the association between
meat consumption and risk of colorectal cancer. Nutr Res. 2013;33:
983-994.
7. Giovannucci E. Obesity, gender, and colon cancer. Gut. 2002;51:
147.
8. Larsson SC, Wolk A. Obesity and colon and rectal cancer risk: a
meta-analysis of prospective studies. Am J Clin Nutr. 2007;86:
556-565.
9. Williams CD, Satia JA, Adair LS, et al. Associations of red meat,
fat, and protein intake with distal colorectal cancer risk. Nutr Can-
cer. 2010;62:701-709.
10. Paul AA, Southgate DAT, Russell J. First Supplement to McCance
and Widdowson’s the Composition of Foods: Amino Acid Composi-
tion (mg per 100 g Food) and Fatty Acid Composition (g per 100 g
Food). Amsterdam, the Netherlands: Elsevier/North-Holland Bio-
medical Press; 1980.
11. Buchman AL. Glutamine: commercially essential or conditionally
essential?. A critical appraisal of the human data. Am J Clin Nutr.
2001;74:25-32.
12. Turowski GA, Rashid Z, Hong F, Madri JA, Basson MD. Gluta-
mine modulates phenotype and stimulates proliferation in human
colon cancer cell lines. Cancer Res. 1994;54:5974-5980.
13. Scheppach W, Loges C, Bartram P, et al. Effect of free glutamine
and alanyl-glutamine dipeptide on mucosal proliferation of the
human ileum and colon. Gastroenterology. 1994;107:429-434.
14. DeBerardinis RJ, Cheng T. Q’s next: the diverse functions of gluta-
mine in metabolism, cell biology and cancer. Oncogene. 2010;29:
313-324.
15. Watford M. Glutamine metabolism and function in relation to pro-
line synthesis and the safety of glutamine and proline supplementa-
tion. J Nutr. 2008;138:2003S-2007S.
16. Tian Y, Wang K, Wang Z, Li N, Ji G. Chemopreventive effect of
dietary glutamine on colitis-associated colon tumorigenesis in mice.
Carcinogenesis. 2013;34:1593-1600.
17. Schlesinger S, Lieb W, Koch M, et al. Body weight gain and risk of
colorectal cancer: a systematic review and meta-analysis of observa-
tional studies. Obes Rev. 2015;16:607-619.
18. Park JH, Choi JC, Sung MH, Kang JH, Chang MJ. High molecular
weight poly-gamma-glutamic acid regulates lipid metabolism in rats fed
a high-fat diet and humans. J Microbiol Biotechnol. 2011;21:766-775.
19. Laviano A, Molfino A, Lacaria MT, et al. Glutamine supplementa-
tion favors weight loss in nondieting obese female patients. A pilot
study. Eur J Clin Nutr. 2014;68:1264-1266.
20. Hofman A, Brusselle GG, Darwish Murad S, et al. The Rotterdam
study: 2016 objectives and design update. Eur J Epidemiol. 2015;30:
661-708.
21. Hofman A, Darwish Murad S, van Duijn CM, et al. The Rotterdam
study: 2014 objectives and design update. Eur J Epidemiol. 2013;28:
889-926.
22. Klipstein-Grobusch K, den Breeijen JH, Goldbohm RA, et al. Die-
tary assessment in the elderly: validation of a semiquantitative food
frequency questionnaire. Eur J Clin Nutr. 1998;52:588-596.
23. Dutch Food Composition Table (NEVO) [in Dutch]. The Hague,
the Netherlands: Voorlichtingsbureau Voor de Voeding; 1996.
24. Altorf-van der Kuil W, Engberink MF, De Neve M, et al. Dietary
amino acids and the risk of hypertension in a Dutch older popula-
tion: the Rotterdam study. Am J Clin Nutr. 2013;97:403-410.
25. Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake
in epidemiologic studies. Am J Clin Nutr. 1997;65:1220S-1228S.
26. World Health Organization. International Statistical Classification of
Diseases and Related Health Problems. Vol 1. 10th rev. Geneva,
Switzerland: World Health Organization; 1992.
27. Caspersen CJ, Bloemberg BP, Saris WH, Merritt RK, Kromhout D.
The prevalence of selected physical activities and their relation with
coronary heart disease risk factors in elderly men: the Zutphen study,
1985. Am J Epidemiol. 1991;133:1078-1092.
28. Mickey RM, Greenland S. The impact of confounder selection crite-
ria on effect estimation. Am J Epidemiol. 1989;129:125-137.
