Content uploaded by Rodrigo Cauduro Oliveira Macedo
Author content
All content in this area was uploaded by Rodrigo Cauduro Oliveira Macedo on Jul 26, 2018
Content may be subject to copyright.
Effects of fructose consumption on postprandial TAG: an update on systematic
reviews with meta-analysis
Rodrigo C. O. Macedo*, Alexandra F. Vieira, Cesar E. J. Moritz and Alvaro Reischak-Oliveira
Grupo de Estudos em Fisiologia e Bioquímica do Exercício (GEFEX), Universidade Federal do Rio Grande do Sul (UFRGS),
Porto Alegre, RS, 90960-200, Brazil
(Submitted 20 November 2017 –Final revision received 30 April 2018 –Accepted 7 May 2018)
Abstract
The aim of this study was to re-examine the chronic effect (>7 d) of fructose consumption on postprandial TAG, in adolescents and adults. The
research was carried out in March 2017 and used different electronic databases, such as Medline
®
(Pubmed
®
), Embase
®
and Cochrane. The
review considered clinical trials (parallel or crossed) that evaluated the effect of fructose consumption for a period longer than 7 d, in humans.
Two investigators independently performed data extraction. The outcome was the absolute delta of TAG concentration in a 4-h postprandial
period. The results were presented with delta mean difference between treatments with 95 % CI. The calculations were made based on
random-effect models. Statistical heterogeneity of treatment effects between studies was assessed by Cochrane’s‘Q Test’and ‘I
2
’inconsistency
test. The meta-analysis of the twelve selected interventions (n318) showed that fructose generated larger variation (δ) of TAG concentrations
during the postprandial period, compared with other carbohydrates (mean difference: 8·02 (95 % CI 0·46, 15·58) mg/dl (0·09 (95 % CI 0·01,
0·18) mmol/l); I
2
: 74 %). High heterogeneity was generated almost exclusively by one study, and its withdrawal did not alter the result. We
concluded that chronic consumption of fructose (>7 d) has a negative role on postprandial TAG in healthy adolescents and adults, as well as in
overweight/obese individuals, but not in diabetics.
Key words: Fructose: Postprandial lipaemia: Sugar: Fat metabolism: Meta-analyses
Postprandial lipaemia (PPL) is a complex and dynamic process
that involves alteration of lipids and lipoproteins after one or
more meals
(1)
. Since 1947, it has been suggested that PPL plays
an atherogenic role
(2)
and, consequently, is related to the
pathogenesis of CVD
(3–7)
. Exaggerated elevation of TAG in the
postprandial period represents an abnormal response from the
metabolism and is associated with increased morbidity and
mortality
(8,9)
owing to reduced sensitivity to insulin
(10)
and
endothelial dysfunction by increasing oxidative stress
(11)
.
Several diet factors, such as sugar-rich meals, can worsen
postprandial hypertriacylglycerolaemic response
(5,12–14)
.Among
these, fructose stands out. Several studies have shown the effect of
fructose-rich diets on postprandial TAG increase
(15,16)
–dose-
related for quantities above 50 g/d
(17)
.Themechanismsseemto
be associated with the stimulation of liver lipogenesis
(14,18)
,
reduction of sensitivity to insulin
(18,19)
and secretion or reduction
of VLDL-TAG clearance
(20)
. For this connection with the increase
of atherogenic lipids and lipoproteins, it was suggested that
fructose has an indirect role in increased CVD risk
(8)
.
The effects of fructose consumption on fasting
(17,21,22)
and
postprandial TAG (in acute and chronic forms) have been
previously analysed by some systematic reviews with meta-
analysis
(17,23,24)
. Two of these studies
(21,24)
presented important
conflicts of interest declared by investigators arising from
funding by the food industry and showed negative effects only
when it contributes to excess of energy in the diet. Therefore,
the aim of this study was to re-examine the chronic effect (>7d)
fructose consumption has on TAG during postprandial period,
in adolescents and adults.
Methods
The entire process used in this study was elaborated in accor-
dance with the guidelines presented in Preferred Reporting
Items for Systematic Reviews and Meta-Analyses
(25,26)
. This
review was registered at http://www.crd.york.ac.uk/prospero
as CRD42017059987.
Eligibility criteria
The review considered studies in humans, such as clinical trials
(parallel or cross-over designs) that evaluated the effect of
fructose consumption (dissolved in liquid or added to some
food and preparation) over a period of >7 d. The intervention
should be compared with any other carbohydrate that does not
contain fructose in the chemical composition, and there is no
requirement for energy balance between the comparisons.
Research was not limited by illnesses or exercise. The effect of
fructose consumption on the postprandial TAG concentration
*Corresponding author: R. C. O. Macedo, email nutricionistarodrigomacedo@gmail.com
British Journal of Nutrition, page 1 of 9 doi:10.1017/S0007114518001538
© The Authors 2018
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
was evaluated by comparing it with another carbohydrate
(fructose-free) in hyperenergetic or isoenergetic conditions. Stu-
dies evaluating acute intervention were discarded. In cases of
studies with several publications, only one was included.
Research strategy
ThesearchwascarriedoutinMarch2017anduseddifferent
electronic databases such as Medline
®
(Pubmed
®
), Embase
®
and
Cochrane, and manual from the references of studies included.
Theresearchwascomposedofandassociatedwiththefollowing
terms (and their respective related terms): ‘fructose’,‘triglycerides’,
‘hyperlipidemia’.Toexpandtheresearch,therewasnodelimita-
tion of type or year of the study. Studies were limited to English,
Portuguese and Spanish languages. Research strategy is detailed
and available as online Supplementary Material.
