The effect of dietary glycemic index and glycemic load on
inflammatory biomarkers: a systematic review and meta-analysis of
randomized clinical trials
Alireza Milajerdi,1 Parvane Saneei,4 Bagher Larijani,2 and Ahmad Esmaillzadeh1,3,5
1 Department of Community Nutrition, School of Nutritional Sciences and Dietetics; 2 Endocrinology and Metabolism Research Center, Endocrinology and
Metabolism Clinical Sciences Institute; 3 Obesity and Eating Habits Research Center, Endocrinology and Metabolism Molecular–Cellular Sciences Institute,
Tehran University of Medical Sciences, Tehran, Iran; 4 Students’ Research Committee; and 5 Food Security Research Center, Department of Community Nu-
trition, School of Nutrition and Food Science, Isfahan University of Medical Sciences, Isfahan, Iran
ABSTRACT
Background: To our knowledge, there is no study available that sum-
marizes earlier findings on the effect of dietary glycemic index (GI)
and glycemic load (GL) on inflammatory biomarkers.
Objective: This systematic review and meta-analysis was conducted
to systematically review the available clinical trials that examined the
effects of low-GI (LGI) and low-GL (LGL) diets on several inflam-
matory biomarkers in adults.
Design: We searched for relevant articles published up to June
2017 through PubMed, Medline, SCOPUS, EMBASE, and Google
Scholar with the use of relevant keywords. Clinical trials that exam-
ined the effect of dietary GI and GL on inflammation in adults were
included.
Results: Overall, 28 randomized controlled trials (RCTs) includ-
ing 2961 participants (59% women, 41% men) were included in
this meta-analysis. By combining findings from 14 studies on high-
sensitivity C-reactive protein (hs-CRP) concentrations, we found
no significant effect of LGI or LGL diets on serum hs-CRP con-
centrations compared with the control diet [weighted mean dif-
ference (WMD) for dietary GI: −0.05 mg/L (95% CI: −0.21,
0.10 mg/L); and WMD for dietary GL: 0.08 mg/L (95% CI: –0.26,
0.42 mg/L), respectively]. After combining effect sizes from 5 stud-
ies, we did not find significant changes in serum tumor necrosis factor
α (TNF-α) concentrations comparing control diets with LGI (WMD:
–0.18 mg/L; 95% CI: –0.43, 0.06 mg/L) or LGL (WMD: –0.20 mg/L;
95% CI: –0.33, 0.07 mg/L) diets. Significant changes were also not
seen in leptin and interleukin 6 (IL-6) concentrations after the con-
sumption of LGI or LGL diets.
Conclusions: We did not find any significant effect of dietary GI
or GL on serum concentrations of inflammatory cytokines, including
hs-CRP, leptin, IL-6, and TNF-α in adults. Additional RCTs—in par-
ticular, feeding trials—are required to shed light on this issue. Am
J Clin Nutr 2018;107:593–606.
Keywords: diet, glycemic index, glycemic load, inflammation,
meta-analysis
Am J Clin Nutr 2018;107:593–606. Printed in USA. © 2018 American Society for Nutrition. All rights reserved.
Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636
by University of Durham user
on 09 April 2018
INTRODUCTION
Inflammation, characterized by increased concentrations of cy-
tokines, plays a central role in innate immunity (1). As an essen-
tial part of host defense, low-grade inflammation has been asso-
ciated with type 2 diabetes, cardiovascular disease (CVD), and
other metabolic disorders (2).
Different lifestyle-related factors, including diet, are associ-
ated with low-grade inflammation (3). The association between
type and amount of dietary carbohydrate and low-grade inflam-
mation has been shown (4). Dietary glycemic index (GI) is de-
fined as the potential of carbohydrate-containing foods to in-
crease postprandial blood glucose. Adherence to low-GI (LGI)
diets was associated with a decreased risk of mortality, CVD, di-
abetes, and cancer (5). Earlier studies also reported the associ-
ation between dietary GI and inflammation (6). In addition, di-
etary glycemic load (GL) as a measure of carbohydrate quality
and quantity was also associated with the risk of several chronic
diseases and inflammation (7, 8). Some investigators found an in-
verse association between dietary GI and GL and serum concen-
trations of inflammatory biomarkers (9), whereas others found no
significant associations (10). Qi and Hu (11) found a direct asso-
ciation between dietary GL and systemic inflammation in patients
with diabetes. However, findings on the association of dietary
GL and inflammation are conflicting (12). In a meta-analysis,
a significant reduction in serum high-sensitivity C-reactive
AE was supported by the Iran National Science Foundation. No external
funding was received for this study.
Address correspondence to AE (e-mail: a-esmaillzadeh@tums.ac.ir).
Abbreviations used: CVD, cardiovascular disease; GI, glycemic index; GL,
glycemic load; HGI, high glycemic index; HGL, high glycemic load; hs-CRP,
high-sensitivity C-reactive protein; LF, low fat; LGI, low glycemic index;
LGL, low glycemic load; RCT, randomized clinical trial; T2D, type 2 dia-
betes; WMD, weighted mean difference.
Received June 25, 2017. Accepted for publication November 21, 2017.
First published online April 9, 2018; doi: https://doi.org/10.1093/ajcn/
nqx042.
593
 594 MILAJERDI ET AL.
protein (hs-CRP) concentrations was found after the consump-
tion of LGI and LGL diets compared with after consumption of
HGI and HGL diets. However, that meta-analysis only included
5 studies, in which the effect of an LGI and LGL diet on serum
hs-CRP concentrations was examined (13).
Despite several publications on the effect of dietary GI and GL
on different inflammatory biomarkers, we are aware of no further
study that summarized the findings from previous studies in this
regard. In the current study, we aimed to systematically review the
available evidence about the effects of diets with different GI and
GL on several inflammatory biomarkers in adults and to conduct
a meta-analysis, if possible.
METHODS
Search strategy
We searched for relevant articles published up to June 2017
through PubMed and Medline: https://www.ncbi.nlm.nih.gov/
pubmed, SCOPUS: https://www.scopus.com, EMBASE: https://
www.elsevier.com, and Google Scholar: https://scholar.google.
com. For this purpose, we used MESH (Medical Subject
Heading) and non-MESH keywords including the following:
(“Glycemic index” OR “Glycemic indices” OR “Glycemic
index number” OR “Glycaemic index” OR “Glycaemic in-
dices” OR “Glycaemic index number” OR “Glycemic Load”
OR “Glyceamic Load”) AND (“Inflammation” OR “Inflamma-
tory biomarker” OR “Interleukin-10” OR “Interleukin-8” OR
“Interleukin-6” OR “Tumor necrosis factor” OR “C- reactive pro-
tein” OR “Transforming growth factor beta” OR “Cytokine” OR
“Acute phase reactant” OR “Matrix metalloproteinase” OR “e-
selectin” OR “p-selectin” OR “Intercellular adhesion molecule-
1” OR “Monocyte chemotactic protein 1” OR “Inflammation
mediator” OR “Neurogenic inflammation” OR “Myokine” OR
“Adipokine” OR “Interleukin-1B” OR “Interleukin” OR “Sys-
temic inflammation” OR “Biological marker”). We restricted the
search to clinical trials published in the English language. In
addition, reference lists of all available studies, including re-
view articles, were also examined to avoid missing any publi-
cation. Unpublished data and gray literature, including congress
abstracts, dissertations, and patents, were not included in this
meta-analysis. Moreover, duplicate citations were removed.
Inclusion criteria
Clinical trials that examined the effect of dietary GI and GL
on inflammation in adults were included in this systematic review
and meta-analysis. If several publications with the same data set
were found, only the more complete one was included. In addi-
tion, if a study reported data for specific subgroups, results for the
whole population were used in this analysis.
Exclusion criteria
Studies were excluded if they 1) were conducted in animals,
pregnant women, children, or patients with chronic diseases; 2)
did not perform randomized allocation; 3) were observational
studies including case-control publications; 4) examined the ef-
fect of other interventions along with changes in dietary GI and
GL; 5) had made alterations in dietary GI or GL through the use
Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636
by University of Durham user
on 09 April 2018
of supplements or syrups; 6) had examined only postprandial in-
flammatory responses; and 7) did not have any control group.
Data extraction
Two independent reviewers extracted required data, includ-
ing the following: the first author’s name, year of publication,
participants’ heath status, sample size, participants’ sex, num-
ber of participants in each group, participants’ mean ± SD age,
type of intervention performance (feeding or free living), weight
changes during the intervention, design of clinical trial (parallel
or crossover), type of the intervention and control diets, duration
of the intervention, mean ± SD concentrations of inflammatory
biomarkers after the intervention in different groups, mean ± SD
changes in inflammatory biomarkers after the intervention in dif-
ferent groups, and adjustment for confounders. All reported SEs,
IQRs, and 95% CIs were converted to SDs. When a cytokine con-
centration was reported in different units, we converted them to
the most frequently used unit.
Statistical analysis
Mean differences ± SEs of cytokine concentrations, compar-
ing LGI and LGL to control diets were used to calculate the over-
all effect size. When mean differences ± SEs were not reported,
we calculated them by considering changes in each cytokine con-
centration throughout the study; otherwise, we considered end-
of-trial means in each group to calculate the mean difference ±
SE. The overall effect size was calculated by using a random-
effects model, which takes between-study variation into account.
Cochran’s Q test and I2 statistic were used to assess between-
study heterogeneity. In addition, we used subgroup analysis to
detect probable sources of heterogeneity with the use of a fixed-
effects model. Sensitivity analysis was used to explore the extent
to which inferences might depend on a particular study or group
of studies. Publication bias was examined by visual inspection of
funnel plots and the application of Egger’s and Begg’s tests. All
of the statistical analyses were conducted by using Stata, version
11.2 (StataCorp). P values <0.05 were considered significant.
RESULTS
Study characteristics
Overall, 28 publications were included in this meta-analysis
(11, 14–40). The flow diagram of the study selection is provided
in Figure 1. All of the studies were randomized clinical trials
(RCTs) published between 2002 and 2015. The duration of the
interventions was between 10 d and 48 wk. In total, these studies
enrolled 2961 participants (59% women, 41% men). Character-
istics of these studies are summarized in Table 1.
Among the publications included, 19 studies (11, 14–17, 21–
24, 26–29, 31, 34–36, 38, 39) had a parallel design, whereas the
9 remaining studies were crossover RCTs (18–20, 25, 30, 32, 33,
37, 40). In addition, 3 studies were conducted in patients with
metabolic syndrome (11, 26, 34), 4 studies in those with type
2 diabetes or insulin resistance (16, 17, 31, 37), and 1 study in
women with polycystic ovary syndrome (21).
Ten studies were feeding trials (18–20, 25, 28, 29, 32, 33, 35,
37), 6 other studies had recommended subjects to choose their
 GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION 595

