Association between Carbohydrate Intake and the Prevalence of Metabolic Syndrome in Korean Women

Abstract

Carbohydrates consist of a large proportion of calories in the Asian diet. Therefore, we aimed to investigate the association between carbohydrate intake and metabolic syndrome in Korean women. A cross-sectional analysis was conducted with a total of 4294 Korean women aged 40–69 years from the Korean Genomic and Epidemiology Study (KoGES). Carbohydrate intake was calculated based on a validated food frequency questionnaire. Metabolic syndrome was defined by using the National Cholesterol Education Program, Adult Treatment Panel III (NCEPIII). Logistic regression was used to estimate the association of carbohydrate intake with metabolic syndrome and its components. In this study, high carbohydrate intake seemed to be associated with low socioeconomic status and an imbalanced diet. After adjusting for confounding factors, subjects with higher carbohydrate intake showed an increased risk of metabolic syndrome (odds ratio (OR) 1.34, 95% confidence interval (CI) 1.08–1.66, p-trend = 0.004, highest vs. lowest quartile [≥75.2 vs. <67.0% of energy]), particularly elevated waist circumference. This association was stronger among those with low levels of C-reactive protein (CRP) and those with low dairy intake. In conclusion, higher carbohydrate intake is associated with a higher risk of metabolic syndrome, particularly abdominal obesity, in Korean women. This association may differ according to individuals’ CRP level and dairy intake.

Full text

nutrients
Article
Association between Carbohydrate Intake and the Prevalence of
Metabolic Syndrome in Korean Women
Young-Ae Cho and Jeong-Hwa Choi *


Citation: Cho, Y.-A.; Choi, J.-H.
Association between Carbohydrate
Intake and the Prevalence of
Metabolic Syndrome in Korean
Women. Nutrients 2021,13, 3098.
https://doi.org/10.3390/nu13093098
Academic Editor: David Rowlands
Received: 5 August 2021
Accepted: 1 September 2021
Published: 3 September 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Department of Food Science and Nutrition, Keimyung University, Daegu 42601, Korea; youngcho914@gmail.com
*Correspondence: jhchoi@kmu.ac.kr; Tel.: +82-53-580-5913
Abstract:
Carbohydrates consist of a large proportion of calories in the Asian diet. Therefore,
we aimed to investigate the association between carbohydrate intake and metabolic syndrome in
Korean women. A cross-sectional analysis was conducted with a total of 4294 Korean women aged
40–69 years
from the Korean Genomic and Epidemiology Study (KoGES). Carbohydrate intake was
calculated based on a validated food frequency questionnaire. Metabolic syndrome was defined by
using the National Cholesterol Education Program, Adult Treatment Panel III (NCEPIII). Logistic
regression was used to estimate the association of carbohydrate intake with metabolic syndrome
and its components. In this study, high carbohydrate intake seemed to be associated with low
socioeconomic status and an imbalanced diet. After adjusting for confounding factors, subjects with
higher carbohydrate intake showed an increased risk of metabolic syndrome (odds ratio (OR) 1.34,
95% confidence interval (CI) 1.08–1.66, p-trend = 0.004, highest vs. lowest quartile [
≥
75.2 vs. <67.0%
of energy]), particularly elevated waist circumference. This association was stronger among those
with low levels of C-reactive protein (CRP) and those with low dairy intake. In conclusion, higher
carbohydrate intake is associated with a higher risk of metabolic syndrome, particularly abdominal
obesity, in Korean women. This association may differ according to individuals’ CRP level and
dairy intake.
Keywords: carbohydrate; metabolic syndrome; women; obesity; dairy; C-reactive protein
1. Introduction
Metabolic syndrome is a cluster of interrelated abnormalities, including abdominal
obesity, insulin resistance, dysglycemia, hypertension, and dyslipidemia [
1
]. Metabolic
syndrome has become an important public health concern because of its association with
cardiovascular disease and type 2 diabetes [
2
]. Low-grade systemic inflammation is also
characteristic of metabolic syndrome. Increased levels of C-reactive protein (CRP) have
been reported to be associated with insulin resistance, adiposity, and other features of
metabolic syndrome [3,4].
It is difficult to directly compare the prevalence of metabolic syndrome in different
studies because of differences in criteria used to define metabolic syndrome, study designs,
and sample size [
2
]. Many dietary factors have been suggested to be associated with
metabolic syndrome [
5
,
6
]. It is difficult to explain the prevalence of metabolic syndrome
according to single nutrients or foods. However, Asians traditionally consume a lot of
rice as a staple food, thus obtaining a large proportion of calories from carbohydrates.
Therefore, carbohydrate intake could play an important role in metabolic abnormalities in
this population. Several studies have investigated the association between carbohydrate
intake and the risk of metabolic syndrome, and they have reported mixed findings [
7
–
10
].
Some of these studies reported a stronger association in the Asian population compared to
the non-Asian population [
7
,
8
], suggesting that this may be attributed to different amounts
of carbohydrate consumption. In a study by Ha et al., the proportion of energy from
carbohydrate was 80–82% in the highest quintile among the Korean adults, compared with
Nutrients 2021,13, 3098. https://doi.org/10.3390/nu13093098 https://www.mdpi.com/journal/nutrients