Selection of studies
Two investigators (A. F. V. and C. E. J. M.) assessed titles and
abstracts independently from all studies found during the
research. Whenever the abstract did not provide sufficient
information about inclusion and exclusion criteria, the full
article was evaluated. Thereafter, the full study was evaluated
and selected by the reviewers independently. The selection of
studies was based on previously adopted eligibility criteria.
Disagreements were settled by consensus, and in the case of
continuing disagreement the evaluation was made by a third
investigator (R. C. O. M.). Sampling duplication criteria was
controlled by screening the period and place of recruitment,
and authors were contacted for clarification when necessary.
Data extraction
Standardised form using the software Microsoft Office Excel
®
was adopted for proper data extraction, executed indepen-
dently by two reviewers (A. F. V. and C. E. J. M.). The main
features of the studies selected, such as author, year of
publication, population and sample, methods, intervention,
outcome and results, were written in detail. Eventual
disagreements were settled by consensus by a third investigator
(R. C. O. M.). Missing data were requested to the corresponding
author of the study. In case of no answer, denying provision or
data loss, the study was excluded. For data that were presented
only graphically, the results were extracted using DigitizeIt
®
software (I. Bormann). The studies in which the comparison
was not made with carbohydrates and/or acute intervention
(<7 d), as well as those using intravenous carbohydrate infu-
sion, were excluded.
The outcome was the absolute delta of TAG concentration in
a 4-h postprandial period. The deltas were calculated from peak
values (4 h) and basal values (immediately before breakfast) for
properly representing the postprandial TAG curve
(6,27–29)
.
Values in mmol/l were transformed to mg/dl, multiplying
by 88·5. Standard δdeviation was imputed by the equation
proposed by Higgins & Green
(31)
.
Studies with two or more comparison or intervention groups
within the same sample were included only once. When the
study presented more than one comparison component
(another carbohydrate) for the intervention (fructose),
the data were extracted only by the following priority:
starch >glucose.
Evaluation of bias risk
Evaluation of the methodological quality of the studies included
proper randomisation generation, allocation concealment,
blinding participants and/or therapist, blinding the assessors of
outcomes and description of losses and exclusions, as proposed
by Cochrane
(31)
. When these characteristics were described in
the published document, it was considered that criteria were
met and that they were satisfied and classified as ‘low risk’and
otherwise as ‘high risk’. The studies that did not describe these
data were classified as ‘unclear risk’. Two reviewers (A. F. V.
and C. E. J. M.) carried out quality assessment independently.
Data analysis
The results were presented with mean δdifferences among
treatments with 95 % CI. Mean difference expresses the differ-
ence of the intervention effect, when outcome values are
standardised. The calculations were made based on random-
effect models. Statistic treatment effect heterogeneity between
studies was evaluated by Cochrane’s‘Q’test and ‘I
2
’incon-
sistency test, where it was considered that values higher than
50 % indicated high heterogeneity
(32)
. Meta-analysis included
comparison of fructose consumption with any other carbohy-
drate (without fructose in composition) on the variation of
postprandial TAG (expressed by the delta values) immediately
before breakfast (0) and at the 4-h peak. The value of α≤0·05
was considered statistically significant.
The following sensitivity analyses were carried out: funding
source, randomisation, energy balance, form of the carbohy-
drate provided, type of comparison (comparator), time of
intervention (follow-up), amount of carbohydrate provided and
length of analysis. The analysis of period >12 h was also per-
formed by the 4-h data extraction, because not all studies
showed the peak values in longer periods. The use of values
other than the delta peak (in this case >12 h) would create a
confounding factor by the use of another way to measure
intervention v. control effect. Subdivision by amounts larger or
smaller than 87 g was based in the 95th percentile of fructose
consumption (p95) by the American population
(33)
. The soft-
ware Review Manager version 5.3 (Cochrane Collaboration)
was used.
Furthermore, the funnel plot of the variable analysed was
carried out to verify the bias of publication. Asymmetry was
tested by the Begg and Egger test, being considered meaningful
when P<0·10. The trim-and-fill test was used to estimate the
publication’s bias effects in interpreting results. The software
comprehensive meta-analysis version 2.0 was used.
Results
Research results
A total of 3337 studies were identified as eligible in the database
search. After duplicates were removed, there were 2805 studies
2 R. C. O. Macedo et al.
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
remaining. Of these, it was determined that 2797 were irrelevant
based on title and/or summary review, and there were eight
studies left. After integral reading, eight studies
(18,34–40)
and
twelve interest interventions (Fig. 1) were included.
Description of the studies
The complete description of the studies is included in Table 1.
Of the twelve interventions selected, five were with healthy
individuals (41·6%)
(35,37,39,40)
, four with diabetics (33·3%)
(34,36)
and three with overweight/obese individuals (25 %)
(38–40)
.
A total of 318 participants were included in this meta-analysis,
of whom 148 were men (46·6 %) and 170 were women (53·4 %),
with a mean age of 31 years (variation, 17–64 years).
Among the interventions selected, most were carried out in a
cross-over (83·3 %) and randomised (75 %) design, and exclu-
sively in an environment external to the laboratory (41·6 %).
In most interventions, all foods and drinks of the diet were
supplied (75 %). The mean period of intervention was 28 d
(variation, 8–70 d).
The amount of fructose provided in the studies had a mean of
92·6 g (variation, 50–182 g) or 20 % of the total energy of the diet
(variation, 10–25 %), mainly provided in a mixed form (solid
and liquid) (58·3 %). Studies containing fructose in the com-
parison component’s composition, such as sucrose and maize
syrup rich in fructose, were not selected. Thus, 50 % of the
interventions used glucose and 50 % used starch as comparison
components. The amount of fat at breakfast (the main meal
analysed) had a variation of 30–40 % (or 10·6–34·7 g) of the total
energy of the meal.