29. Sterne JA, White IR, Carlin JB, et al. Multiple imputation for miss-
ing data in epidemiological and clinical research: potential and pit-
falls. BMJ. 2009;338:b2393.
30. Wallace C, Keast D. Glutamine and macrophage function. Metabo-
lism. 1992;41:1016-1020.
31. Tapiero H, Mathe G, Couvreur P, Tew KD. II. Glutamine and glu-
tamate. Biomed Pharmacother. 2002;56:446-457.
32. Cruzat VF, Bittencourt A, Scomazzon SP, Leite JS, de Bittencourt
PI Jr, Tirapegui J. Oral free and dipeptide forms of glutamine sup-
plementation attenuate oxidative stress and inflammation induced by
endotoxemia. Nutrition. 2014;30:602-611.
Original Article
906 Cancer March 15, 2016
33. Oba M, Baldwin RL IV, Bequette BJ. Oxidation of glucose, gluta-
mate, and glutamine by isolated ovine enterocytes in vitro is
decreased by the presence of other metabolic fuels. J Anim Sci.
2004;82:479-486.
34. Blachier F, Boutry C, Bos C, Tome D. Metabolism and functions
of
L-glutamate in the epithelial cells of the small and large intestines.
Am J Clin Nutr. 2009;90:814S-821S.
35. Wang WW, Qiao SY, Li DF. Amino acids and gut function. Amino
Acids. 2009;37:105-110.
36. Wang B, Wu G, Zhou Z, et al. Glutamine and intestinal barrier
function. Amino Acids. 2015;47:2143-2154.
37. Ren W, Duan J, Yin J, et al. Dietary
L-glutamine supplementation
modulates microbial community and activates innate immunity in
the mouse intestine. Amino Acids. 2014;46:2403-2413.
38. Rutter MD, Riddell RH. Colorectal dysplasia in inflammatory bowel
disease: a clinicopathologic perspective. Clin Gastroenterol Hepatol.
2014;12:359-367.
39. Huang F, Zhang Q, Ma H, Lv Q, Zhang T. Expression of gluta-
minase is upregulated in colorectal cancer and of clinical significance.
Int J Clin Exp Pathol. 2014;7:1093-1100.
40. Medina MA. Glutamine and cancer. JNutr. 2001;131:2539S-
2542S.
41. Gamcsik MP, Kasibhatla MS, Teeter SD, Colvin OM. Glutathione
levels in human tumors. Biomarkers. 2012;17:671-691.
42. Freedman LS, Schatzkin A, Midthune D, Kipnis V. Dealing with di-
etary measurement error in nutritional cohort studies. J Natl Cancer
Inst. 2011;103:1086-1092.
43. Goldbohm RA, van’t Veer P, van den Brandt PA, et al. Reproduci-
bility of a food frequency questionnaire and stability of dietary hab-
its determined from five annually repeated measurements. Eur J Clin
Nutr. 1995;49:420-429.
44. Hu FB, Stampfer MJ, Rimm E, et al. Dietary fat and coronary heart
disease: a comparison of approaches for adjusting for total energy
intake and modeling repeated dietary measurements. Am J Epidemiol.
1999;149:531-540.
45. Amersi F, Agustin M, Ko CY. Colorectal cancer: epidemiology, risk
factors, and health services. Clin Colon Rectal Surg. 2005;18:133-140.
46. Triantafillidis JK, Nasioulas G, Kosmidis PA. Colorectal cancer and
inflammatory bowel disease: epidemiology, risk factors, mechanisms
of carcinogenesis and prevention strategies. Anticancer Res. 2009;29:
2727-2737.
47. Lieberman DA, Rex DK, Winawer SJ, et al. Guidelines for colono-
scopy surveillance after screening and polypectomy: a consensus
update by the US Multi-Society Task Force on Colorectal Cancer.
Gastroenterology. 2012;143:844-857.
48. Wallace K, Grau MV, Ahnen D, et al. The association of lifestyle
and dietary factors with the risk for serrated polyps of the colorec-
tum. Cancer Epidemiol Biomarkers Prev. 2009;18:2310-2317.
Dietary Glutamic Acid and Colorectal Cancer/Veloso et al
Cancer March 15, 2016 907