Nutrients 2021,13, 3098 2 of 12
64–65% in the US adults [
7
]. In addition, higher proportion of carbohydrate could result in
imbalanced diet. Furthermore, socioeconomic status (SES) has been suggested to contribute
to health status through the intake of macronutrients. Sakurai et al. reported that older age
and some aspects of SES (e.g., income and education levels) were associated with a high
carbohydrate/low fat intake [
11
]. Therefore, SES should be considered to investigate the
role of diet in metabolic syndrome.
The prevalence of metabolic syndrome increases with age, but it has different patterns
according to sex. The prevalence in women is lower but catches up to that in men after
the age of 60 years [
2
]. Because men and women have different dietary habits, work-
related activities, and socioeconomic status, it is appropriate to study the role of dietary
habits separately [
2
,
12
]. In addition, some studies have demonstrated stronger associations
between dietary carbohydrate intake and metabolic disease in women than in men [
10
,
13
].
Based on this information, we aimed to investigate the association of carbohydrate
intake with the prevalence of metabolic syndrome and its individual components in Korean
women. We also investigated whether socioeconomic status affected carbohydrate intake
and whether the association of carbohydrate intake with metabolic syndrome differed
according to participants’ CRP level and other food group intakes that were consumed as
side dishes or snacks.
2. Materials and Methods
2.1. Study Population
Korean Genomic and Epidemiology Study (KoGES) is a prospective cohort study that
investigated the environmental and genetic factors affecting prevalent chronic diseases in
the Korean population. As part of the KoGES, two community-based cohort study started
in 2001 and since then, conducted biennial follow-up studies. Subjects aged 40–60 years
were randomly selected from residents of two communities: Ansung (a rural community)
or Ansan (an urban community). A total of 10,030 participants were initially enrolled and
8840 participants (4182 men and 4658 women) completed baseline survey, anthropometric
and biochemical measurements, and genotyping. We used this baseline data obtained from
2001 to 2002. Detailed information on the study procedure is described elsewhere [
14
,
15
].
Of the 4658 women participants, participants were excluded due to an incomplete food
frequency questionnaire (FFQ) (n= 164), unreliable energy intake of
≤
500 kcal or >5000 kcal
(n= 51), and missing information regarding metabolic syndrome criteria (n= 149). Finally,
a total of 4294 women were included in this analysis.
2.2. Data Collection
Information regarding the demographic and lifestyle factors (e.g., age, sex, alcohol
consumption, tobacco smoking, marital status, education level, and physical activity level)
of the study population was obtained by trained interviewers using a questionnaire [
14
].
Subjects’ physical activity levels were assessed in metabolic equivalents (METs) computed
as the sum of METs for five levels of activity [16].
Dietary intake was assessed by using a validated semi-quantitative FFQ with 103 items,
which was developed and validated among Koreans [
17
]. Each subject provided their
average frequencies of consumption and typical portion sizes based on the year preceding
the interview. Study participants were asked to report the frequency of their consump-
tion of each food based on nine response options (never or barely, 1 time/month, 2 to
3 times/month, 1 to 2 times/week, 3 to 4 times/week, 5 to 6 times/week, 1 time/day,
2 times/day, or
≥
3 times/day) and three portion sizes (small, medium or large). To esti-
mate the intake of seasonal foods (i.e., fruits), participants were also asked to record the
period of consumption (3, 6, or 9 months or a year). Nutritional intake was estimated by
using the Food Composition Table that was developed by the Korean Health and Industry
of Development Institute (7th edition).
Anthropometry and metabolic parameters were measured by trained medical staff,
with subjects wearing light clothes. Body mass index (BMI) was calculated as weight (kg)

Nutrients 2021,13, 3098 3 of 12
divided by the square of the height (m
2
). Waist circumference was measured three times.
Blood pressure was measured at least twice under comfortable conditions. Blood samples
were obtained after participants had fasted for 8 h and were analyzed for biochemical
markers, such as high-density lipoprotein (HDL) cholesterol, triglycerides, fasting blood
glucose, and CRP.
2.3. Diagnostic Criteria for Metabolic Syndrome
This study defined metabolic syndrome using the updated National Cholesterol
Education Program Adult Treatment Panel III (NCEP-ATPIII) (NCEP-ATPIII, 2002). To
define abdominal obesity, a modified waist circumference cutoff was used [
18
]. Sub-
jects were diagnosed with metabolic syndrome if they had three or more of the follow-
ing: (a) central obesity (defined as waist circumference)
≥
85 cm in women; (b) blood
pres
sure ≥13
0/85 mmHg or use of antihypertensive medication; (c) trigly
ceride ≥15
0 mg/dL;
(d) fasting gluc
ose ≥100 m
g/dL or use of antidiabetic medication; and (e) HDL chole-
sterol < 50 mg/dL in women.
2.4. Statistical Analysis
The levels of dietary carbohydrate intake were categorized into quartiles (Qs) based
on the distribution of the metabolic syndrome-free participants (Q1: <67.0, Q2: 67.0–71.4,
Q3: 71.4–75.2, Q4:
≥
75.2% of energy) because the presence of metabolic syndrome could
change participants’ dietary habits including the amount of carbohydrate consumed.
General characteristics of the study population were examined by metabolic syndrome
status. Data are expressed as the mean
±
standard deviation for continuous variables
and frequency and percentage for categorical variables. Logistic regression analysis was
performed to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) of the risk
of metabolic syndrome (or its individual components) according to the quartiles of carbo-
hydrate intake, taking the lowest quartile group as the reference group. To select which
confounders would be controlled, a backward elimination strategy was used [
19
]. The
characteristics of the participants were categorized as follows: education level (
≤
middle
school, high school, and
≥
college), marital status (yes/no), residential location (urban
area/rural area), job (professional/office worker, service/sales, agriculture/manufacturing,
housewife/other), household income (low, medium, and high), alcohol intake (never/ever),
smoking status (never/ever), BMI (<23, 23–25,
≥
25 kg/m
2
) and physical activity (Q1: <720,
Q2: 720–1155 Q3: 1155–1965, Q4:
≥
1965 METs/d). Finally, a multivariable model was
adjusted for age, residential area, and education level. A test for trend across quartile
categories was conducted by including the median intake of each quartile as a continuous
variable in the logistic regression model.
To identify factors relating carbohydrate consumption, general characteristics and the
distributions of certain food and nutrient intakes among metabolic syndrome-free partici-
pants were examined across quartiles of carbohydrate intake. The significant differences for
several variables across levels of carbohydrate intake were examined using the generalized
linear model. Food and nutritional intake data were adjusted for total caloric intake using
Willet’s residual method and were then included in the analyses [20].
To test whether these associations were affected by individuals’ inflammatory status,
we conducted stratified analyses by CRP concentration. The subjects were categorized
into two groups (low/high, cutoff point: 1 mg/dL). Additionally, we investigated whether
intake of other food groups may affect the risk of metabolic syndrome. The subjects were
categorized into two groups (low/high) based on the median intake levels of carbohydrates
and other food groups in the metabolic syndrome-free participants, respectively. Then,
stratified analyses were carried out according to the intake of other food groups, including
fruits, kimchi (traditional fermented cabbage product), vegetables other than kimchi,
vegetables, dairy foods (milk and any foods made from milk), and meats: these food
groups were selected based on the amount of daily intake.