The participants’diets were mainly composed of 55 %
carbohydrates, 15 % proteins and 30 % fats (83·3 %). The energy
balance was estimated as neutral (eight interventions) and
positive (two interventions). Two interventions had distinct
periods (positive and neutral), but were not separated because
they have equal amounts of energy and macronutrients
between the intervention and the comparator.
The most common analysis period for the referred variable,
postprandial TAG, was 4 h (41·6 %). This measure was chosen,
primarily, for most studies presenting peak value of this mea-
sure and because it provides a good evaluation of lipaemic
curve and is a simple protocol that can be used for clinical
purpose
(6,27–29)
.
Funding source was extracted and detailed from interven-
tions, wherein 41·6 % were detailed, which could generate
conflict of interest and alter the outcome of the study
(41,42)
. Data
were extracted as agency or agency/industry funding.
Risk of bias
Among the studies included, 75 % showed proper randomisa-
tion, 16·6 % reported allocation concealment, 33·3 % had blin-
ded the participants and investigators, 16·6 % had blinded the
assessors to the outcomes and 33·3 % reported description of
sample losses (Fig. 2 and 3).
Postprandial TAG
Compared with other carbohydrates (starch or glucose), fruc-
tose generated higher variation (delta) on the concentration of
TAG in the postprandial period, evaluated in 4 h (mean differ-
ence: 8·02 (95 % CI 0·46, 15·58) mg/dl (0·09 (95 % CI 0·01, 0·18)
mmol/l); I
2
: 74 %). The high heterogeneity found (P<0·0001)
was almost exclusively generated by one study
(39)
. Its exclusion
did not alter the total result of the meta-analysis (Fig. 4). There
was no difference in publication bias analysis (Egger’s regres-
sion, P=0·128; online Supplementary Fig. S1).
3337 articles identified through
database search
2805 articles after the removal of
duplicates
2797 exclusions based on the title and/or
summary review
1699 did not study the target population
736 were not clinical trials
299 did not perform the intervention
39 did not evaluate the interest variables
20 did not have the adequate comparator
3 irretrievable reports
1 without translation
8 full-text articles assessed
for eligibility
8 articles included 4 studies with two different
populations
Identification
Selection
Eligibility
Inclusion
12 intervention included in the meta-analysis
Fig. 1. Flow chart of studies included.
Fructose and postprandial TAG 3
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
Tab le 1 . Characteristics of included studies
(Mean values and standard deviations)
Age (years)
Control of
Studies Subjects Mean SD Place of intervention Study design consumption* Randomisation
Obese/overweight
Stanhope et al.
(18)
†32 (16M, 16W) 53 (SD 10) (M)
54 (SD 5·6) (W)
Ext/Lab Parallel Dietetic (Lab)
Supplementation
(External)
No
Swarbrick et al.
(38)
7 (0M, 7W) 64 7·9 Lab Crossed Dietetic No
Heden et al.
(40)
‡20 (11M, 9W) 17 (SD 0·5) (M)
17 (SD 0·6) (W)
Ext Crossed Dietetic Yes
Healthy
Bantle et al.
(35)
‡12W 29 (SD 7·3) (<40 years)
51 (SD 4·9) (>40 years)
Ext Crossed Dietetic Yes
Bantle et al.
(35)
‡12H 31 (SD 7·3) (<40 years)
54 (SD 9·8) (>40 years)
Ext Crossed Dietetic Yes
Stanhope et al.
(39)
†32 (18M, 14W) 27 7·0 Ext/Lab Parallel Dietetic (Lab)
Supplementation
(External)
No
Swanson et al.
(37)
14 (7H, 7M) 34 19–60 Ext Crossed Dietetic Yes
Heden et al.
(40)
‡20 (9M, 11W) 18 (SD 0·6) (M)
18 (SD 0·4) (W)
Ext Crossed Supplementation Yes
Diabetics
Bantle et al.
(34)
12 DM1 (6H, 6M) 23 15–32 Lab Crossed Dietetic Yes
Bantle et al.
(34)
12 DM2 (5H, 7M) 62 36–80 Lab Crossed Dietetic Yes
Bantle et al.
(36)
6 DM1 (3M, 3W) 23 18–34 Ext/Lab Crossed Dietetic Yes
Bantle et al.
(36)
12 DM2 (4M, 8W) 62 40–72 Ext/Lab Crossed Dietetic Yes
Dose of fructose§
Form of Comparison Intervention Energetic Analysis
g % consumption|| component¶ period (d) Diet** balance†† period (h) Financial support
Obese/overweight
Stanhope et al.
(18)
~182 ≥25 Liquid Glucose 70 55:15:30 Neutral/positive 24 Agency
Swarbrick et al.
(38)
~125 25 Liquid Starch 70 55:15:30 Neutral 14 Agency
Heden et al.
(40)
50 10 Liquid Glucose 15 50:16:34 Positive 12 Agency
Healthy
Bantle et al.
(35)
70 14 Mixed Glucose 42 55:15:30 Neutral 24 Agency
Bantle et al.
(35)
70 14 Mixed Glucose 42 55:15:30 Neutral 24 Agency
Stanhope et al.
(39)
~145 25 Liquid Glucose 15 55:15:30 Neutral/positive 24 Agency
Swanson et al.
(37)
100 20 Mixed Starch 28 55:15:30 Neutral 4 Agency/industr y
Heden et al.
(40)
50 10 Liquid Glucose 15 50:16:34 Positive 12 Agency
Diabetics
Bantle et al.
(35)
85·25 21 Mixed Starch 8 55:15:30 Neutral 4 Agency/industry
Bantle et al.
(35)
85·25 21 Mixed Starch 8 55:15:30 Neutral 4 Agency/industry
Bantle et al.