569 Potentially relevant publications identified and screened 6 Articles identified from reference lists

496 Articles excluded on the basis of title/abstract

52 Articles excluded for the following reasons

1. irrelevant (6 studies)
2. on animal models (1 study)
3. observational (12 studies)
4. review (3 studies)
5. changed GI with an individual food (4 studies)
80 Articles were retrieved 6. evaluated postprandial or post exercise biomarkers (5 studies)
for detailed evaluation 7. done on children (2 studies)
8. done on pregnant women (3 studies)
9. reduced dietary GI with a specific diet (7 studies)
10. experimental (1 study)
11. changed dietary GI along with other interventions (3 studies)
12. did not report determined inflammatory biomarkers (5 studies)

28 Randomized clinical trials included in the systematic review

5 Studies included in the meta-analysis of GI and TNF-α

3 Studies included in the meta-analysis of GL and TNF-α

14 Studies included in the meta-analysis of GI and CRP 8 Studies included in the meta-analysis of GI and leptin
7 Studies included in the meta-analysis of GL and CRP 2 Studies included in the meta-analysis of GL and leptin

5 Studies included in the meta-analysis of GI and IL-6

4 Studies included in the meta-analysis of GL and IL-6

FIGURE 1 Flow diagram of study selection. CRP, C-reactive protein; GI, glycemic index; GL, glycemic load.

foods from a list of LGI or LGL foods (semi-feeding) on each
day (15, 16, 22, 26, 34, 40), and in 11 studies participants were
given a recommendation (free-living) to consume LGI foods (14,
17, 21, 22, 24, 27, 30, 31, 36, 38, 39). In 20 studies, an LGI diet
(14–17, 19–27, 33–38, 40), in 7 studies an LGL diet (11, 28–32,
39), and in 1 study an LGI and an LGL regimen (18) were used
as the intervention. Control diets consisted of a high-GI (HGI)
diet (15, 16, 19, 20, 22, 23, 26, 27, 33–36, 40), a high-GL (HGL)
diet (11, 29, 32, 39), a normal diet (14, 21, 37, 38), a low-fat
(LF) diet (25, 30, 31), an LF and HGL diet (28), an HGI and
HGL diet (18), a high-cereal-fiber diet (17), and a low-energy-
density diet (24). Nine studies also had a third intervention arm
(14–16, 22, 23–25, 27, 35), which was not entered in the current
meta-analysis. Among the included studies, 12 publications did
not control their findings for potential confounders (11, 14, 18,
23, 24, 30, 33–35, 38–40).
Biochemical analyses were conducted by using commer-
cial kits, including a radioimmunoassay (DPC, USA) (36) kit,
ELISA [(R&D Systems, Minneapolis, MN) and (invitrogen and
Bio-Rad)] kits (34, 35), a Linco kit (St. Charles, MO) (40) and

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

by University of Durham user
on 09 April 2018
 596

on 09 April 2018
TABLE 1
General characteristics of included studies1
Type of Diet type Outcome

by University of Durham user

dietary Adjustment
First author, Weight interven- Int (name and Cnt (name and Any other or
year (ref) Participants, n Age, y loss tion Design composition) composition) Duration Int group Cnt group intervention Notes on subjects matching2
Bouché, 2002 F: 0; M: 11; 46 ± 33 No Semi- RCT, LGI: carbohydrate HGI: carbohydrate 5 wk LGI—before int: leptin: HGI—before int: leptin: No 11 healthy men, No
(40) both: 11 feeding crossover items with a GI items with a GI 22.88 ± 4.61 ng/mL; 22.42 ± 4.80 ng/mL; age 46 ± 3 y;
<45% were >60% were after int: 19.40 ± 4.27 after int: 20.86 ± 4.21 BMI: 28 ± 1
recommended recommended ng/mL ng/mL
Pereira, 2004 F: 30; M: 9; both: LGL: 28.8 ± Both Feeding RCT, LGL: Providing more LF, HGL: low in fat, 6–10 wk Before int—CRP: 2.8 ± Before int—CRP: 1.9 ± No 39 overweight or 1, 2
(28) 39; LGL diet: 22 6.3; LF, groups parallel than ample high in 2.80 mg/L; after 2.46 mg/L; after int: obese young
(17 F/5 M); HGL: carbohydrate to carbohydrate and int—CRP: 1.0 ± CRP: 1.3 ± 1.64 mg/L; adults
LF diet: 17 32.6 ± 4.3 prevent ketosis, by GL, and consistent 1.40 mg/L; changes changes (percent adj): (BMI ≥27) aged
(13 F/4 M) modifications of both with heart-healthy (percent adj): −47.7 ± −5.1 ± 13.61 mg/L 18–40 y; weight
the amount and type diet of NCEP 11.94 mg/L <135 kg
of carbohydrate guidelines
Thompson, F: 51; M: 9; both: Stand: Both Free-living RCT, Energy deficit of 500 Energy deficit of 48 wk hs-CRP hs-CRP change—stand: Third arm: 90 obese subjects No
2005 (14) 60; Stand: 29 (25 42.0 ± 8.8; groups parallel kcal with an 500 kcal; 2 change—HF/LGI: −10 ± 32 mg/L high-dairy diet (age: 25–70 y,
F/4 M); HF/LGI: HF/LGI: increased amount of servings of dairy −16 ± 40 mg/L BMI: 30–40)
31 (26 F/5 M) 41.1 ± 8.6 fiber and reduction in (4.59 servings/d
GI (1.92 servings/d HGI)
HGI)
Pittas, F: 25; M: 7; both: LGL: 35.0 ± Both Feeding RCT, LGL: 40% HGL: (60% 24 wk CRP—before int: 3.1 ± CRP—before int: 2.2 ± No 32 healthy adults 1, 3
2006 (29) 32; LGL: 16 (12 1.5; HGL: groups parallel carbohydrate, 30% carbohydrate, 20% 2.79 mg/L; changes: 2.39 mg/L; changes: aged 24–42 y;

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

F/4 M); HGL: 34.3 ± 1.2 protein, 30% fat, and protein, 20% fat, −1.44 ± 1.75 mg/L 0.41 ± 3.63 mg/L BMI = 25–29.9
16 (13 F/3 M) mean GI of 53 and and mean GI of 86
GL of 45 g/1000 kcal and GL of 116
g/1000 kcal
McMillan- F: 48; M: 16; LGI: 30.5 ± 4.2– Semi- RCT, LGI: high HGI: high 12 wk LGI changes—leptin: HGI changes—leptin: Third arm: 129 overweight or 1, 2, 4, 5
MILAJERDI ET AL.