Nutrients 2021,13, 3098 4 of 12
All statistical analyses were performed by using SAS 9.2 software (SAS Institute Inc.,
Cary, NC, USA). A two-sided p-value of less than 0.05 was considered statistically significant.
3. Results
3.1. Association of Carbohydrate Intake with Metabolic Syndrome and Its Components
The mean age of the subjects in the current study was 52.3 years. A total of 1221 cases
of metabolic syndrome were identified, and the prevalence of metabolic syndrome was
28.4% in this population. The participants’ characteristics according to metabolic syndrome
status are presented in Supplementary Materials Table S1. Compared to those without
metabolic syndrome, those with metabolic syndrome are more likely to be older (p< 0.001),
have a higher BMI (p< 0.001), live in rural areas (p< 0.001), and be unmarried (p< 0.001).
They were more likely to have jobs that required physical labor and were less professional
(p< 0.001), had lower education levels (p< 0.001), and had lower household income
(
p< 0.001
). Those in this category were less likely to drink alcohol (
p< 0.
001) and more
likely to smoke tobacco (p= 0.018). Figure 1shows the prevalence of metabolic syndrome
and its components according to the quartiles of carbohydrate intake. Carbohydrate intake
was positively associated with the prevalence of metabolic syndrome and its components.
Low HDL cholesterol was the most common component of metabolic syndrome, followed
by high blood pressure.
Nutrients 2021, 13, x FOR PEER REVIEW 4 of 12
Then, stratified analyses were carried out according to the intake of other food groups,
including fruits, kimchi (traditional fermented cabbage product), vegetables other than
kimchi, vegetables, dairy foods (milk and any foods made from milk), and meats: these
food groups were selected based on the amount of daily intake.
All statistical analyses were performed by using SAS 9.2 software (SAS Institute Inc.,
Cary, NC, USA). A two-sided p-value of less than 0.05 was considered statistically signif-
icant.
3. Results
3.1. Association of Carbohydrate Intake with Metabolic Syndrome and Its Components
The mean age of the subjects in the current study was 52.3 years. A total of 1221 cases
of metabolic syndrome were identified, and the prevalence of metabolic syndrome was
28.4% in this population. The participants’ characteristics according to metabolic syn-
drome status are presented in Supplementary Materials Table S1. Compared to those
without metabolic syndrome, those with metabolic syndrome are more likely to be older
(p < 0.001), have a higher BMI (p < 0.001), live in rural areas (p < 0.001), and be unmarried
(p < 0.001). They were more likely to have jobs that required physical labor and were less
professional (p < 0.001), had lower education levels (p < 0.001), and had lower household
income (p < 0.001). Those in this category were less likely to drink alcohol (p < 0.001) and
more likely to smoke tobacco (p = 0.018). Figure 1 shows the prevalence of metabolic syn-
drome and its components according to the quartiles of carbohydrate intake. Carbohy-
drate intake was positively associated with the prevalence of metabolic syndrome and its
components. Low HDL cholesterol was the most common component of metabolic syn-
drome, followed by high blood pressure.
Table 1 shows the association of carbohydrate intake with the risk of metabolic syn-
drome and its components. A higher carbohydrate intake was associated with an in-
creased risk of metabolic syndrome after adjusting for covariates (OR 1.34, 95% CI 1.08–
1.66, highest vs. lowest quartile, p-trend = 0.004). Among the components of metabolic
syndrome, this association was observed for waist circumference (OR 1.32, 95% CI 1.08–
1.62, highest vs. lowest quartile, p-trend = 0.007) and HDL cholesterol (OR 1.19, 95% CI
0.99–1.43, highest vs. lowest quartile, p-trend = 0.040). There was no significant association
with other metabolic components.
Figure 1. Prevalence of metabolic syndrome and its components according to the quartiles of carbo-
hydrate intake. Q, quartile; TG, triglycerides; HDL-C, high-density lipoprotein/cholesterol; WC,
waist circumference.
Figure 1.
Prevalence of metabolic syndrome and its components according to the quartiles of
carbohydrate intake. Q, quartile; TG, triglycerides; HDL-C, high-density lipoprotein/cholesterol;
WC, waist circumference.
Table 1shows the association of carbohydrate intake with the risk of metabolic syn-
drome and its components. A higher carbohydrate intake was associated with an increased
risk of metabolic syndrome after adjusting for covariates (OR 1.34, 95% CI 1.08–1.66, highest
vs. lowest quartile, p-trend = 0.004). Among the components of metabolic syndrome, this
association was observed for waist circumference (OR 1.32, 95% CI 1.08–1.62, highest vs.
lowest quartile, p-trend = 0.007) and HDL cholesterol (OR 1.19, 95% CI 0.99–1.43, high-
est vs. lowest quartile, p-trend = 0.040). There was no significant association with other
metabolic components.

Nutrients 2021,13, 3098 5 of 12
Table 1.
Association of carbohydrate intake with the risk of metabolic syndrome and its components.
Carbohydrate Intake 1No. (%) Crude
OR (95% CI)
Adjusted
OR (95% CI) 2
Controls Cases
Metabolic syndrome
Q1 768 (25.0) 200 (16.4) 1.0 (ref) 1.0 (ref)
Q2 768 (25.0) 225 (18.4) 1.13 (0.91–1.40) 0.99 (0.79–1.25)
Q3 768 (25.0) 278 (22.8) 1.39 (1.13–1.71) 1.00 (0.80–1.25)
Q4 769 (25.0) 518 (42.4) 2.59 (2.14–3.13) 1.34 (1.08–1.66)
pfor trend <0.001 0.004
Elevated waist circumference
Q1 718 (25.9) 250 (16.4) 1.0 (ref) 1.0 (ref)
Q2 717 (25.9) 276 (18.1) 1.11 (0.91–1.35) 1.04 (0.84–1.28)
Q3 697 (25.1) 349 (22.9) 1.44 (1.19–1.74) 1.02 (0.83–1.26)
Q4 640 (23.1) 647 (42.5) 2.90 (2.42–3.48) 1.32 (1.08–1.62)
pfor trend <0.001 0.007
High blood pressure
Q1 637 (26.4) 331 (17.6) 1.0 (ref) 1.0 (ref)
Q2 618 (25.6) 375 (19.9) 2.17 (0.97–1.41) 0.99 (0.81–1.21)
Q3 594 (24.6) 452 (24.0) 1.46 (1.22–1.75) 0.98 (0.80–1.19)
Q4 564 (23.4) 723 (38.4) 2.47 (2.08–2.93) 1.07 (0.87–1.31)
pfor trend <0.001 0.623
High triglycerides
Q1 721 (23.9) 247 (19.4) 1.0 (ref) 1.0 (ref)
Q2 719 (23.8) 274 (21.5) 1.11 (0.91–1.36) 1.01 (0.82–1.23)
Q3 755 (25.0) 291 (22.9) 1.13 (0.92–1.37) 0.95 (0.78–1.17)
Q4 826 (27.3) 461 (36.2) 1.63 (1.36–1.96) 1.19 (0.97–1.47)
pfor trend <0.001 0.118
High fasting blood glucose
Q1 861 (23.0) 107 (19.4) 1.0 (ref) 1.0 (ref)
Q2 870 (23.3) 123 (22.3) 1.14 (0.86–1.50) 1.03 (0.78–1.37)
Q3 905 (24.2) 141 (25.5) 1.25 (0.96–1.64) 1.08 (0.82–1.42)
Q4 1106 (29.6) 181 (32.8) 1.32 (1.02–1.70) 1.08 (0.82–1.44)
pfor trend 0.027 0.508
Low HDL cholesterol
Q1 536 (24.9) 432 (20.2) 1.0 (ref) 1.0 (ref)
Q2 512 (23.8) 481 (22.5) 1.17 (0.98–1.39) 1.11 (0.93–1.33)
Q3 511 (23.7) 535 (25.0) 1.30 (1.09–1.55) 1.18 (0.99–1.42)
Q4 595 (27.6) 692 (32.3) 1.44 (1.22–1.71) 1.19 (0.99–1.43)
pfor trend <0.001 0.040
Abbreviations: CI, confidence interval; HDL, high-density lipoprotein; OR, odds ratio; ref, reference; Q, quartile.
1
The intake levels of carbohydrates were categorized into quartiles according to the distribution of the control
group (Q1: < 67.0, Q2: 67.0–71.4, Q3: 71.4–75.2, Q4:
≥
75.2% of energy).
2
Tests of association were from logistic
regression and were adjusted for age, residence area, and education.
3.2. Carbohydrate Intake, Socioeconomic Status, and Diet Quality
The participants’ characteristics according to the quartiles of carbohydrate intake are
presented in Table 2. Compared to those in the lowest quartile of carbohydrate intake,
those in the highest quartile were more likely to be older (p< 0.001), live in rural areas
(
p< 0.00
1), and not be married (p= 0.001). They were more likely to have jobs that required
physical labor and were less professional (p< 0.001), had lower education levels (p< 0.001),
and had lower household income (p< 0.001). Those in this category were less likely to
drink alcohol (p< 0.001). In terms of physical activity, carbohydrate intake was inversely
associated with physical activity in the first to third physical activity groups. However,
those in the highest quartiles of physical activity consumed more carbohydrates (p< 0.001).
No differences were observed in BMI and smoking status.