(36)
100 20 Mixed Starch 28 55:15:30 Neutral 4 Agency/industr y
Bantle et al.
(36)
100 20 Mixed Starch 28 55:15:30 Neutral 4 Agency/industr y
M, men; W, women; Ext, external to the laboratory; Lab, laboratory; DM1, type 1 diabetes; DM2, type 2 diabetes.
* Consumption control: dietetic, when all foods, beverages and supplements were provided. Supplementation, when only the intervention carbohydrate was provided by the investigator.
†Two studies presented periods in laboratory (neutral energy balance) and external environment (positive energy balance); however, there was no energy intake difference between the protocols (fructose v. glucose) of the study (isoenergetic).
‡Two studies characterised participants according to age, amount of fat or BMI. The outcome was analysed with all included, but the group or subgroup’s total mean age was not disclosed.
§ The amount of carbohydrates administered in g/d and total diet energy percentage (%). When preceded by ‘~’, it represents the mean amount estimated reported indirectly by the study. In cases where data were unavailable, the value was calculated from 25% of the
total of a diet of 8368kJ (2000 kcal).
|| Fructose could be supplied in liquid form (in the form of sweetened drink) or mixed (from solid foods and sweetened drinks).
¶ Comparison component refers to another carbohydrate (control) provided with the intervention (fructose), regardless of whether or not it is hyperenergetic.
** Macronutrients’energy values for carbohydrates: proteins: fats informed in the study.
†† Represents the ratio between participants’energy consumption and output. Positive when there was energy surplus. Neutral when both were considered equivalent.
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
The analysis of subgroups showed higher variation of
TAG concentration for overweight/obese individuals (mean
difference: 11·47 (95 % CI 4·51, 18·44) mg/dl (0·13 (95 % CI
0·05, 0·21) mmol/l); I
2
: 0 %) and healthy individuals (mean
difference: 13·55 (95 % CI 1·60, 25·49) mg/dl (0·15 (95 % CI
0·02, 0·29) mmol/l); I
2
: 76%) but not in diabetics (mean differ-
ence: −2·77 (95 % CI −10·54, 4·99) mg/dl (−0·03 (95 % CI −0·12,
0·06) mmol/l); I
2
: 0 %). High heterogeneity in healthy indivi-
duals (P=0·002) was exclusively generated by one study
(39)
. Its
withdrawal did not alter the result of the subgroup.
Owing to the high heterogeneity, sensitivity analyses were
carried out in the interventions for the following: (a) funding,
industry/agency (mean difference: −2·63 (95 % CI −9·86, 4·60)
mg/dl (−0·03 (95 % CI −0·11, 0·05) mmol/l); P=0·48; I
2
:0%)
and agency (mean difference: 15·20 (95 % CI 7·40, 23·00) mg/dl
(0·17 (95 % CI 0·08, 0·26) mmol/l); P=0·0001; I
2
: 65 %) (online
Supplementary Fig. S2); (b) amount of carbohydrate, <87 g
(mean difference: 6·97 (95 % CI 2·14, 11·80) mg/dl (0·08 (95 %
CI 0·02, 0·13) mmol/l); P=0·005; I
2
: 0 %) and >87 g (mean
difference: 7·55 (95 % CI −6·40, 21·49) mg/dl (0·09 (95 % CI
−0·07, 0·24) mmol/l); P=0·29; I
2
: 81 %) (subdivision by
amounts larger or smaller than 87 g was based in the
p95 fructose consumption by the American population)
(33)
(online Supplementary Fig. S3); (c) randomisation, yes (mean
difference: 3·93 (95 % CI −1·42, 9·29) mg/dl (0·04 (95 % CI
−0·02, 0·10) mmol/l); P=0·26; I
2
: 21 %) and no (mean differ-
ence: 22·8(95%CI17·96, 27·64) mg/dl (0·26 (95 % CI 0·20, 0·31)
mmol/l); P<0·00001; I
2
: 0 %) (online Supplementary Fig. S4); (d)
energy balance, positive (mean difference: 14·78 (95 % CI 3·94,
25·62) mg/dl (0·17 (95 % CI 0·04, 0·29) mmol/l); P=0·008;
I
2
:82%)andneutral(meandifference:3·11 (95 % CI −4·80, 11·01)
mg/dl (0·04 (95 % CI −0·05, 0·12) mmol/l); P=0·44; I
2
:36%)
(online Supplementary Fig. S5); (e) form of fructose, liquid
(mean difference: 14·78 (95 % CI 5·94, 23·62) mg/dl (0·17 (95 %
CI 0·07, 0·27) mmol/l); P=0·001; I
2
: 76 %) and mixed (mean
difference: −0·42 (95 % CI −7·58, 6·75) mg/dl (0·00 (95 % CI
−0·09, 0·08) mmol/l); P=0·91; I
2
: 6 %) (online Supplementary
Fig. S6); (f) type of comparison (comparator), starch (mean dif-
ference: 1·19 (95 % CI −7·73, 9·51) mg/dl (0·01 (95 % CI −0·09,
0·11) mmol/l); P=0·78; I
2
: 40 %) and glucose (mean difference:
15·33 (95 % CI 6·02, 24·64) mg/dl (0·17 (95 % CI 0·07, 0·28) mmol/l);
P=0·001; I
2
: 70 %) (online Supplementary Fig. S7); (g) follow-
up, <30 d (mean difference: 4·92 (95 % CI −4·41, 14·25) mg/dl
(0·06 (95 % CI −0·05, 0·16) mmol/l); P=0·30; I
2
: 83 % and >30 d
(mean difference: 16·85 (95 % CI 6·35, 27·34) mg/dl (0·19 (95 %
CI 0·07, 0·31) mmol/l); P=0·002; I
2
: 0 %) (online Supplementary
Fig. S8); (h) length of analysis, 4 h (mean difference: −2·63 (95 %
CI −9·86, 4·60) mg/dl (−0·03 (95 % CI −0·11, 0·05) mmol/l);
P=0·48; I
2
: 0 %) and >12 h (mean difference: 15·20 (95 % CI
7·40, 23·00) mg/dl (0·17 (95 % CI 0·08, 0·26) mmol/l);
P=0·0001; I
2
: 65 %) (online Supplementary Fig. S9).