Price, 2006 both: 64; LGI: 1.4; HGI: 6.2% feeding parallel carbohydrate; GI = carbohydrate; GI −7.5 ± 8.48 ng/mL; −1.9 ± 8.48 ng/mL; HP/LGI or HGI obese young
(15) 32; HGI: 32 31.8 ± 1.7 40; GL = 65 = 65, GL = 116 CRP: −1.1 ± 2.26 CRP: −0.8 ± 2.26 diets adults, age
mg/L mg/L 18–40 y, BMI
≥25
McMillan- F: 50; M: 15; LGI: 34.6 ± 4.2– Semi- RCT, LGI: HP; GI = 34, HGI: HP; GI = 68, 12 wk LGI changes—leptin: HGI changes—leptin: Third arm: High- 129 overweight or 1, 2, 4, 5
Price, 2006 both: 65; LGI: 1.5; HGI: 6.2% feeding parallel GL = 43 GL = 84 −1.0 ± 8.04 ng/mL; −5.4 ± 8.48 ng/mL; carbohydrate/LGI obese young
(15) 33; HGI: 32 30.2 ± 1.5 CRP: −0.01 ± 2.29 CRP: −0.8 ± 2.26 or HGI diets adults, age
pg/mL pg/mL 18–40 y, BMI
≥25
Clapp, F: 7; M: 0; both: 7 35 ± 8 LGI: Feeding RCT, LGI: GI = 57–63 HGI: GI = 87–95 20 d LGI—after int: leptin: HGI—after int: leptin: No 7 healthy women, No
2007 (33) −23 ± crossover 0.017 ± 0.01 ng/mL; 0.021 ± 0.01 ng/mL; aged 18–55 y;
27 g/d; TNF-α: 2.2 ± 0.8 TNF-α: 2.0 ± 0.6 BMI: 20–35
HGI: pg/mL pg/mL
+10 ±
27 g/d
Wolever, 2008 Both: 103; LGI: LGI: 60.6 ± No Semi- RCT, LGI: GI = 55.1 ± 0.4; HGI: GI = 63.2 ± 48 wk LGI—before int: CRP: HGI—before int: CRP: Third arm: low- 162 participants 1, 2, 6, 7, 8
(16) 48; HGI: 55 1.0; HGI: feeding parallel GL = 133 ± 2 0.4; GL = 135 ± 3 2.64 ± 3.03 mg/L; after 3.34 ± 3.05 mg/L; after carbohydrate with T2D
60.4 ± 1.1 int: CRP: 1.95 ± 0.99 int: CRP: 2.75 ± diet managed by diet
mg/L 1.63 mg/L alone; aged
35–75 y; BMI:
24–40
(Continued)
on 09 April 2018
TABLE 1 Continued
Type of Diet type Outcome

by University of Durham user

dietary Adjustment
First author, Weight interven- Int (name and Cnt (name and Any other or
year (ref) Participants, n Age, y loss tion Design composition) composition) Duration Int group Cnt group intervention Notes on subjects matching2
Jenkins, 2008 F: 82; M: 128; LGI: 60 ± LGI: 3.6 Free-living RCT, LGI: GI = 69.6 HCF: GI = 83.5 24 wk LGI—before int: CRP: HCF—before int: CRP: No 210 participants 2, 3, 7, 8, 9,
(17) both: 210; LGI: 10; HCF: kg; parallel (67.7–71.4); GL = (82.4–84.7), 4.62 mg/L (mean); after 4.59 mg/L; after int: with T2D aged 10, 11, 12,
106 (41 F/65 61 ± 9 HCF: 128.9 (120.5–137.3) GL = 166 int: CRP: 3.02 mg/L; CRP: 2.82 mg/L; 60 ± 10 y and 13
M); HCF: 104 2.3 kg (155.5–176.4) changes: CRP: −1.6 ± changes: −1.8 ± BMI of 30.6 ± 6
(41 F/63 M) 6.82 mg/L 11.18 mg/L in LGI; aged
61 ± 9 y and
BMI of 31.2 ±
5.8 in HCF
Kallio, 2008 F: 23; M: 24; 55.1 ± 6.4 No Semi- RCT, LGI: Subjects replaced HGI: Subjects 12 wk LGI—Before int: IL-1β: HGI—before int: IL-1β: No 47 overweight or No
(34) both: 47; LGI: feeding parallel their normally replaced their 0.19 ± 0.09 pg/mL; 0.23 ± 0.10 pg/mL; obese men and
21; HGI: 26 consumed breads normally IL-10: 4.81 ± IL-10: 4.91 ± women with the
with the wheat consumed 1.19 pg/mL; TNF-α: 1.32 pg/mL; TNF-α: metabolic
breads; 50% of the breads with the 1.29 ± 5.95 pg/mL; 1.63 ± 0.56 pg/mL; syndrome, aged
daily bread oat breads, 50% IL-6: 2.11 ± IL-6: 2.38 ± 55.1 ± 6.4 y;
consumption was to of the daily 1.19 pg/mL; after int: 1.22 pg/mL; after int: BMI: 31.9 ± 3.8
be oat and wheat bread IL-1β: 0.17 ± IL-1β: 0.26 ±
bread consumption 0.09 pg/mL; IL-10: 0.15 pg/mL; IL-10:
was to be 5.19 ± 1.46 pg/mL; 5.23 ± 1.22 pg/mL;
low-fiber TNF-α: 1.46 ± TNF-α: 1.62 ±

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

endosperm rye 0.50 pg/mL; IL-6: 0.50 pg/mL; IL-6:
bread 1.83 ± 0.87 pg/mL 2.95 ± 2.29 pg/mL
Abete, 2008 F: 14; M: 18; 36 ± 7 LGI: Free-living RCT, LGI: GI = 40–45 HGI: GI = 60–65 8 wk LGI changes: leptinadj : HGI changes: leptinadj : No 32 obese subjects, 14, 15
(36) both: 32; LGI: 5.3% ± parallel −8.18 ± 0.69 −8.69 ± 0.85 aged 36 ± 7 y;
16 (8 F/8 M); 2.6%; BMI: 32.5 ± 4.3
HGI: 16 (6 F/10 HGI:
M) 7.5% ±
2.9%
Shikany, 2009 F: 0; M: 24; both: 34.5 ± 8.1 No Feeding RCT, LGI/LGL: mean GI = HGI/HGL: mean 4 wk LGI/GL changes: CRP: High GI/GL No 24 healthy men, No
(18) 24 crossover 49.5; GL = 158.3 GI = 75; GL = 0.0 ± 0.7 mg/L; IL-6: changes—CRP: aged 20–50 y;
245.5 −0.2 ± 1.8 pg/mL; −0.4 ± 2.0 mg/L; IL-6: BMI: 25–33
TNF-α: −0.1 ± −0.1 ± 3.5 pg/mL;
0.7 pg/mL; TNFR-2: TNF-α: 0.1 ±
−70,000.0 ± 210,000.0 0.5 pg/mL; TNFR-2:
ng/L; PAI-1: 400.0 ± −40,000.0 ± 300,000.0
GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION

52,800.0 ng/mL ng/L; PAI-1: 4200.0 ±

30,800.0 ng/mL
Klemsdal, 2010 F: 118; M: 84; LGL: 50.1 ± LGL: Free-living RCT, LGL: 35–40% fat, LF: <30% fat, 48 wk LGL—before int: LF—before int: hs-CRP: No 202 participants, No
(12) both: 202; LGL: 9.3; LF: 4.0 ± parallel 25–30% protein, ∼15% protein, hs-CRP: 4.13 ± 4.28 ± 2.9 mg/L; after aged 30–65 y;
100; LF: 102 49.9 ± 8.4 5.5 kg; 30–35% 55–60% 3.4 mg/L; after int: int: hs-CRP: 3.35 ± BMI of 28–40 in
LF: carbohydrate; advice carbohydrate; hs-CRP: 3.67 ± 3.0 mg/L men and 28–35
4.3 ± to consume LGI without advice 3.7 mg/L in women
6.2 kg carbohydrates to consume LGI
carbohydrates
(Continued)
597
 598
TABLE 1 Continued

on 09 April 2018
Type of Diet type Outcome
dietary Adjustment
First author, Weight interven- Int (name and Cnt (name and Any other or
year (ref) Participants, n Age, y loss tion Design composition) composition) Duration Int group Cnt group intervention Notes on subjects matching2

by University of Durham user

Botero, 2009 F: 0; M: 12; 29.4 ± 4.4 Not re- Feeding RCT, LGI: selecting these HGI: selecting 10 d LGI—after int: CRP: HGI—after int: CRP: No 12 overweight or 16, 17
(19) both: 12 ported crossover carbohydrate- these 0.24 ± 0.20 mg/L 0.26 ± 0.34 mg/L obese male
containing foods: carbohydrate- subjects, aged
uncooked cornstarch, containing 18–35 y; BMI:
pasta, and Guardian foods: corn 27–45
(a commercial cereal) syrup, instant
potatoes, and
Bran Flakes
(Kellogg’s
Australia)
Hartman, 2010 64 M 54.5 ± 7.8 No Feeding RCT, LGI: GI and GL were HGI: GI and GL 4 wk in Change—CRP: −0.25 ± Change—CRP: −0.23 ± No 64 men, aged 1, 7, 8, 18, 19
(20) crossover 38 and 84; contained were 69 and 152 each 1.03 mg/L; sTNFR-1: 0.80 mg/L; MD: 35–75 y; BMI:
∼250 g/d legumes period −35.85 ± 97.68 ng/L; −0.02 ± 1.33 mg/L; 28.7 ± 3.5;
(1.5 cups) sTNFR-2: −71.09 ± sTNFR-1: −42.95 ± of them, 28
296.48 ng/L 95.28 ng/L; MD: subjects with IR
7.09 ± 142.48 ng/L;
sTNFR-2: −95.94 ±
178.0 ng/L; MD:
24.85 ± 356.56 ng/L
Marsh, 2010 F: 49; M: 0; both: LGI: 31.0 ± Both Free-living RCT, LGI: GI = 40, Cnt: GI = 59, 48 wk LGI changes—CRP: Cnt changes—CRP: No 96 women with 1, 8, 20, 21
(21) 49; LGI: 29; 0.7; Cnt: groups parallel GL = 74 GL = 109 −1.2 ± 1.61 mg/L; −0.6 ± 3.13 mg/L; PCOS, aged
Cnt: 20 29.3 ± 0.8 PAI-1: −19.9 ± 51.15 PAI-1: 6.6 ± 36.22 18–40 y; BMI

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

ng/mL ng/mL ≤25
Vrolix, 2010 F: 6; M: 9; F: 50 ± 14; No Semi- RCT, LGL: bread: GI = 40, HGL: bread: GI = 11 wk After int—hs-CRP: After int—hs-CRP: No 15 patients with No
(30) both: 15 M: 55 ± 9 feeding crossover GL = 19 ± 10; drink: 69, GL = 36 ± 2.76 ± 1.85 mg/L; 2.46 ± 1.71 mg/L; MetS, aged
GI = 48, GL = 15 ± 20; drink: GI = TNF-α: 900.0 ± TNF-α: 780.0 ± 30–65 y
1; cake: GI = 20, GL 86, GL = 26 ± 360.0 pg/mL; IL-6: 370.0 pg/mL; IL-6:
MILAJERDI ET AL.