Nutrients 2021,13, 3098 6 of 12
Table 2.
General characteristics of the study participants according to the quartiles of carbohydrate
intake in controls 1.
Quartiles of Carbohydrate Intake 3
Q1 (n= 768) Q2 (n= 768) Q3 (n= 768) Q4 (n= 769) p-Value 4
Age (years)
<50 546 (71.1) 454 (59.1) 414 (53.9) 263 (34.2) <0.001
50–60 147 (19.1) 186 (24.2) 186 (24.2) 235 (30.6)
≥60 75 (9.8) 128 (16.7) 168 (21.9) 271 (35.2)
BMI (kg/m2)
<23 300 (39.1) 229 (33.7) 262 (34.1) 301 (39.1) 0.063
23–25 227 (29.6) 238 (31.0) 221 (28.8) 206 (26.8)
≥25 241 (31.4) 271 (35.3) 285 (37.1) 262 (34.1)
Residential location
Rural area 230 (30.0) 201 (26.2) 318 (41.4) 558 (72.6) <0.001
Urban area 538 (70.1) 567 (73.8) 450 (58.6) 211 (27.4)
Marital status
No 69 (9.0) 102 (13.1) 100 (13.0) 120 (15.6) 0.001
Yes 699 (91.0) 664 (86.7) 664 (86.9) 644 (84.3)
Unknown 0 (0) 2 (0.3) 4 (0.5) 5 (0.7)
Occupation
Professional, Office 29 (3.8) 26 (3.4) 12 (1.7) 13 (1.7) <0.001
Service, Sales 114 (14.8) 81 (10.5) 78 (10.2) 56 (7.3)
Agriculture, manufacturing 107 (14.0) 109 (14.2) 175 (22.8) 309 (40.2)
Housewife, others 517 (67.3) 552 (71.9) 501 (65.2) 385 (50.1)
Unknown 1 (0.1) 0 (0) 2 (0.3) 3 (0.8)
Education level
Middle school or less 343 (44.7) 406 (52.9) 501 (65.2) 629 (81.8) <0.001
High school 345 (44.9) 289 (37.6) 212 (27.6) 110 (14.3)
College or more 78 (10.2) 68 (8.9) 49 (6.4) 24 (3.1)
Unknown 0 (0.3) 5 (0.7) 6 (0.8) 6 (0.8)
Household income 2
Low (<200 won) 385 (50.1) 429 (55.9) 512 (66.7) 625 (81.4) <0.001
Medium (200–400 won) 296 (38.5) 260 (33.9) 196 (25.5) 104 (13.5)
High (≥400 won) 77 (10.0) 64 (8.3) 46 (6.0) 15 (2.0)
Unknown 10 (1.3) 15 (2.0) 14 (1.8) 25 (3.3)
Alcohol consumption
Never 458 (59.6) 541 (70.4) 521 (67.8) 574 (74.6) <0.001
Ever 309 (40.3) 223 (29.2) 245 (32.0) 189 (24.8)
Unknown 1 (0.1) 4 (0.5) 2 (0.3) 6 (0.8)
Smoking status
Never 712 (92.7) 731 (95.2) 731 (95.2) 722 (93.9) 0.335
Ever 41 (5.4) 29 (3.8) 33 (4.3) 28 (3.7)
Unknown 15 (2.0) 8 (1.0) 4 (0.5) 19 (2.5)
Physical activity (METs/day)
Q1 (<720) 191 (24.9) 193 (25.1) 169 (22.0) 204 (26.5) <0.001
Q2 (720–1155) 223 (29.0) 243 (31.6) 206 (26.8) 134 (17.4)
Q3 (1155–1965) 250 (32.6) 214 (27.9) 187 (24.3) 133 (17.3)
Q4 (≥1965) 87 (11.3) 101 (13.2) 189 (24.6) 262 (34.1)
Unknown 17 (2.2) 17 (2.2) 17 (2.2) 36 (4.7)
Abbreviations: BMI, body mass index; METs, metabolic equivalents.
1
Data are presented as n(%).
2
Unit is
10,000 won in Korean currency ($1 = 1103.50 Korean won as of 24 December 2020).
3
Subjects were divided into
four groups based on carbohydrate intake among those having no metabolic syndrome (Q1: < 67.0, Q2: 67.0–71.4,
Q3: 71.4–75.2, Q4: ≥75.2% of energy). 4Tests of association by chi-square test (categorical variables).
The daily intakes of all macronutrients, except energy and vitamin C, were signifi-
cantly lower in the highest carbohydrate group than in the lowest carbohydrate group
(Supplementary Materials Table S2). The percentage of energy intake obtained from fat
and protein was significantly lower in the group with higher carbohydrate intake. In
addition, we examined the distribution of the intakes of other foods across the quartiles
of carbohydrate intake (Supplementary Materials Table S3). Most of the intakes of the
examined food groups decreased across the lowest to highest quartiles of carbohydrate