Apparently, authors or studies that received funding support
from the industry had a tendency to show no increase in
TAG concentration after fructose consumption. Non-randomised
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participants and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
0 % 25 % 50 % 75 % 100 %
Fig. 2. Risk of bias of the studies included (percentage). , Low risk of bias; , unclear risk of bias; , high risk of bias.
Bantle et al., 1986a DM1 +–
+–
+–
+–
+–
+–
+++–
+–
+
++++
++++
++
+
+–
––––
Random sequence generation (selection bias)
Allocation concealment (selection bias)
Blinding of participatns and personnel (performance bias)
Blinding of outcome assessment (detection bias)
Incomplete outcome data (attrition bias)
Bantle et al., 1992a DM1
Bantle et al., 2000a Females
Heden et al., 2014a Obese
Heden et al., 2014b Lean
Stanhope et al., 2009
Stanhope et al., 2011
Swanson et al., 1992
Swarbrick et al., 2008
Bantle et al., 1986b DM2
Bantle et al., 1992b DM2
Bantle et al., 2000b Males
Fig. 3. Summary of risk of bias of the studies included. DM1, type 1 diabetes;
DM2, type 2 diabetes.
Fructose and postprandial TAG 5
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
interventions, positive energy balance, liquid fructose, glucose
comparison component, total period of intervention >30 d (follow-
up) and length of analysis >12 h influenced the results of the study.
Discussion
The main finding of this systematic review with meta-analysis is
that chronic fructose consumption (>7 d) causes higher variation
of TAG concentration in the postprandial period when compared
with other carbohydrates (glucose or starch). This variation (delta)
from fasting (immediately before breakfast) to 4-h peak TAG was
about 8·02 mg/dl (0·09 mmol/l). This effect occurred in healthy
and overweight/obese individuals, but not in diabetics.
Fructose is absorbed in the final portion of the duodenum
and ileum, in the small intestine, from non-dependent Na pro-
cess. From portal circulation, the monosaccharide is transported
to the liver, where it can be converted to glucose, lactate, gly-
cogen, glycerol and fatty acids
(43–45)
, regardless of insulin
secretion
(39)
. In healthy individuals, fructose is oxidised at
approximately 45 %, within a period from 3 to 6 h after inges-
tion, including the entry of carbons in the lipogenesis path-
way
(43)
. The postprandial hyperlipidic effect of fructose seems
to originate directly from the synthesis of fatty acids and gly-
cerol in hepatocytes
(14,20)
, and indirectly by the smaller with-
drawal of TAG from the plasma by reduction in the activation of
lipoprotein lipase from the adipose tissue
(20)
.
Different studies showed that fructose-rich diets induce
alterations in the lipid metabolism in eutrophic and overweight/
obese individuals
(15,18,39,46)
. The addition of fructose (1·0–3·0g/kg
per d) to the diet increases fasting
(15,47,48)
and postprandial
(16,18,39)
lipaemia. Moreover, it is associated with the reduction of liver
sensitivity to insulin
(49)
, even at moderate doses (40 g/d)
(19)
,and
decrease of fat oxidation
(50)
.
A series of sensitivity analyses was carried out in order
to verify heterogeneity on the results found. One analysis
accentuated the relation between industry funding and null
results of fructose on postprandial TAG. Clearly, the conflict of
interests can interfere in the conclusions of a particular
study
(42)
. Currently, the influence of funding by the food
industry on several health outcomes has been discussed,
as well as how it could modify guidelines of nutrition
(41,51,52)
.
Some systematic reviews with meta-analysis have been
published on the effects of fructose on blood lipids
(17,21,53)
,
glycaemic control on diabetes
(54)
, blood pressure
(55)
, markers
of non-alcoholic liver steatosis
(56)
, weight gain
(57)
, uric acid
(58)
and postprandial TAG
(24)
. These studies also compared fructose
with fructose-containing carbohydrates, including sucrose and
high-fructose maize syrup (HFCS), which are more likely to
show a small or no effect on different metabolic outcomes
(59)
.
Despite the fact that all these studies received funding support
from the food industry (which produces or uses fructose
in products), results remain feasible but must be interpreted
with caution.