= 1 ± 0; cookie: GI 2; cake: GI = 1410.0 ± 640.0 pg/mL; 1420.0 ± 720.0 pg/mL;

= 37, GL = 1 ± 0 51, GL = 3 ± 1; MCP-1: 178,500.0 ± MCP-1: 174,900.0 ±
cookie: GI = 63, 46,700.0 pg/mL 39,400.0 pg/mL; MD
GL = 3 ± 1; (after int): hs-CRP:
total GL −0.30 ± 1.01 mg/L
difference (95% CI: −0.26,
between 2 0.86 mg/L); TNF-α:
groups: 32 units −120.0 ± 330.0 pg/mL
(−60.0, 300.0 pg/mL);
IL-6: 20.0 ±
520.0 pg/mL (−310.0,
720.0 pg/mL); MCP-1:
2000.0 ± 1690.0 pg/mL
(−16,800.0,
24,100.0 pg/mL)
Jebb, F: 318; M: 230; M: 52 ± 10; LF: 0.8 Semi- RCT, The target differentials between the HGI 24 wk Baseline—CRP: 0.4 ± Baseline—CRP: 0.54 ± Third arm: a 548 men and 2, 3, 6, 7, 22,
2010 (22) both: 548; F: 51 ± 9 kg feeding parallel and LGI groups were 11 and 13 GI 0.45 mg/L; ICAM-1: 0.77 mg/L; ICAM-1: high-SFA and women, aged 23, 24
HM/HGI: 107; points, respectively 242,200 ± 240,000 ± HGI diet 30–70 y; BMI of
HM/LGI: 108 185,925.92 pg/mL; 185,555.55 pg/mL; 28.3 in men and
follow-up—CRP: 0.7 ± follow-up—CRP: 28.6 in women
0.81 mg/L; ICAM-1: 0.65 ± 0.92 mg/L;
242,000 ± ICAM-1: 247,600 ±
179,259.25 pg/mL 182,222.22 pg/mL
(Continued)
on 09 April 2018
TABLE 1 Continued
Type of Diet type Outcome

by University of Durham user

dietary Adjustment
First author, Weight interven- Int (name and Cnt (name and Any other or
year (ref) Participants, n Age, y loss tion Design composition) composition) Duration Int group Cnt group intervention Notes on subjects matching2
Jebb, F: 318; M: 230; M: 52 ± 10; LF: 0.8 Semi- RCT, The target differentials between the HGI 24 wk Baseline—CRP: 0.57 ± Baseline—CRP: 0.5 ± Third arm: a 548 men and 2, 3, 6, 7, 22,
2010 (22) both: 548; F: 51 ± 9 kg feeding parallel and LGI groups were 11 and 13 GI points, 0.76 mg/L; ICAM-1: 0.75 mg/L; ICAM-1: high-SFA and women, aged 23, 24
LF/HGI: 109; respectively 247,000 ± 240,000 ± HGI diet 30–70 y, BMI of
LF/LGI: 119 181,851.85 pg/mL; 181,111.11 pg/mL; 28.3 in men and
follow-up—CRP: 0.6 ± follow-up—CRP: 0.7 ± 28.6 in women
0.70 mg/L; ICAM-1: 0.96 mg/L; ICAM-1:
249,000 ± 242,000 ±
182,962.96 pg/mL 180,370.37 pg/mL
Gögebakan, F: 400; M: 219; <65 8% in- Free-living RCT, LGI HGI: 15 points 26 wk hs-CRP—change: hs-CRP—change: Low protein vs. 773 overweight No
2011 (23) both: 619; LGI: crease parallel higher GI than −0.64 ± 2.10 mg/L −0.18 ± 2.06 mg/L; high protein in adults who lost
309 (200 F/109 in all LGI MD: −0.46 ± each group vs. body weight
M); HGI: 310 partici- 2.96 mg/L; P < 0.001 Cnt with an 8-wk
(200 F/110 M) pants low-calorie diet
before this study
Kelly, F: 13; M: 15; LGI: 67 ± 1; Both Feeding RCT, LGI: a mean GI of 40 HGI: a mean GI 12 wk Changes—TNF-α: Changes—TNF-α: 60 min of aerobic 28 insulin- No
2011 (35) both: 28; LGI: HGI: 66 ± 1 groups parallel of 80 −1.06 ± 1.04 pg/mL; 0.57 ± 1.11 pg/mL; exercise, 5 d/wk resistant
13 (6 F/7 M); IL-6: −1.20 ± IL-6: −0.26 ± sedentary adults,
HGI: 15 1.04 pg/mL; MCP-1: 0.88 pg/mL; MCP-1: aged 66 ± 1 y;
(7 F/8 F) before int: 22.8 ± before int: 20.6 ± BMI: 34.2 ± 0.7

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

22.71 pg/mL; after int: 17.81 pg/mL; after int:
15.2 ± 11.17 pg/mL 21.8 ± 18.97 pg/mL
Zhang, F: 0; M: 64; 35–75 No Feeding RCT, LGI: ∼1.5 cups Cnt: isocaloric 4 wk MD between LGI and MD between LGI and No 36 insulin- 25
2011 (37) both: 64 crossover cooked bean healthy Cnt—leptin: −0.2 ± Cnt—leptin: −0.2 ± sensitive and 28
mixture/2000 kcal American diet 0.96 ng/mL; ghrelin: 0.96 ng/mL; ghrelin: insulin-resistant
−34 ± 176.35 pg/mL; −34 ± 176.35 pg/mL; men, aged
changes—leptin: changes—leptin: 35–75 y
−1.5 ± 2.51 ng/mL; −1.3 ± 2.51 ng/mL;
ghrelin: −15 ± ghrelin: 19 ±
129.84 pg/mL 151.16 pg/mL
Fabricatore, F: 63; M: 16; LGL: 52.8 ± LGL: Free-living RCT, LGL: According to LF: according to 40 wk LGI changes—hs-CRP: LF changes—hs-CRP: No 79 obese adults 26
2011 (31) both: 79; LGL: 1.4; LF: 6.7% ± parallel “LGL Pyramid”; “LF Pyramid”; −2.6 ± 14.54 mg/L −3.3 ± 13.73 mg/L with T2D, aged
40 (32 F/8 M); 52.5 ± 1.3 4.4%; consuming ≤3 and consuming 18–65 y; BMI:
LF: 39 LF: ≤1 servings of ≤30% of energy 27–45
(31 F/8 M) 5.7% ± moderate-GL and from fat
GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION

3.7% HGL items/d

Neuhouser, F: 41; M: 41; All: 29.5 ± No Feeding RCT, LGL: GL = 125, HGL: GL = 250, 4 wk LGL—after int: hs-CRP: HGL—after int: hs-CRP: No 80 healthy 1, 2, 7, 8, 16,
2012 (32) both: 82 8.1 crossover fiber = 55 g/d fiber = 28 g/d 0.6 ± 0.43 mg/L; SAA: 0.6 ± 0.21 mg/L; SAA: participants 18, 27
2.4 ± 1.75 mg/L; IL-6: 2.2 ± 1.53 mg/L; IL-6: (n = 40 with
1.3 ± 0.65 pg/mL; 1.2 ± 0.43 pg/mL; BMI of
leptin: 8.7 ± 3.94 leptin: 9.2 ± 1.97 18.5–24.9;
ng/mL; adiponectin: ng/mL; adiponectin: n = 40 with BMI
4100.00 ± 877.45 3900.00 ± 658.09 of 28.0–40.0);
ng/mL ng/mL aged 18–45 y
(Continued)
599
 600

on 09 April 2018
TABLE 1 Continued
Type of Diet type Outcome
dietary Adjustment
First author, Weight interven- Int (name and Cnt (name and Any other or
year (ref) Participants, n Age, y loss tion Design composition) composition) Duration Int group Cnt group intervention Notes on subjects matching2