Nutrients 2021,13, 3098 7 of 12
intake. The intake of grains, kimchi, and fruit/fruit juice tended to be higher in the highest
quartiles than in the lowest quartiles (pfor all < 0.001).
3.3. Association between Carbohydrate Intake and Metabolic Syndrome According to CRP Level
and Other Food Intake
When the data were stratified by individuals’ CRP level, the association between carbo-
hydrate intake and metabolic syndrome was significant only among those with a low level
of CRP (OR 1.84, 95% CI 1.21–2.80, highest vs. lowest quartile,
p-trend = 0.003
) (
Table 3
).
Among metabolic components, this association was observed only for elevated waist cir-
cumference (OR 1.76, 95% CI 1.24–2.51, highest vs. lowest quartile,
p-trend = 0.0
02), high
triglycerides (OR 1.53, 95% CI 1.03–2.26, highest vs. lowest quartile,
p-trend = 0.026
), and
low HDL cholesterol (OR 1.57, 95% CI 1.15–2.14, highest vs. lowest quartile,
p-trend = 0.001
).
However, none of the associations was significant among those in the high-CRP group.
Table 3.
Association of carbohydrate intake with the risk of metabolic syndrome and its components, stratified by CRP level.
Carbohydrate Intake 1Low CRP 2High CRP
No. Controls/Cases OR (95% CI) 3No. Controls/Cases OR (95% CI) 3
Metabolic syndrome
Q1 341/43 1.0 (ref) 427/157 1.0 (ref)
Q2 298/55 1.40 (0.90–2.20) 470/170 0.86 (0.66–1.12)
Q3 295/78 1.65 (1.08–2.52) 473/200 0.82 (0.63–1.07)
Q4 290/132 1.84 (1.21–2.80) 479/386 1.17 (0.91–1.51)
pfor trend 0.003 0.147
Elevated waist circumference
Q1 308/76 1.0 (ref) 410/174 1.0 (ref)
Q2 276/77 1.09 (0.76–1.58) 441/199 0.99 (0.76–1.28)
Q3 278/95 1.12 (0.78–1.61) 419/254 0.95 (0.73–1.23)
Q4 234/188 1.76 (1.24–2.51) 406/459 1.13 (0.87–1.45)
pfor trend 0.002 0.344
High blood pressure
Q1 284/100 1.0 (ref) 353/231 1.0 (ref)
Q2 249/104 1.06 (0.74–1.50) 369/271 0.95 (0.74–1.21)
Q3 242/131 1.06 (0.75–1.49) 352/321 0.94 (0.74–1.21)
Q4 208/214 1.16 (0.82–1.64) 356/509 1.01 (0.79–1.30)
pfor trend 0.464 0.972
High triglyceride
Q1 325/59 1.0 (ref) 396/188 1.0 (ref)
Q2 282/71 1.29 (0.87–1.90) 437/203 0.89 (0.70–1.14)
Q3 286/87 1.49 (1.02–2.18) 469/204 0.78 (0.61–1.00)
Q4 311/111 1.53 (1.03–2.26) 515/350 1.06 (0.83–1.35)
pfor trend 0.026 0.729
High fasting blood glucose
Q1 359/25 1.0 (ref) 502/82 1.0 (ref)
Q2 324/29 1.15 (0.65–2.02) 546/94 0.98 (0.71–1.36)
Q3 334/39 1.37 (0.80–2.35) 571/102 0.98 (0.71–1.35)
Q4 371/51 1.45 (0.84–2.51) 735/130 0.97 (0.70–1.34)
pfor trend 0.165 0.892
Low HDL cholesterol
Q1 248/136 1.0 (ref) 288/296 1.0 (ref)
Q2 190/163 1.51 (1.12–2.03) 322/318 0.91 (0.73–1.14)
Q3 183/190 1.80 (1.34–2.42) 328/345 0.92 (0.73–1.15)
Q4 213/209 1.57 (1.15–2.14) 382/483 1.00 (0.79–1.26)
pfor trend 0.001 0.976
Abbreviations: CI, confidence interval; CRP, c-reactive protein; HDL, low-density lipoprotein; OR, odds ratio; ref, reference; Q, quartile.
1
The intake levels of carbohydrates were categorized into quartiles according to the distribution of the control group (Q1: < 67.0, Q2:
67.0–71.4, Q3: 71.4–75.2, Q4:
≥
75.2% of energy).
2
The level of CRP was categorized into two groups (cutoff: 1 mg/dL).
3
Tests of association
were from logistic regression and were adjusted for age, residence area, and education.
Furthermore, we investigated whether the intake of other food groups (kimchi, fruits,
vegetables without kimchi, vegetables, dairy foods, and meat) may affect the prevalence of
metabolic syndrome. When the data were stratified by the amount of intake of the other
food groups, those with high dairy intake showed a reduced risk of metabolic syndrome

Nutrients 2021,13, 3098 8 of 12
in the group of participants with high carbohydrate intake (OR 0.78, 95% CI 0.64–0.95,
p= 0.0
14, high vs. low dairy intake) (Supplementary Materials Table S4). Among the
metabolic components, this association was observed only for elevated waist circumference
(OR 0.72, 95% CI 0.59–0.87, p< 0.001), high triglycerides (OR 0.72, 95% CI 0.59–0.87,
p= 0.001), and low HDL cholesterol (OR 0.80, 95% CI 0.67–0.96, p= 0.015) (Table 4).
Table 4.
Association between dairy food intake and the risk of metabolic syndrome and its components, stratified by the
level of carbohydrate intake 1.
Low Carbohydrate Intake High Carbohydrate Intake
Dairy Food Intake Low High p-Value Low High p-Value
Metabolic syndrome
No. controls/cases 563/168 973/257 0.333 973/571 564/225 0.014
OR (95% CI)21.0 (ref) 0.89 (0.71–1.13) 1.0 (ref) 0.78 (0.64–0.95)
Elevated waist circumference
No. controls/cases 510/221 925/305 0.062 818/726 519/270 <0.001
OR (95% CI) 21.0 (ref) 0.81 (0.65–1.01) 1.0 (ref) 0.72 (0.59–0.87)
High blood pressure
No. controls/cases 461/270 794/436 0.444 732/812 426/363 0.264
OR (95% CI) 21.0 (ref) 0.92 (0.75–1.14) 1.0 (ref) 0.90 (0.74–1.09)
High triglycerides
No. controls/cases 531/200 909/321 0.431 1006/538 575/214 0.001
OR (95% CI) 21.0 (ref) 0.92 (0.74–1.14) 1.0 (ref) 0.72 (0.59–0.87)
High fasting glucose
No. controls/cases 634/97 1097/133 0.051 1335/209 676/113 0.601
OR (95% CI) 21.0 (ref) 0.75 (0.57–1.00) 1.0 (ref) 1.07 (0.83–1.38)
Low HDL cholesterol
No. controls/cases 384/347 664/566 0.624 701/843 405/384 0.015
OR (95% CI) 21.0 (ref) 0.95 (0.79–1.15) 1.0 (ref) 0.80 (0.67–0.96)
Abbreviations: CI, confidence interval; HDL, high-density lipoprotein; OR, odds ratio.
1
The intake levels of carbohydrate and dairy food
were categorized into two groups according to the distribution in the control group.
2
Tests of association were from logistic regression and
were adjusted for age, residence area, and education.
4. Discussion
In the present study, carbohydrate intake was associated with metabolic syndrome,
and this association may differ according to individuals’ CRP levels and dairy food intake.
Those with high carbohydrate level were more likely to be in low socioeconomic status,
and their diet lacked variety and balance.
4.1. Carbohydrate Intake and Metabolic Syndrome
Previous studies have reported mixed findings regarding the association between
carbohydrate intake and the risk of metabolic syndrome [
7
–
10
]. The results from different
countries are hard to compare because of differences in the criteria used to define metabolic
syndrome, study design, and many other factors. The present study showed that high
carbohydrate intake was associated with an increased risk of metabolic syndrome. Among
the five components of metabolic syndrome, abdominal obesity showed the strongest
association with carbohydrate intake. Abdominal obesity indicates the existence of adipose
tissue dysfunction and is independently associated with hyperlipidemia (low HDL choles-
terol and high triglycerides), increased insulin resistance, elevated risk of diabetes, and
subclinical atherosclerosis. Therefore, identifying the determinants of abdominal obesity
may reveal strategies to prevent these metabolic abnormalities [21].
Asians are quite prone to visceral fat accumulation, which may explain their greater
tendency to develop metabolic complications of obesity at relatively low BMI
values [22,23]
.
Some studies have reported the association of carbohydrate intake with abdominal obe-
sity [
24
]. In a randomized controlled study of obese subjects with type 2 diabetes mellitus,