Study or subgroup
1.1.1 Obese/overweight
1.1.2 Healthy
1.1.3 Diabetics
Heden et al., 2014a Obese
Bantle et al., 2000a Women
Bantle et al., 1992b DM2
Bantle et al., 1992a DM1
Bantle et al., 1986b DM2
Bantle et al., 1986a DM1
Bantle et al., 2000b Men
Heden et al., 2014b Eutrophic
Heterogeneity: 2=0
.00; 2=1
.27, df = 2 (P=0
.53); I2=0%
Heterogeneity: 2=108.52; 2=16
.53, df = 4 (P=0
.002); I2=76%
Heterogeneity: 2=0
.00; 2=2
.71, df = 3 (P=0
.44); I2=0%
Heterogeneity: 2=106.46; 2=41
.61, df = 11 (P<0
.0001); I2=74%
Test for overall effect: Z= 3.23 (P=0
.001)
Test for overall effect: Z= 2.22 (P=0
.03)
Test for overall effect: Z= 0.70 (P=0
.48)
Test for overall effect: Z= 2.08 (P=0
.04)
Test for subgroup differences: 2=8
.77, df = 2 (P=0
.01), I2=77
.2%
Stanhope et al., 2009
Stanhope et al., 2011
Swanson et al., 1992
Subtotal (95 % CI)
Subtotal (95 % CI)
Subtotal (95 % CI)
Total (95 % CI)
Swarbrick et al., 2008
31.5
37.1
49
74
23.9
39.0
7.3
14.5
7.3
14.5
12
12
6
12
52
65
29.2
50.4
20.2
27.6
20.2
27.6
12
12
6
12
42
10.3
7.9
8.1
7.9
34.242
75.2
16.5
46.6
31.0
41.9
41.9
11.6
8.4
28.5
24.7
49.6
9.0
22.5
32.7
36.4
36.4
12.9
6.8
25.1
12
12
20
16
14
12
12
20
16
14
74 74
4.1
4.1
12.2
13.1
7.1
40.6
61.4
59.3
9.9
57.1
13.3
22.5
34.9
44.3
16.9
48.0
10.6
20
15
7
11.89
.00 0.42, 17.58
26.50 –9.93, 62.93
15.00 2.40, 27.60
11.47 4.51, 18.44
12.40 –19.00, 43.80
25.60 –5.80, 57.00
7.50 –0.10, 15.10
24.10 18.80, 29.40
–1.70 –21.59, 18.19
13.55 1.60, 25.49
–3.00 –15.15, 9.15
9.00 –8.64, 26.64
–5.30 –22.49, 11.89
–11.40 –29.04, 6.24
–2.77 –10.54, 4.99
158 100.0160 8.02 0.46, 15.58
–100 –50 0
Fructose Any CHO
50 100
3.3
10.1
20
44 42 25.2
17
7
Fructose Any CHO Mean difference
Mean SD Total
Weight
(%) IV, random, 95 % CI
Mean difference
IV, random, 95 % CIMean TotalSD
Fig. 4. Forest plot of the effect of fructose or any carbohydrate (CHO) consumption on postprandial TAG in obese/overweight, healthy and diabetic individuals. The
estimation for each group (subtotal) and combined effect (total) are detailed. The data are in mean difference and 95 % CI (mg/dl) of the δbetween fasting and 4-h peak
TAG. Significance values for the random-effect model. Values in mg/dl can be transformed to mmol/l by multiplying by 0·0113. DM, diabetes mellitus.
6 R. C. O. Macedo et al.
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
The median amount of fructose among the studies selected
was approximately 93 g, which represents approximately
86 % higher than AHA recommendations, but only approxi-
mately 7 % above p95 of the consumption by the American
population
(33)
. This meta-analysis found that quantities smaller
than 87 g of fructose/d are enough to promote variation
in postprandial TAG, as demonstrated in a previous
meta-analysis that observed the threshold of 50 g/d for
the general population
(17)
and 60 g/d for type 2 diabetes
(53)
.
These thresholds are very close to the average consumption (49 g)
by the American
(33)
and Dutch
(60)
population.However,some
authors stated that this effect only occurs when there is positive
energy balance or when fructose is generating a hyperenergetic
conditionincomparisonwithanothercarbohydrate
(21,24)
.This
effect was equally found in our study from sensitivity analysis.
The recent meta-analysis by Evans et al.
(23)
found no differ-
ence in postprandial TAG with acute fructose consumption
when compared with glucose or sucrose. Furthermore, the
authors also propose replacing those sugars for fructose, once
they did not find TAG alterations, and glycaemia and insuli-
naemia reduction were found in the postprandial period. Our
view is contrary, as our sensitivity analysis showed that a total
time of intervention (follow-up) extended for over 30 d influ-
enced TAG variation. Although increasingly studied, the effects
of chronic fructose consumption are still discordant and are not
fully elucidated in the literature because of a number of con-
founding factors (e.g. fructose dose, excess of energy in the
diet, diabetes, obesity)
(61)
.
Some limitations are present in our study. (1) Data showed
high heterogeneity; sensitivity analysis demonstrated that this
effect was being almost exclusively generated by one study
(39)
but its removal does not alter the result of subgroups and total
results of the meta-analysis. (2) The inclusion of a study with
adolescents
(40)
could create a confusion factor, but it was
considered important for representing the metabolic interven-
tion effect on different age groups. (3) One study
(38)
did not
clearly show the amount of fructose provided (in grams or % of
the diet) and, in this case, had to be removed. (4) The amount
of fructose varied to a great extent among studies (50–182 g),
showing lack of intervention standardisation and often the
provision of supraphysiological doses
(62)
. (5) All interventions
that evaluated diabetics analysed plasma TAG over a period of
4 h, which may represent a bias; studies with longer analysis
periods especially in diabetic subjects are needed. (6) The
quality of studies varied between groups, presenting higher risk
of bias in diabetes interventions.
Findings from this systematic review with meta-analysis
update the results previously described by Wang et al.
(24)
and
showed the negative effects of chronic (>7 d) fructose con-
sumption on postprandial TAG, in healthy adolescents and
adults, as well as in overweight/obese individuals, but not in
diabetics. As chronic ingestion of fructose may promote lipae-
mic alterations, and hypertriacylglycerolaemia in the post-
prandial period is associated with increased morbimortality,
recommendations for the population are needed to limit
intake, especially from liquids (e.g. sweetened beverages).
Longitudinal studies (>30 d), well-controlled, with habitual
doses of consumption between 49 and 87 g (close to the p95 of
the population) and in different forms (free fructose, HFCS,
sucrose) are necessary to clarify the interrelationship between
fructose, lipaemia and CVD.
Acknowledgements
The authors thank authors Timothy Daniel Heden, Jill Kanaley
and John Sievenpiper for responding our emails and Josianne
Krause and Daniel Umpierre for aiding with data analysis.