by University of Durham user

Heggen, 2012 Both: 181; LGL: LGL: 50.3 ± Both Free-living RCT, parallel LGL: HP and LGI HGL: LF and 12 wk Changes—IL-6: 1.67 ± Changes—IL-6: 1.13 ± No 181 men and No
(39) 93; HGL: 88 9.4; HGL: groups carbohydrates, high carbohydrates 4.81 pg/mL; MCP-1: 4.39 pg/mL; MCP-1: women aged
49.8 ± 8.1 in fiber (30–35% high in fiber 4.14 ± 59.70 pg/mL; 18.50 ± 104.86 pg/mL; 30–65 y; BMI:
carbohydrate) (55–60% PAI-1: 4.47 ± 8.55 PAI-1: 3.96 ± 8.39 28–40; ≥1 MetS
carbohydrate) ng/mL; TNF-α: 0.68 ± ng/mL; TNF-α: 0.87 ± components
1.77 pg/mL; leptin: 2.29 pg/mL; leptin:
5.75 ± 9.19 ng/mL; 5.52 ± 11.93 ng/mL;
adiponectin: 410.0 ± adiponectin: −710.0 ±
4452.80 ng/mL; 4762.20 ng/mL;
resistin: −0.04 ± 4.27 resistin: 1.38 ± 6.84
ng/mL ng/mL
Melanson, 2012 Both: 73; LGI: LGL: 39.1 ± Both Free-living RCT, parallel LGL: GI = 42.43, GL Cnt: LED: GI = 12 wk Changes—LGI: CRP: Changes—Cnt: CRP: Third arm: 116 healthy No
(24) 35; Cnt: 38 7.1; Cnt: groups = 44.75 40.15, GL = −5.40 ± 21.40 mg/L −3.90 ± 22.20 portion- overweight and
38.8 ± 7 54.39 controlled obese adults,
diet aged 38.7 ±
6.7 y; BMI:
31.8 ± 2.2
Ebbeling, 2012 F: 8; M: 13; 30.3 ± 5.7 10–15% Feeding RCT, LGI: moderate GL by LF: had an HGL, 4 wk LGI—before int: leptin: LF—before int: leptin: Third arm: 21 overweight and 2, 4, 7, 16,
(25) both: 21 crossover replacing some grain reduce dietary 29.2 ± 11.98 ng/mL; 29.2 ± 11.98 ng/mL; very-low- obese young 28, 29, 30,
products and starchy fat, emphasize CRP: 1.75 ± CRP: 1.75 ± carbohydrate adults, aged 31
vegetables with whole-grain 4.62 mg/L; after int: 4.62 mg/L; after int: diet 18–40 y, BMI
sources of healthful products, and leptin: 12.7 ± 5.88 leptin: 14.9 ± 6.99 ≥27

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

fat and LGI include a variety ng/mL; CRP: 0.76 ± ng/mL; CRP: 0.78 ±
vegetables, legumes, of vegetables 1.88 mg/L 1.70 mg/L
and fruit and fruit
Giacco, 2013 Both: 123; 40–65 No Semi- RCT, parallel LGI: a fixed amount of HGI: a fixed 12 wk hs-CRP—before int: hs-CRP—before int: No 123 individuals of 2, 7, 8
(26) whole-grain feeding whole-grain cereal amount of 19.5 ± 18.00 mg/L; 19.5 ± 13.00 mg/L; both sexes; age
MILAJERDI ET AL.

(LGI): 62; Cnt products; average GI refined cereal, after int: 13.6 ± after int: 17.4 ± range: 40–65 y,
(HGI): 61 = 46% contained 14.66 mg/L; IL-6: 14.70 mg/L; IL-6: with MetS
commercial before: 1.42 ± before: 1.41 ±
products; 1.23 pg/mL; after: 1.12 pg/mL; after:
average GI = 1.54 ± 1.24 pg/mL; 1.43 ± 1.17 pg/mL;
72% IL-1ra: before: 300 ± IL-1ra: before: 251 ±
271.11 pg/mL; after: 193.70 pg/mL; after:
298 ± 207.77 pg/mL; 239 ± 210.74 pg/mL;
TNF-α: before: 0.73 ± TNF-α: before: 0.62 ±
0.54 pg/mL; after: 0.54 pg/mL; after:
0.68 ± 0.53 pg/mL 0.63 ± 0.48 pg/mL
Juanola- F: 66; M: 15; LGI: 42.5 ± Both Free-living RCT parallel LGI: GI = 34 HGI: GI = 62 24 wk Change—CRP: −0.19 ± Change—CRP: −0.07 ± Third arm: LF and 122 overweight 1, 2, 7, 32
Falgarona, both: 81, HGI: 1.1; HGI: groups 1.31 mg/L; IL-6: 2.02 mg/L; IL-6: HGI diet (LF) and obese adults
2014 (27) 40 (33 F/7 M); 44.0 ± 1.3 −0.27 ± 0.63 pg/mL; 0.12 ± 0.67 pg/mL;
LGI: 41 leptin: −5.64 ± 5.35 leptin: −6.03 ± 5.04
(33 F/8 M) ng/mL; MCP-1: ng/mL; MCP-1:
−5.39 ± 19.27 pg/mL; −2.87 ± 19.28 pg/mL;
adiponectin: 1.95 ± adiponectin: 0.33 ±
16.11 ng/mL; ICAM-1: 19.91 ng/mL; ICAM-1:
0.01 ± 0.06 pg/mL; 0.01 ± 0.10 pg/mL;
VCAM: 0.19 ± VCAM: −0.07 ±
1.28 pg/mL 1.26 pg/mL
(Continued)
 GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION 601

load; HM, high-MUFA; HP, high-protein; hs-CRP, high-sensitivity C-reactive protein; ICAM-1, intracellular adhesion molecule 1; IL-1ra, IL-1 receptor antagonist; int, intervention; IR, insulin resistance; LED,
low energy density; LF, low-fat; LGI, low glycemic index; LGL, low glycemic load; MCP-1, monocyte chemoattractant protein 1; MD, mean difference; MetS, metabolic syndrome; NCEP, National Cholesterol

23, baseline waist circumference; 24, (log) HDL; 25, subjects’ IR status; 26, changes in antidiabetic medications; 27, race/ethnicity; 28, weight after run-in phase; 29, within-participant covariance; 30, correlation
1 BMI is presented as kg/m2 . adj, adjusted; Cnt, control; CRP, C-reactive protein; GI, glycemic index; GL, glycemic load; HCF, high cereal fiber; HF, high fiber; HGI, high glycemic index; HGL, high glycemic

Education Program; PAI-1, plasminogen activator inhibitor 1; PCOS, polycystic ovarian syndrome; RCT, randomized clinical trial; ref, reference; SAA, serum amyloid A; Stand, standard; sTNFR, soluble

2 Adjustments were as follows: 1, baseline values; 2, sex; 3, weight changes; 4, weight; 5, baseline differences; 6, center; 7, age; 8, BMI; 9, carbohydrate or fiber intake; 10, GI; 11, change in fiber; 12, change

in GI; 13, glycated hemoglobin; 14, fat mass; 15, fat-free mass; 16, sequence of diets; 17, potentially confounding dietary factors; 18, period; 19, treatment order; 20, metformin use; 21, weight loss; 22, ethnicity;
a Milliplex Map Plex Kit (Merck Millipore) (27). Furthermore,

Adjustment

matching2
nephelometry (16, 17), reagents (22), immunoturbidimetric and
enzymatic assays (11), immunoassay (20, 38, 39), radioim-

No
or
munoassay and ELISA [(Diagnostic Systems, Webster TX) and
Notes on subjects
26 overweight or
(EZHL-80SK, Linco, Billerica, MA, USA)] (33, 37), the gas

18–35 y; BMI:
obese infertile
women, aged
chromatography–mass spectrometry–single ion monitoring
method (26), microparticle enzyme immunometric assay and

25–40
near-infrared immunonephelometry (15), ELISA kits (R&D Sys-
tems) and nephlometry (18, 32), particle-enhanced turbidimetric
assay (14, 19), immunoturbidimetric assay (28), and solid-phase
intervention

chemiluminescent immunometric assay (29) were other methods

Any other

used to quantify serum concentrations of inflammatory biomark-

No

ers. In addition, 6 studies did not report laboratory tests in

detail (21, 23–25, 30, 31).
−0.30 ± 16.03 ng/mL;

−0.24 ± 31.31 pg/mL

In the LGI and LGL groups, the range for dietary GI was 34–
Cnt changes—leptin:

acylated ghrelin:
Cnt group

71.4, whereas GL varied between 43 and 158.3. The correspond-

ing data for HGI and HGL protocols were 60–95 (GI) and 84–250
(GL).
In 4 studies, a significant reduction in serum hs-CRP concen-
Outcome

trations was reported in the LGI and LGL group compared with
the control diet (14, 23, 28, 29), whereas changes in hs-CRP con-
LGI changes—leptin:

centrations were nonsignificant in other RCTs (11, 15–17, 19–22,

ghrelin: 10.72 ±
−21.43 ± 33.45
ng/mL; acylated
Int group

24–27, 30–32). With regard to TNF-α, only one study reported a

33.93 pg/mL

significant reduction in LGI and LGL group compared with the

control diet (35). Others failed to find significant changes in TNF-
α concentrations (11, 18, 27, 34, 39). Compared with the control
Duration

diet, a significant reduction was seen in leptin concentrations af-

12 wk

ter consumption of an LGI and LGL diet in some (15, 25, 33, 38)
but not all (27, 32, 36, 37, 39, 40) studies. In addition, reductions
their usual diets

in serum IL-6 concentrations were significant only in one study

Cnt: maintained
Cnt (name and
composition)

after consumption of an LGI and LGL diet compared with the

TNF-receptor; T2D, type 2 diabetes; TNFR, TNF-receptor; VCAM, vascular cell adhesion molecule.

control diet (35). Others did not find significant changes (11, 18,
26, 27, 32, 34, 39).
Diet type

Participants’ weight did not change during the intervention in

between 3 daily measures; 31, order of measurement period; 32, antidiabetic medication use.
LGI: hypocaloric diet,

9 studies (11, 16, 18, 20, 26, 32, 34, 37, 40). One study did not
report weight changes (19), and 2 other studies found weight
daily GI <55
Int (name and
composition)

changes only in the LGI group (22, 38).