Nutrients 2021,13, 3098 9 of 12
a greater decrease in visceral fat was observed in the low-carbohydrate group than in the
high-carbohydrate group. These results imply that, among isocaloric diets, a low carbohy-
drate diet might be more effective in reducing visceral fat, improving insulin sensitivity,
and increasing HDL cholesterol levels than a high carbohydrate diet in obese subjects
with type 2 diabetes mellitus [
24
]. Based on this evidence, higher carbohydrate intake and
a higher tendency of accumulating visceral fat among Asians may increase their risk of
metabolic syndrome.
4.2. Carbohydrate Intake and Socioeconomic Status
The proportion of carbohydrates consumed seems to be associated with socioeconomic
characteristics [
11
]. In the current study, those with a higher carbohydrate intake were
more likely to be older, have a lower level of education and household income, and live
in rural areas than those with a lower carbohydrate intake. Accordingly, the association
of carbohydrate intake with metabolic syndrome prevalence was attenuated after mul-
tivariable adjustment of these factors [
21
]. Low socioeconomic status is possibly linked
to a high carbohydrate intake and a higher prevalence of metabolic syndrome [
25
]. It
has been suggested that socioeconomic status may affect the health status of individuals
due to its effect on macronutrient balance [
11
]. Analyses that have considered education,
occupation, income, and employment status have shown that education is usually the
strongest determinant of socioeconomic differences [
26
]. Higher levels of education may
increase the ability of individuals to understand health-related information and develop
health-promoting behaviors. Additionally, older age was strongly associated with high
carbohydrate intake, possibly because food preferences and socioeconomic status (e.g.,
household income and occupation type) change with age. Finally, rural residents had
a higher level of carbohydrate intake than urban residents. Generally, rural residents
are older and have low socioeconomic status [
27
]. Based on these findings, strategies to
improve diet quality need to consider socioeconomic status.
4.3. Carbohydrate Intake and Diet Quality
Carbohydrate intake may play an important role in diet quality in this population
because the proportion of carbohydrates was more than 70% of the total calories. Therefore,
we examined differences in food and nutrient intake across quartiles of carbohydrate intake.
Individuals with a high intake of carbohydrates had a higher intake of grains, but the
intakes of other foods, even carbohydrate-rich foods, except for kimchi and fruit/fruit
juice, were lower than those in the lower-carbohydrate groups. Lower intake of side dishes
and higher intake of rice may lead to an imbalanced intake of macronutrients and other
nutrients [
11
]. As a result, their diets lack protein, fat, and other essential minerals and
vitamins. The diet of those with higher carbohydrate levels lacked variety and balance
and was thus poor in regard to diet quality. Even though both the quality and quantity of
carbohydrates are important, reducing the amount of carbohydrates (especially refined
carbohydrates) is more effective in lowering glycemic load than reducing the overall dietary
glycemic index alone [
6
]. High carbohydrate intake is a major characteristic of the Korean
diet because rice is a staple food among Koreans. Therefore, exploring foods that are
optimal alternatives for white rice to reduce metabolic syndrome incidence could be an
important focus.
Therefore, other food choices, such as side dishes and snacks, may help to reduce the
risk of metabolic syndrome. To examine the differences in metabolic syndrome prevalence
by the amount of consumption of other food groups, we conducted stratified analyses.
We found that those with higher dairy intake had a reduced risk of metabolic syndrome
in the high-carbohydrate group. Because dairy foods (e.g., milk, yogurt, and cheese)
are usually not included in the traditional Korean diet, their intake is lower than that in
Western countries [
28
]. Many studies have also reported an inverse association between
dairy intake and the risk of metabolic syndrome [
29
]. Various nutrients in milk (e.g.,
calcium and dairy protein) may synergistically protect against metabolic syndrome and

Nutrients 2021,13, 3098 10 of 12
its individual components. Calcium may increase the binding of fatty acids and bile acids
in the intestine, thus increasing fecal fat excretion and inhibiting fat reabsorption [
30
].
Additionally, calcium may regulate body fat deposition by affecting adipocyte intracellular
calcium concentrations and reducing fatty acid synthesis while increasing lipolysis and
thus utilizing triglyceride stores [
31
]. Dairy protein may also improve metabolic health
by promoting changes in body composition in favor of increased lean body mass and
decreased fat mass [
32
,
33
]. To improve the health status of those with higher carbohydrate
intake, substituting carbohydrate-rich foods with proteins and fats from healthy sources
may make it easier to dramatically change their diet [21,34].
4.4. Metabolic Syndrome and Inflammation
Metabolic syndrome is a proinflammatory state characterized by increased CRP lev-
els [
4
]. In the present study, we observed a stronger association between carbohydrate
intake and metabolic syndrome among those with low levels of CRP. This association
was also observed for central obesity and dyslipidemia. Carbohydrates and other dietary
components in the high-carbohydrate group may affect proinflammatory status and the
level of CRP [
35
,
36
]. A higher level of CRP is related to an increased risk of metabolic
syndrome and its components, which are associated with underlying inflammatory pro-
cesses [
3
,
4
]. CRP is produced in the liver, primarily in response to interleukin-6, tumor
necrosis factor-
α
and interleukin-1, each of which has been implicated in insulin-resistance
pathways. A higher carbohydrate diet may induce higher rates of interleukin-6 secretion
from adipocytes and thus increase the level of CRP [
36
]. Our findings show that the role of
diet could be more significant among those with a low inflammatory status, implying the
importance of prevention in managing metabolic syndrome.
4.5. Study Limitation
This study had several limitations that should be considered. First, our data are based
on a cross-sectional study; thus, it is difficult to explain the causal relationship between
carbohydrate intake and metabolic syndrome. Second, dietary information was obtained
from a self-reported FFQ; thus, measurement errors in dietary assessment are inevitable.
Because underreporting of energy intake is a major source of bias in dietary surveys, we
conducted energy adjustment to mitigate the effects of measurement error in data collected
via self-reported dietary assessment instruments [
20
]. Third, carbohydrate intake is very
high in this population. Therefore, variation in the intake of carbohydrates (typically
providing 60% to 70% of daily energy) is smaller than that for other nutrients. Fourth,
our data did not provide the information of menopausal status. However, our analyses
were adjusted by age, so it could mitigate the lack of this information. Finally, this study
included only Korean women. Thus, further studies are required to examine the association
between carbohydrate intake and metabolic syndrome according to race and gender, and
the underlying mechanism.
5. Conclusions
The present study found that high carbohydrate intake was associated with an in-
creased prevalence of metabolic syndrome and abdominal obesity. This association could
differ according to individuals’ inflammatory status and other food choices. Socioeconomic
factors affect carbohydrate intake, and a higher proportion of carbohydrate intake results
in an imbalanced diet with poor diet quality. These findings may help to develop more
specialized strategies to prevent metabolic syndrome in vulnerable social groups. Further
investigations are needed to determine an appropriate carbohydrate intake level and diet
composition. The implementation of dietary interventions, such as providing food vouch-
ers, could be one of the effective strategies to allow those with low socioeconomic status to
achieve a balanced diet with variety.