C. E. J. M. has received PhD scholarship from the Coorde-
nação de Aperfeiçoamento de Pessoal de Nível Superior.
R. C. O. M. formulated the research question; A. F. V. and
C.E.J.M.readandextracteddata.A.F.V.,C.E.J.M.,R.C.O.M.
analysed the data. R. C.O.M.,A.F.V.,C.E.J.M.andA.R.-O.
wrote, reviewed, performed and perfected this study.
The authors declare that there are no conflicts of interest to
declare.
Supplementary material
For supplementary material/s referred to in this article, please
visit https://doi.org/10.1017/S0007114518001538
References
1. Ooi TC & Nordestgaard BG (2011) Methods to study post-
prandial lipemia. Current Vasc Pharmacol 9, 302–308.
2. Moreton JR (1947) Atherosclerosis and alimentary hyperlipemia.
Science 106, 190–191.
3. Stefanutti C, Labbadia G & Athyros VG (2014)
Hypertriglyceridaemia, postprandial lipaemia and non-HDL
cholesterol. Curr Pharm Des 20, 6238–6248.
4. Lairon D & Defoort C (2011) Effects of nutrients on post-
prandial lipemia. Current Vasc Pharmacol 9, 309–312.
5. Lairon D, Lopez-Miranda J & Williams C (2007) Methodology
for studying postprandial lipid metabolism. Eur J Clin Nutr 61,
1145–1161.
6. Kolovou GD, Mikhailidis DP, Kovar J, et al. (2011) Assessment
and clinical relevance of non-fasting and postprandial trigly-
cerides: an expert panel statement. Current Vasc Pharmacol
9, 258–270.
7. Mora S, Rifai N, Buring JE, et al. (2008) Fasting compared with
nonfasting lipids and apolipoproteins for predicting incident
cardiovascular events. Circulation 118, 993–1001.
8. Miller M, Stone NJ, Ballantyne C, et al. (2011) Triglycerides
and cardiovascular disease: a scientific statement from the
American Heart Association. Circulation 123, 2292–2333.
9. Nordestgaard BG & Varbo A (2014) Triglycerides and
cardiovascular disease. Lancet 384, 626–635.
10. Bansal S, Buring JE, Rifai N, et al. (2007) Fasting compared
with nonfasting triglycerides and risk of cardiovascular events
in women. JAMA 298, 309–316.
11. Bae JH, Bassenge E, Kim KB, et al. (2001) Postprandial
hypertriglyceridemia impairs endothelial function by
enhanced oxidant stress. Atherosclerosis 155, 517–523.
12. Cohen JC, Noakes TD & Benade AJ (1989) Postprandial
lipemia and chylomicron clearance in athletes and in
sedentary men. Am J Clin Nutr 49, 443–447.
13. Chong MF, Fielding BA & Frayn KN (2007) Metabolic
interaction of dietary sugars and plasma lipids with a focus
on mechanisms and de novo lipogenesis. Proc Nutr Soc
66,52–59.
Fructose and postprandial TAG 7
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
14. Parks EJ, Skokan LE, Timlin MT, et al. (2008) Dietary sugars
stimulate fatty acid synthesis in adults. J Nutr 138, 1039–1046.
15. Abdel-Sayed A, Binnert C, Le KA, et al. (2008) A high-fructose
diet impairs basal and stress-mediated lipid metabolism in
healthy male subjects. Br J Nutr 100, 393–399.
16. Bidwell AJ, Fairchild TJ, Redmond J, et al. (2014) Physical
activity offsets the negative effects of a high-fructose diet. Med
Sci Sports Exerc 46, 2091–2098.
17. Livesey G & Taylor R (2008) Fructose consumption and
consequences for glycation, plasma triacylglycerol, and
body weight: meta-analyses and meta-regression models of
intervention studies. Am J Clin Nutr 88, 1419–1437.
18. Stanhope KL, Schwarz JM, Keim NL, et al. (2009) Consuming
fructose-sweetened, not glucose-sweetened, beverages
increases visceral adiposity and lipids and decreases insulin
sensitivity in overweight/obese humans. J Clin Invest 119,
1322–1334.
19. Aeberli I, Hochuli M, Gerber PA, et al. (2013) Moderate
amounts of fructose consumption impair insulin sensitivity in
healthy young men: a randomized controlled trial. Diabetes
Care 36, 150–156.
20. Chong MF, Fielding BA & Frayn KN (2007) Mechanisms for
the acute effect of fructose on postprandial lipemia. Am J Clin
Nutr 85, 1511–1520.
21. Chiavaroli L, de Souza RJ, Ha V, et al. (2015) Effect of fructose
on established lipid targets: a systematic review and
meta-analysis of controlled feeding trials. J Am Heart Assoc 4,
e001700.
22. Evans RA, Frese M, Romero J, et al. (2017) Chronic fructose
substitution for glucose or sucrose in food or beverages
has little effect on fasting blood glucose, insulin, or triglycer-
ides: a systematic review and meta-analysis. Am J Clin Nutr
106, 519–529.
23. Evans RA, Frese M, Romero J, et al. (2017) Fructose
replacement of glucose or sucrose in food or beverages
lowers postprandial glucose and insulin without raising
triglycerides: a systematic review and meta-analysis. Am J Clin
Nutr 106, 506–518.
24. David Wang D, Sievenpiper JL, de Souza RJ, et al. (2014)
Effect of fructose on postprandial triglycerides: a systematic
review and meta-analysis of controlled feeding trials. Athero-
sclerosis 232, 125–133.
25. Liberati A, Altman DG, Tetzlaff J, et al. (2009) The PRISMA
statement for reporting systematic reviews and meta-analyses
of studies that evaluate health care interventions: explanation
and elaboration. Ann Intern Med 151, W65–W94.