Findings from the meta-analysis on serum hs-CRP

parallel
Design

concentrations
RCT,

Because 2 publications reported their findings separately by

Free-living

sex, we considered each sex subgroup as an individual study

interven-
Type of
dietary

(15, 22). Combining findings from 14 studies, we found no

tion

significant effect of an LGI diet on serum hs-CRP concentra-

tions compared with the control diet [weighted mean difference
Weight

5.5%

(WMD): −0.05 mg/L; 95% CI: −0.21, 0.10 mg/L; I2 = 64.5%]

LGI: 31.36 ± LGI:
loss

(Figure 2). However, significant between-study heterogeneity

was found. Subgroup analysis by study design (parallel or
0.89; Cnt:
31.25 ±

crossover RCT), sex (male, female, or both sexes), country

Age, y

0.78

3 Mean ± SD (all such values).

(United States or another country), or study duration (<10 wk

or ≥10 wk) did not provide any explanation for between-study
F: 26; M: 0; both:

heterogeneity. However, subgroup analysis by the control group

Participants, n

26; LGI: 14;

(HGI, normal diet, high cereal fiber, LF, or low energy density)
TABLE 1 Continued

Cnt: 12

showed a significant reduction in serum hs-CRP concentrations

after consumption of an LGI diet compared with the HGI diet
(WMD: −0.03 mg/L; 95% CI: −0.08, 0.03 mg/L) (Table 2).
Becker, 2015
First author,

With regard to dietary GL, again we did not find any sig-
year (ref)

nificant effect of an LGL diet on serum hs-CRP concentrations

(38)

compared with the control diet (WMD: 0.08 mg/L; 95% CI:

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

by University of Durham user
on 09 April 2018
 602 MILAJERDI ET AL.

FIGURE 2 Forest plot for the effect of a low–glycemic index diet on serum hs-CRP concentrations, expressed as mean differences between the intervention
and the control diets. The area of each square is proportional to the inverse of the variance of the WMD. Horizontal lines represent 95% CIs. Diamonds rep-
resent pooled estimates from random-effects analysis. 1 High-carbohydrate and low–glycemic index diet compared with high-carbohydrate and high–glycemic
index diet; 2 high-protein and low–glycemic index diet compared with high-protein and high–glycemic index diet; 3 high-MUFA and low–glycemic index diet
compared with high-MUFA and high–glycemic index diet; 4 low-fat and low–glycemic index diet compared with low-fat and high–glycemic index diet. hs-CRP,
high-sensitivity C-reactive protein; WMD, weighted mean difference.

TABLE 2
Subgroup analysis for the effect of LGI diet on serum concentrations of hs-CRP1

Subgroup Effect sizes, n Participants, n Effect size (95% CI) I2 Within2 Between3
Overall 16 2011 −0.05 (−0.21, 0.10) 64.5 <0.001
Study design 0.58
Parallel RCT 12 1890 −0.14 (−0.40, 0.12) 72.4 <0.001
Crossover RCT 4 121 −0.01 (−0.10, 0.08) 0 0.56
Study location 0.84
United States 5 230 −0.02 (−0.11, 0.07) 0 0.97
Country other than United States 11 1781 −0.08 (−0.33, 0.18) 76 <0.001
Participants’ sex 0.66
Male 3 100 −0.01 (−0.11, 0.09) 2.1 0.36
Female 1 49 −0.60 (−2.09, 0.89)
Both sexes 12 1862 −0.12 (−0.36, 0.13) 72 <0.001
Study duration, wk 0.58
≥10 12 1890 −0.14 (−0.40, 0.12) 72.4 <0.001
<10 4 121 −0.01 (−0.10, 0.08) 0 0.56
Control group 0.90
HGI 11 1598 −0.03 (−0.08, 0.03) 75.7 <0.001
Normal diet 2 109 −0.60 (−1.76, 0.56) 0 1.00
HCF 1 210 0.20 (−2.31, 2.71)
LF 1 21 −0.02 (−0.61, 0.57)
LED 1 73 −1.50 (−11.50, 8.50)
1 HCF, high cereal fiber; HGI, high glycemic index; hs-CRP, high-sensitivity C-reactive protein; LED, low energy density; LF, low fat; LGI, low glycemic
index; RCT, randomized controlled trial.
2 P values were obtained by random-effects analysis.
3 P values were obtained by fixed-effects analysis.

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

by University of Durham user
on 09 April 2018
 GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION 603

FIGURE 3 Forest plot for the effect of a low–glycemic load diet on serum hs-CRP concentrations, expressed as mean differences between the intervention
and the control diets. The area of each square is proportional to the inverse of the variance of the WMD. Horizontal lines represent 95% CIs. Diamonds represent
pooled estimates from random-effects analysis. hs-CRP, high-sensitivity C-reactive protein; WMD, weighted mean difference.

−0.26, 0.42 mg/L; I2 = 66.3%). Due to an inadequate num- Publication bias

ber of studies, we did not perform any subgroup analysis in this
case (Figure 3).
Findings from the meta-analysis on serum TNF-α, leptin,
and IL-6 concentrations
We did not find significant changes in serum TNF-α con-
centrations between LGI and control diets after combining ef-
fect sizes from 5 studies (WMD: −0.18 mg/L; 95% CI: −0.43,
0.06 mg/L; I2 = 80.1%). Significant changes were also not seen
in leptin (WMD: −0.53 mg/L; 95% CI: − 1.08, 0.02 mg/L;
I2 = 66.3%) and IL-6 (WMD: −0.39 mg/L; 95% CI: −0.81,
0.03 mg/L; I2 = 80.4%) concentrations between the LGI diet
and the control diet (Figure 4). In addition to dietary GI, no sig-
nificant differences in serum TNF-α concentrations were found
comparing participants who consumed an LGL diet with those
who were consuming a control diet (WMD: −0.20 mg/L; 95%
CI: −0.33, 0.07 mg/L; I2 = 0.0%). The same results were
also found for leptin (WMD: −0.48 mg/L; 95% CI: −0.98,
0.02 mg/L; I2 = 0.0%) and IL-6 (WMD: 0.10 mg/L; 95% CI:
0.02, 0.18 mg/L; I2 = 0.0%) concentrations (Figure 5).
Sensitivity analysis
To investigate the influence of each individual study on the
overall findings, we excluded studies from the analysis, stage by
stage. We found no significant impact of any individual study on
the overall effect size for studies either on GI or GL.
The funnel plots indicated moderate asymmetry, suggesting
that publication bias cannot be completely excluded as a factor
of influence on the present meta-analysis (data not shown). How-
ever, the Begg’s (P = 0.20) and Egger’s (P = 0.49) regression
tests provided no evidence of substantial publication bias.
DISCUSSION
In the current study, we failed to find significant effects of di-
etary GI or GL on inflammatory cytokines, including hs-CRP,
leptin, IL-6, and TNF-α concentrations. To the best of our knowl-
edge, this is the first study summarizing earlier publications on
the effect of GI and GL on serum concentrations of inflammatory
cytokines, including leptin, IL-6, and TNF-α; however, a previ-
ous meta-analysis in 2013 investigated the effect of dietary GI
and GL on serum CRP concentrations. Findings from that study
were different from those of our meta-analysis; the authors found
a significant reduction in serum CRP concentrations after the con-
sumption of an LGI and LGL diet compared with an HGI and
HGL diet (WMD: −0.43 mg/dL; 95% CI: − 0.78, −0.09 mg/dL;
P = 0.01) (12), respectively. It should be noted that only 5 stud-
ies, which assessed the effect of an LGI and LGL diet on serum
CRP concentrations, were included in that meta-analysis.
Chronic low-grade inflammation might be involved in insulin
resistance and endothelial dysfunction (41), which may sub-
sequently cause obesity-linked chronic diseases such as T2D
and CVD (42). CRP is an indicator of general low-grade in-
flammation (43, 44). Macrophage infiltration into adipose tis-
sue, which occurs as a result of obesity, increases the production
and release of macrophage-derived proinflammatory cytokines,

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

by University of Durham user
on 09 April 2018
 604 MILAJERDI ET AL.