Nutrients 2021,13, 3098 11 of 12
Supplementary Materials:
The following are available online at https://www.mdpi.com/article/
10.3390/nu13093098/s1. Table S1: General characteristics of the study participants according to
metabolic syndrome status. Table S2: Daily intake of energy and nutrients in the subjects according to
carbohydrate intake in controls. Table S3: Food-group intake according to the level of carbohydrate
intake in controls. Table S4: Association between carbohydrate intake and the risk of metabolic
syndrome, stratified by the intakes of different food groups.
Author Contributions:
Conceptualization, Y.-A.C. and J.-H.C.; methodology, Y.-A.C. and J.-H.C.;
software, Y.-A.C.; validation, Y.-A.C. and J.-H.C.; formal analysis, Y.-A.C.; investigation, Y.-A.C.
and J.-H.C.; resources, J.-H.C.; data curation, Y.-A.C.; writing—original draft preparation, Y.-A.C.;
writing—review and editing, Y.-A.C. and J.-H.C.; visualization, Y.-A.C.; supervision, J.-H.C.; project
administration, J.-H.C.; funding acquisition, J.-H.C. All authors have read and agreed to the published
version of the manuscript.
Funding:
This research was funded by the National Research Foundation of Korea (NRF), grant
number NRF-2018R1A1A1A05019155 and NRF-2021R1A2C1008635.
Institutional Review Board Statement:
The KoGES was conducted according to the guidelines of
the Declaration of Helsinki and approved by the Institutional Review Board of Korea Centers for
Disease Control and Prevention. This study was approved by the Institutional Review Board of
University (40525-201802-HR-121-07).
Informed Consent Statement:
This study analyzed the data of KoGES. Informed consent was
obtained from all subjects involved in the study.
Acknowledgments:
This work was conducted with bioresources from the National Biobank of Korea,
the Korea Disease Control and Prevention Agency, Republic of Korea (KBN-2018-018).
Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design
of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or
in the decision to publish the results.
References
1.
Liu, J.; Grundy, S.M.; Wang, W.; Smith, S.C.; Vega, G.L.; Wu, Z.; Zeng, Z.; Wang, W.; Zhao, D. Ethnic-specific criteria for the
metabolic syndrome: Evidence from China. Diabetes Care 2006,29, 1414–1416. [CrossRef] [PubMed]
2.
Cornier, M.A.; Dabelea, D.; Hernandez, T.L.; Lindstrom, R.C.; Steig, A.J.; Stob, N.R.; Van Pelt, R.E.; Wang, H.; Eckel, R.H. The
metabolic syndrome. Endocr. Rev. 2008,29, 777–822. [CrossRef]
3.
Ridker, P.M.; Bruing, J.E.; Cook, N.R.; Rifai, N. C-reactive protein, the metabolic syndrome, and risk of incident cardiovascular
event. Circulation 2003,107, 391–397. [CrossRef]
4.
Devaraj, S.; Singh, U.; Jialal, I. Human C-reactive protein and the metabolic syndrome. Curr. Opin. Lipidol.
2009
,20, 182–189.
[CrossRef]
5.
Pitsavos, C.; Panagiotakos, D.; Weinem, M.; Stefanadis, C. Diet, exercise and the metabolic syndrome. Rev. Diabet. Stud.
2006
,3,
118–126. [CrossRef]
6.
Schulze, M.B.; Hu, F.B. Dietary approaches to prevent the metabolic syndrome: Quality versus quantity of carbohydrates. Diabetes
Care 2004,27, 613–614. [CrossRef] [PubMed]
7.
Ha, K.; Kim, K.; Chun, O.K.; Joung, H.; Song, Y. Differential association of dietary carbohydrate intake with metabolic syndrome
in the US and Korean adults: Data from the 2007–2012 NHANES and KNHANES. Eur. J. Clin. Nutr.
2018
,72, 848–860. [CrossRef]
[PubMed]
8.
Liu, Y.; Wu, Q.; Xia, Y.; Zhang, J.; Jiang, Y.; Chang, Q.; Zhao, Y. Carbohydrate intake and risk of metabolic syndrome: A
dose-response meta-analysis of observational studies. Nutr. Metab. Cardiovasc. Dis. 2019,29, 1288–1298. [CrossRef]
9.
McKeown, N.M.; Meigs, J.B.; Liu, S.; Saltzman, E.; Wilson, P.W.F.; Jacques, P.F. Carbohydrate nutrition, insulin resistance, and the
prevalence of the metabolic syndrome in the Framingham Offspring Cohort. Diabetes Care 2004,27, 538–546. [CrossRef]
10.
Song, S.; Lee, J.E.; Song, W.O.; Paik, H.; Song, Y. Carbohydrate intake and refined-grain consumption are associated with
metabolic syndrome in the Korean adult population. J. Acad. Nutr. Diet. 2014,114, 54–62. [CrossRef] [PubMed]
11.
Sakurai, M.; Nakagawa, H.; Kadota, A.; Yoshita, K.; Nakamura, Y.; Okuda, N.; Nishi, N.; Miyamoto, Y.; Arima, H.; Ohkubo, T.;
et al. Macronutrient intake and socioeconomic status: NIPPON DATA2010. J. Epidemiol.
2018
,28 (Suppl. 3), S17–S22. [CrossRef]
12.
Irala-Estevez, J.D.; Groth, M.; Johansson, L.; Oltersdorf, U.; Prattala, R.; Martinez-Gonzalez, M.A. A systematic review of
socio-economic differences in food habits in Europe: Consumption of fruit and vegetables. Eur. J. Clin. Nutr.
2000
,54, 706–714.
[CrossRef] [PubMed]

Nutrients 2021,13, 3098 12 of 12
13.
Park, S.H.; Lee, K.S.; Park, H.Y. Dietary carbohydrate intake is associated with cardiovascular disease risk in Korean: Analysis
of the third Korean National Health and Nutrition Examination Survey (KNHANES III). Int. J. Cardiol.
2010
,139, 234–240.
[CrossRef] [PubMed]
14.
Kim, Y.; Han, B.G.; KoGES Group. Cohort Profile: The Korean Genome and Epidemiology Study (KoGES) Consortium. Int. J.
Epidemiol. 2017,46, 1350. [CrossRef]
15.
Cho, Y.S.; Go, M.J.; Kim, Y.J.; Heo, J.Y.; Oh, J.H.; Ban, H.; Yoon, D.; Lee, M.H.; Kim, D.; Park, M.; et al. A large-scale genome-wide
association study of Asian populations uncovers genetic factors influencing eight quantitative traits. Nat. Genet.
2009
,41, 527–534.
[CrossRef]
16.
Ainsworth, B.E.; Haskell, W.L.; Whitt, M.C.; Irwin, L.; Swartz, A.M.; Strath, S.J.; O’BRIEN, W.L.; Bassett, D.R.; Schmitz, K.H.;
Emplaincourt, R.O.; et al. Compendium of physical activities: An update of activity codes and MET intensities. Med. Sci. Sports
Exerc. 2000,32 (Suppl. 9), S498–S504. [CrossRef]
17.
Ahn, Y.; Kwon, E.; Shim, J.E.; Park, M.K.; Joo, Y.; Kimm, K.; Park, C.; Kim, D.H. Validation and reproducibility of food frequency
questionnaire for Korean genome epidemiologic study. Eur. J. Clin. Nutr. 2007,61, 1435–1441. [CrossRef] [PubMed]
18.
Lee, S.; Park, H.S.; Kim, S.M.; Kwon, H.S.; Kim, D.Y.; Kim, D.J.; Cho, G.J.; Han, J.H.; Kim, S.R.; Park, C.Y.; et al. Cut-off points of
waist circumference for defining abdominal obesity in the Korean population. Korean J. Obes. 2006,15, 1–9.
19.
Bursac, Z.; Gauss, C.H.; Williams, D.K.; Hosmer, D.W. Purposeful selection of variables in logistic regression. Source Code Biol.
Med. 2008,3, 17. [CrossRef]
20.
Willet, W.C.; Howe, G.R.; Kushi, L.H. Adjustment for total energy intake in epidemiologic studies. Am. J. Clin. Nutr.
1997
,65
(Suppl. 4), 1220S–1228S. [CrossRef]
21.
Merchant, A.T.; Anand, S.S.; Vuksan, V.; Jacobs, R.; Davis, B.; Teo, K.; Yusuf, S. Protein intake is inversely associated with
abdominal obesity in a multi-ethnic population. J. Nutr. 2005,135, 1196–1201. [PubMed]
22.
Kadowaki, T.; Sekikawa, A.; Murata, K.; Maegawa, H.; Takamiya, T.; Okamura, T.; EI-Saed, A.; Miyamatsu, N.; Edmundowicz,
D.; Kita, Y.; et al. Japanese men have larger areas of visceral adipose tissue than Caucasian men in the same levels of waist
circumference in a population-based study. Int. J. Obes. (Lond.) 2006,30, 1163–1165. [CrossRef] [PubMed]
23.
Tan, C.E.; Ma, S.; Wai, D.; Chew, S.; Tai, E. Can we apply the National Cholesterol Education Program Adult Treatment Panel
definition of the metabolic syndrome to Asians? Diabetes Care 2004,27, 1182–1186. [CrossRef]
24.
Miyashita, Y.; Koide, N.; Ohtsuka, M.; Ozaki, H.; Itoh, Y.; Oyama, T.; Uetake, T.; Ariga, K.; Shirai, K. Beneficial effect of low
carbohydrate in low calorie diets on visceral fat reduction in type 2 diabetic patients with obesity. Diabetes Res. Clin. Pract.
2004
,
65, 235–241. [CrossRef]
25.
Wamala, S.P.; Lynch, J.; Horsten, M.; Mittleman, M.A.; Schenck-Gustafsson, K.; Orth-Gomer, K. Education and the metabolic
syndrome in women. Diabetes Care 1999,22, 1999–2003. [CrossRef]
26.
Roos, E.; Prattala, R.; Lahelma, E.; Kleemola, P.; Pietinen, P. Modern and healthy? Socioeconomic differences in the quality of diet.
Eur. J. Clin. Nutr. 1996,50, 753–760.
27.
Lee, S.; Shin, Y.; Kim, Y. Risk of metabolic syndrome among middle-aged Koreans from rural and urban areas. Nutrients
2018
,10,
859. [CrossRef] [PubMed]
28.
Shin, S.; Lee, H.W.; Kim, C.E.; Kim, J.; Lee, J.; Kang, D. Association between milk consumption and metabolic syndrome among
Korean adults: Results from the Health Examinees Study. Nutrients 2017,9, 1102. [CrossRef] [PubMed]
29.
Chen, G.C.; Szeto, I.M.Y.; Chen, L.H.; Han, S.; Li, Y.; Hekezen, R.; Qin, L. Dairy products consumption and metabolic syndrome
in adults: Systematic review and metabolic analysis of observational studies. Sci. Rep. 2015,5, 14606. [CrossRef]
30.
Boon, N.; Hul, G.B.J.; Stegen, J.H.C.H.; Sluijsmans, W.E.M.; Valle, C.; Langin, D.; Viguerie, N.; Saris, W.H.M. An intervention
study of the effects of calcium intake on faecal fat excretion, energy metabolism and adipose tissue mRNA expression of
lipid-metabolism related proteins. Int. J. Obes. 2007,31, 1704–1712. [CrossRef]
31.
Zhang, F.; Ye, J.; Zhu, X.; Wang, L.; Gao, P.; Shu, G.; Jiang, Q.; Wang, S. Anti-obesity effects of dietary calcium: The evidence and
possible mechanisms. Int. J. Mol. Sci. 2019,20, 307.
32.
McGregor, R.A.; Poppitt, S.D. Milk protein for improved metabolic health: A review of evidence. Nutr. Metab.
2013
,10, 46.
[CrossRef] [PubMed]
33.
So, E.; Choi, S.K.; Joung, H. Impact of dietary protein intake and obesity on lean mass in middle-aged individuals after a 12-year
follow-up: The Korean Genome and Epidemiology Study (KoGES). Br. J. Nutr. 2019,122, 322–330. [CrossRef]
34. Van Dam, R.M.; Seidel, J.C. Carbohydrate intake and obesity. Eur. J. Clin. Nutr. 2017,61 (Suppl. 1), S75–S99. [CrossRef]
35.
Shivappa, N.; Steck, S.E.; Hurley, T.G.; Hussey, J.R.; Hebert, J.R. Designing and developing a literature-derived, population-based
dietary inflammatory index. Public Health Nutr. 2014,17, 1689–1696. [CrossRef] [PubMed]
36.
Huffmann, K.M.; Orenduff, M.C.; Samsa, G.P.; Houmard, J.A.; Kraus, W.E.; Bales, C.W. Dietary carbohydrate intake and high
sensitivity C reactive protein in at-risk women and men. Am. Heart J. 2008,154, 962–968. [CrossRef]