FIGURE 4 Forest plot for the effect of a low–glycemic index diet on serum concentrations of TNF-α, leptin, and IL-6, expressed as mean differences
between the intervention and the control diets. The area of each square is proportional to the inverse of the variance of the WMD. Horizontal lines represent 95%
CIs. Diamonds represent pooled estimates from random-effects analysis. 1 High-carbohydrate and low–glycemic index diet compared with high-carbohydrate
and high–glycemic index diet; 2 high-protein and low–glycemic index diet compared with high-protein and high–glycemic index diet. hs-CRP, high-sensitivity
C-reactive protein; WMD, weighted mean difference.

including CRP (45, 46). Dietary GI was associated with inflam-
mation in some earlier studies (5, 6), whereas others reached
no significant links (8). Similar contradictory results were found
for dietary GL, which is a measure of carbohydrate quality and
quantity (9, 11). For instance, Qi and Hu (11) found a direct
association between dietary GL and systemic inflammation in
patients with T2D. However, other studies did not find such
associations (10, 31).
In addition to hs-CRP concentrations, we failed to find a sig-
nificant effect of dietary GI and GL on serum concentrations
of leptin, IL-6, and TNF-α. Elevated concentrations of proin-
flammatory cytokines, including IL-6 and TNF-α, can be ex-
plained by macrophage infiltration, similar to CRP (47, 48).
Findings from previous studies on the effect of dietary GI
and GL on serum IL-6 and TNF-α concentrations are con-
flicting. In a parallel RCT, Kelly et al. (35) reported signifi-
cant reductions in serum IL-6 and TNF-α concentrations af-
ter consumption of an LGI diet, compared with an HGI regi-
men, whereas most other studies did not find such significant
effects (11, 18, 27).
We did not find a significant change in serum leptin con-
centrations after consumption of an LGI and LGL diet com-
pared with the control diet. Leptin is an adipokine synthesized
mostly by white adipose tissue cells (49). Leptin is known as
a hormone that regulates food intake and energy balance (50).
However, it also plays a dual role in inflammatory processes
by interacting with the immune system (51). Leptin activates
monocytes and macrophages to release proinflammatory cy-
tokines including TNF-α and IL-6 (52). Conversely, leptin
also exerts anti-inflammatory activities by inducing IL-1 recep-
tor antagonist and increasing IL-4 production (52, 53). Con-
flicting findings are available on the effect of an LGI and
LGL diet on serum leptin concentrations. Although signifi-
cant reductions in serum leptin concentrations were reported in
some publications (15, 25, 33, 38), this was not the case in
others (27, 32, 36, 37, 39, 40).
This is the first systematic review and meta-analysis, to our
knowledge, on the effect of dietary GI and GL on inflammatory
cytokines, including leptin, IL-6, and TNF-α, and an updated
meta-analysis on the effect of dietary GI and GL on serum CRP
concentrations. Egger’s regression tests for the effect of an LGI
and LGL diet provided no evidence of substantial publication bias
in this meta-analysis. However, some limitations must be kept in
mind, such as different methods used to prescribe the LGI and
LGL diets in different studies. In addition, different diets, includ-
ing HGI and HGL or normal diets, were considered control diets
in different studies. Moreover, the lack of controlling for differ-
ent confounders and the use of different study designs may also
explain the different findings.
In conclusion, we did not find a significant effect of dietary GI
and GL on serum concentrations of inflammatory cytokines, in-
cluding CRP, leptin, IL-6, and TNF-α in adults. Additional RCTs,

Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636

by University of Durham user
on 09 April 2018
 GLYCEMIC INDEX, GLYCEMIC LOAD, AND INFLAMMATION
FIGURE 5 Forest plot for the effect of a low–glycemic load diet on serum concentrations of TNF-α, leptin, and IL-6, expressed as mean differences
between the intervention and the control diets. The area of each square is proportional to the inverse of the variance of the WMD. Horizontal lines represent
95% CIs. Diamonds represent pooled estimates from random-effects analysis. WMD, weighted mean difference.
in particular feeding trials, are warranted to shed light on this
issue.
The authors’ responsibilities were as follows—AM, PS, and AE: con-
ducted the research; PS: analyzed the data; AM and AE: wrote the manuscript;
AE: had primary responsibility for the final content; and all authors: designed
the research, and read and approved the final manuscript. None of the authors
had a conflict of interest to declare.
REFERENCES
1. Rose-John S, Schooltink H. Cytokines are a therapeutic target for the
prevention of inflammation-induced cancers. Recent Results Cancer
Prev 2007;174:57–66.
2. Goncalves de Carvalho CM, Ribeiro SM. Aging, low-grade systemic
inflammation and vitamin D: a mini-review. Eur J Clin Nutr
2017;71:434–40.
3. Vasto S, Candore G, Balistreri CR, Caruso M, Colonna-Romano G,
Grimaldi MP, Listi F, Nuzzo D, Lio D, Caruso C. Inflammatory
networks in ageing, age-related diseases and longevity. Mech Ageing
Dev 2007;128:83–91.
4. Nicklas BJ, You T, Pahor M. Behavioural treatments for chronic
systemic inflammation: effects of dietary weight loss and exercise
training. Can Med Assoc J 2005;172:1199–209.
5. Du H, Van Der A DL, Feskens EJ. Dietary glycaemic index: a review
of the physiological mechanisms and observed health impacts. Acta
Cardiol 2006;61:393–7.
6. Buyken AE, Goletzke J, Joslowski G, Felbick A, Cheng G,
Herder C, Brand-Miller JC. Association between carbohydrate quality
and inflammatory markers: systematic review of observational and
interventional studies. Am J Clin Nutr 2014;99:813–33.
7. Liu S, Willett WC, Stampfer MJ, Hu FB, Franz M, Sampson L,
Hennekens CH, Manson JE. A prospective study of dietary glycemic
load, carbohydrate intake, and risk of coronary heart disease in US
women. Am J Clin Nutr 2000;71:1455–61.
8. Du H, van Bakel MM, van der Kallen CJ, Blaak EE, van Greevenbroek
MM, Jansen EH, Nijpels G, Stehouwer CD, Dekker JM, Feskens EJ.
Glycemic index and glycemic load in relation to food and nutrient
Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636
by University of Durham user
on 09 April 2018
605
intake and metabolic risk factors in a Dutch population. Am J Clin Nutr
2008;87:655–61.
9. Bulló M, Casas R, Portillo MP, Basora J, Estruch R, García-Arellano
A, Lasa A, Juanola-Falgarona M, Arós F, Salas-Salvadó J. Dietary
glycemic index/load and peripheral adipokines and inflammatory
markers in elderly subjects at high cardiovascular risk. Nutr Metab
Cardiovasc Dis 2013;23:443–50.
10. Mager DR, Iñiguez IR, Gilmour S, Yap J. The effect of a low
fructose and low glycemic index/load (FRAGILE) dietary intervention
on indices of liver function, cardiometabolic risk factors, and
body composition in children and adolescents with nonalcoholic
fatty liver disease (NAFLD). J Parenter Enteral Nutr 2015;39:
73–84.
11. Qi L, Hu FB. Dietary glycemic load, whole grains, and systemic
inflammation in diabetes: the epidemiological evidence. Curr Opin
Lipidol 2007;18:3–8.
12. Klemsdal TO, Holme I, Nerland H, Pedersen TR, Tonstad S. Effects of a
low glycemic load diet versus a low-fat diet in subjects with and without
the metabolic syndrome. Nutr Metab Cardiovasc Dis 2010;20:195–201.
13. Schwingshackl L, Hoffmann G. Long-term effects of low glycemic
index/load vs. high glycemic index/load diets on parameters of obesity
and obesity-associated risks: a systematic review and meta-analysis.
Nutr Metab Cardiovasc Dis 2013;23:699–706.
14. Thompson WG, Holdman NR, Janzow DJ, Slezak JM, Morris KL,
Zemel MB. Effect of energy-reduced diets high in dairy products and
fiber on weight loss in obese adults. Obes Res 2005;13:1344–53.
15. McMillan-Price J, Petocz P, Atkinson F, O’Neill K, Samman S,
Steinbeck K, Caterson I, Brand-Miller J. Comparison of 4 diets of
varying glycemic load on weight loss and cardiovascular risk reduction
in overweight and obese young adults: a randomized controlled trial.
Arch Intern Med 2006;166:1466–75.
16. Wolever TM, Gibbs AL, Mehling C, Chiasson J-L, Connelly PW,
Josse RG, Leiter LA, Maheux P, Rabasa-Lhoret R, Rodger NW. The
Canadian Trial of Carbohydrates in Diabetes (CCD), a 1-y controlled
trial of low-glycemic-index dietary carbohydrate in type 2 diabetes: no
effect on glycated hemoglobin but reduction in C-reactive protein. Am
J Clin Nutr 2008;87:114–25.
17. Jenkins DJ, Kendall CW, McKeown-Eyssen G, Josse RG, Silverberg
J, Booth GL, Vidgen E, Josse AR, Nguyen TH, Corrigan S. Effect of
 606 MILAJERDI ET AL.
a low–glycemic index or a high–cereal fiber diet on type 2 diabetes: a
randomized trial. J Am Med Assoc 2008;300:2742–53.
18. Shikany JM, Phadke RP, Redden DT, Gower BA. Effects of low-and
high-glycemic index/glycemic load diets on coronary heart disease risk
factors in overweight/obese men. Metabolism 2009;58:1793–801.
19. Botero D, Ebbeling CB, Blumberg JB, Ribaya-Mercado JD, Creager
MA, Swain JF, Feldman HA, Ludwig DS. Acute effects of dietary
glycemic index on antioxidant capacity in a nutrient-controlled feeding
study. Obesity 2009;17:1664–70.
20. Hartman TJ, Albert PS, Zhang Z, Bagshaw D, Kris-Etherton PM,
Ulbrecht J, Miller CK, Bobe G, Colburn NH, Lanza E. Consumption
of a legume-enriched, low-glycemic index diet is associated with
biomarkers of insulin resistance and inflammation among men at risk
for colorectal cancer. J Nutr 2010;140:60–7.
21. Marsh KA, Steinbeck KS, Atkinson FS, Petocz P, Brand-Miller
JC. Effect of a low glycemic index compared with a conventional
healthy diet on polycystic ovary syndrome. Am J Clin Nutr 2010;92:
83–92.
22. Jebb SA, Lovegrove JA, Griffin BA, Frost GS, Moore CS, Chatfield
MD, Bluck LJ, Williams CM, Sanders TAB. Effect of changing the
amount and type of fat and carbohydrate on insulin sensitivity and
cardiovascular risk: the RISCK (Reading, Imperial, Surrey, Cambridge,
and Kings) trial. Am J Clin Nutr 2010;92:748–58.
23. Gögebakan Ö, Kohl A, Osterhoff MA, van Baak MA, Jebb SA,
Papadaki A, Martinez JA, Handjieva-Darlenska T, Hlavaty P, Weickert
MO. Effects of weight loss and long-term weight maintenance with diets
varying in protein and glycemic index on cardiovascular risk factors:
clinical perspective. Circulation 2011;124:2829–38.
24. Melanson KJ, Summers A, Nguyen V, Brosnahan J, Lowndes J,
Angelopoulos TJ, Rippe JM. Body composition, dietary composition,
and components of metabolic syndrome in overweight and obese adults
after a 12-week trial on dietary treatments focused on portion control,
energy density, or glycemic index. Open Nutr J 2012;11:57.
25. Ebbeling CB, Swain JF, Feldman HA, Wong WW, Hachey DL,
Garcia-Lago E, Ludwig DS. Effects of dietary composition on
energy expenditure during weight-loss maintenance. J Am Med Assoc
2012;307:2627–34.
26. Giacco R, Lappi J, Costabile G, Kolehmainen M, Schwab U, Landberg
R, Uusitupa M, Poutanen K, Pacini G, Rivellese AA. Effects of
rye and whole wheat versus refined cereal foods on metabolic risk
factors: a randomised controlled two-centre intervention study. Clin
Nutr 2013;32:941–9.
27. Juanola-Falgarona M, Salas-Salvadó J, Ibarrola-Jurado N, Rabassa-
Soler A, Díaz-López A, Guasch-Ferré M, Hernández-Alonso P, Balanza
R, Bulló M. Effect of the glycemic index of the diet on weight loss,
modulation of satiety, inflammation, and other metabolic risk factors: a
randomized controlled trial. Am J Clin Nutr 2014;100:27–35.
28. Pereira MA, Swain J, Goldfine AB, Rifai N, Ludwig DS. Effects of a
low–glycemic load diet on resting energy expenditure and heart disease
risk factors during weight loss. J Am Med Assoc. 2004;292:2482–90.
29. Pittas AG, Roberts SB, Das SK, Gilhooly CH, Saltzman E, Golden J,
Stark PC, Greenberg AS. The effects of the dietary glycemic load on
type 2 diabetes risk factors during weight loss. Obesity 2006;14:2200–
9.
30. Vrolix R, Mensink RP. Effects of glycemic load on metabolic risk
markers in subjects at increased risk of developing metabolic syndrome.
Am J Clin Nutr 2010;92:366–74.
31. Fabricatore A, Wadden T, Ebbeling C, Thomas J, Stallings V, Schwartz
S, Ludwig DS. Targeting dietary fat or glycemic load in the treatment
of obesity and type 2 diabetes: a randomized controlled trial. Diabetes
Res Clin Pract 2011;92:37–45.
32. Neuhouser ML, Schwarz Y, Wang C, Breymeyer K, Coronado G,
Wang C-Y, Noar K, Song X, Lampe JW. A low-glycemic load diet
reduces serum C-reactive protein and modestly increases adiponectin
in overweight and obese adults. J Nutr 2012;142:369–74.
33. Clapp JF, Lopez B. Low-versus high-glycemic index diets in women:
effects on caloric requirement, substrate utilization and insulin
sensitivity. Metab Syndr Relat Disord 2007;5:231–42.
Downloaded from https://academic.oup.com/ajcn/article-abstract/107/4/593/4964636
by University of Durham user
on 09 April 2018
34. Kallio P, Kolehmainen M, Laaksonen DE, Pulkkinen L, Atalay M,
Mykkänen H, Uusitupa M, Poutanen K, Niskanen L. Inflammation
markers are modulated by responses to diets differing in postprandial
insulin responses in individuals with the metabolic syndrome. Am J Clin
Nutr 2008;87:1497–503.
35. Kelly KR, Haus JM, Solomon TP, Patrick-Melin AJ, Cook M, Rocco
M, Barkoukis H, Kirwan JP. A low-glycemic index diet and exercise
intervention reduces TNF-α in isolated mononuclear cells of older,
obese adults. J Nutr 2011;141:1089–94.
36. Abete I, Parra D, Martinez JA. Energy-restricted diets based on a distinct
food selection affecting the glycemic index induce different weight loss
and oxidative response. Clin Nutr 2008;27:545–51.
37. Zhang Z, Lanza E, Ross A, Albert P, Colburn N, Rovine M, Bagshaw
D, Ulbrecht JS, Hartman TJ. A high-legume low-glycemic index diet
reduces fasting plasma leptin in middle-aged insulin-resistant and -
sensitive men. Eur J Clin Nutr 2011;65:415–8.
38. Becker GF, Passos EP, Moulin CC. Short-term effects of a hypocaloric
diet with low glycemic index and low glycemic load on body adiposity,
metabolic variables, ghrelin, leptin, and pregnancy rate in overweight
and obese infertile women: a randomized controlled trial. Am J Clin
Nutr 2015;102:1365–72.
39. Heggen E, Klemsdal TO, Haugen F, Holme I, Tonstad S. Effect of a
low-fat versus a low-glycemic-load diet on inflammatory biomarker and
adipokine concentrations. Metab Syndr Relat Disord 2012;10:437–42.
40. Bouché C, Rizkalla SW, Luo J, Vidal H, Veronese A, Pacher
N, Fouquet C, Lang V, Slama G. Five-week, low–glycemic index
diet decreases total fat mass and improves plasma lipid profile in
moderately overweight nondiabetic men. Diabetes Care 2002;25:
822–8.
41. Stulnig TM. Inflammation as a trigger for insulin resistance and
cardiometabolic disease. Vienna (Austria); Springer; 2015.
42. Esser N, Legrand-Poels S, Piette J, Scheen AJ, Paquot N. Inflammation
as a link between obesity, metabolic syndrome and type 2 diabetes.
Diabetes Res Clin Pract 2014;105:141–50.
43. Devaraj S, Singh U, Jialal I. Human C-reactive protein and the metabolic
syndrome. Curr Opin Lipidol 2009;20:182.
44. Ridker PM, Morrow DA. C-reactive protein, inflammation, and
coronary risk. Cardiol Clin 2003;21:315–25.
45. Lee D-E, Kehlenbrink S, Lee H, Hawkins M, Yudkin JS. Getting the
message across: mechanisms of physiological cross talk by adipose
tissue. Am J Physiol Endocrinol Metab 2009;296:1210–29.
46. Xu H, Barnes GT, Yang Q, Tan G, Yang D, Chou CJ, Sole J, Nichols A,
Ross JS, Tartaglia LA. Chronic inflammation in fat plays a crucial role
in the development of obesity-related insulin resistance. J Clin Invest
2003;112:1821–30.
47. Ono M, Torisu H, J-i Fukushi, Nishie A, Kuwano M. Biological
implications of macrophage infiltration in human tumor angiogenesis.
Cancer Chemother Pharmacol 1999;43:69–71.
48. Kanda H, Tateya S, Tamori Y, Kotani K, Hiasa K-i, Kitazawa R,
Kitazawa S, Miyachi H, Maeda S, Egashira K. MCP-1 contributes
to macrophage infiltration into adipose tissue, insulin resistance, and
hepatic steatosis in obesity. J Clin Invest 2006;116:1494–505.
49. Fantuzzi G. Adipose tissue, adipokines, and inflammation. J Allergy
Clin Immunol 2005;115:911–9.
50. Berthoud HR. Homeostatic and non-homeostatic pathways involved in
the control of food intake and energy balance. Obesity 2006;14:197–
200.
51. Lago R, Gómez R, Lago F, Gómez-Reino J, Gualillo O. Leptin
beyond body weight regulation—current concepts concerning its role in
immune function and inflammation. Cell Immunol 2008;252:139–45.
52. Agrawal S, Gollapudi S, Su H, Gupta S. Leptin activates human
B cells to secrete TNF-α, IL-6, and IL-10 via JAK2/STAT3 and
p38MAPK/ERK1/2 signaling pathway. J Clin Immunol 2011;31:472–
8.
53. Patraca I, Martínez N, Busquets O, Martí A, Pedrós I, Beas-Zarate C,
Marin M, Ettcheto M, Sureda F, Auladell C. Anti-inflammatory role of
leptin in glial cells through p38 MAPK pathway inhibition. Pharmacol
Rep 2017;69:409–18.