26. Shamseer L, Moher D, Clarke M, et al. (2015) Preferred
reporting items for systematic review and meta-analysis
protocols (PRISMA-P) 2015: elaboration and explanation.
BMJ 349, g7647.
27. Kolovou GD, Mikhailidis DP, Nordestgaard BG, et al. (2011)
Definition of postprandial lipaemia. Current Vasc Pharmacol
9, 292–301.
28. Mihas C, Kolovou GD, Mikhailidis DP, et al. (2011) Diagnostic
value of postprandial triglyceride testing in healthy subjects: a
meta-analysis. Current Vasc Pharmacol 9, 271–280.
29. Weiss EP, Fields DA, Mittendorfer B, et al. (2008) Reprodu-
cibility of postprandial lipemia tests and validity of an abbre-
viated 4-hour test. Metabolism 57, 1479–1485.
30. Liberati A, Altman DG, Tetzlaff J, et al. (2009) The PRISMA
statement for reporting systematic reviews and meta-analyses
of studies that evaluate healthcare interventions: explanation
and elaboration. BMJ 339, b2700.
31. Higgins JPT & Green S (2008) Cochrane Handbook for
Systematic Reviews of Interventions, Cochrane Book Series,
Cochrane Collaboration. Chichester; Hoboken, NJ: Wiley-
Blackwell.
32. Higgins JP, Thompson SG, Deeks JJ, et al. (2003) Measuring
inconsistency in meta-analyses. BMJ 327, 557–560.
33. Marriott BP, Cole N & Lee E (2009) National estimates of
dietary fructose intake increased from 1977 to 2004 in the
United States. J Nutr 139, 1228S–1235S.
34. Bantle JP, Laine DC & Thomas JW (1986) Metabolic effects of
dietary fructose and sucrose in types I and II diabetic subjects.
JAMA 256, 3241–3246.
35. Bantle JP, Raatz SK, Thomas W, et al. (2000) Effects of dietary
fructose on plasma lipids in healthy subjects. Am J Clin Nutr
72, 1128–1134.
36. Bantle JP, Swanson JE, Thomas W, et al. (1992) Metabolic
effects of dietary fructose in diabetic subjects. Diabetes Care
15, 1468–1476.
37. Swanson JE, Laine DC, Thomas W, et al. (1992) Metabolic
effects of dietary fructose in healthy subjects. Am J Clin Nutr
55, 851–856.
38. Swarbrick MM, Stanhope KL, Elliott SS, et al. (2008) Con-
sumption of fructose-sweetened beverages for 10 weeks
increases postprandial triacylglycerol and apolipoprotein-B
concentrations in overweight and obese women. Br J Nutr
100, 947–952.
39. Stanhope KL, Bremer AA, Medici V, et al. (2011) Consumption
of fructose and high fructose corn syrup increase postprandial
triglycerides, LDL-cholesterol, and apolipoprotein-B in
young men and women. J Clin Endocrinol Metab 96,
E1596–E1605.
40. Heden TD, Liu Y, Park YM, et al. (2014) Moderate amounts of
fructose- or glucose-sweetened beverages do not differentially
alter metabolic health in male and female adolescents. Am J
Clin Nutr 100, 796–805.
41. Mozaffarian D (2017) Conflict of interest and the role of the
food industry in nutrition research. JAMA 317, 1755–1756.
42. Lundh A, Lexchin J, Mintzes B, et al. (2017) Industry
sponsorship and research outcome. Cochrane Database
Syst Rev, issue 2, MR000033.
43. Sun SZ & Empie MW (2012) Fructose metabolism in humans –
what isotopic tracer studies tell us. Nutr Metab (Lond) 9, 89.
44. Elliott SS, Keim NL, Stern JS, et al. (2002) Fructose, weight
gain, and the insulin resistance syndrome. Am J Clin Nutr 76,
911–922.
45. Bray GA, Nielsen SJ & Popkin BM (2004) Consumption of
high-fructose corn syrup in beverages may play a role in the
epidemic of obesity. Am J Clin Nutr 79, 537–543.
46. Faeh D, Minehira K, Schwarz JM, et al. (2005) Effect of
fructose overfeeding and fish oil administration on hepatic de
novo lipogenesis and insulin sensitivity in healthy men.
Diabetes 54, 1907–1913.
47. Egli L, Lecoultre V, Theytaz F, et al. (2013) Exercise prevents
fructose-induced hypertriglyceridemia in healthy young
subjects. Diabetes 62, 2259–2265.
48. Le KA, Faeh D, Stettler R, et al. (2006) A 4-wk high-fructose
diet alters lipid metabolism without affecting insulin sensitivity
or ectopic lipids in healthy humans. Am J Clin Nutr 84,
1374–1379.
49. Ter Horst KW, Schene MR, Holman R, et al. (2016) Effect of
fructose consumption on insulin sensitivity in nondiabetic
subjects: a systematic review and meta-analysis of diet-
intervention trials. Am J Clin Nut 104, 1562–1576.
50. Cox CL, Stanhope KL, Schwarz JM, et al. (2012) Consumption
of fructose-sweetened beverages for 10 weeks reduces net fat
oxidation and energy expenditure in overweight/obese men
and women. Eur J Clin Nutr 66, 201–208.
8 R. C. O. Macedo et al.
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
51. Bes-Rastrollo M, Schulze MB, Ruiz-Canela M, et al. (2013)
Financial conflicts of interest and reporting bias regarding the
association between sugar-sweetened beverages and weight
gain: a systematic review of systematic reviews. PLoS Med 10,
e1001578; discussion e1001578.
52. Chartres N, Fabbri A & Bero LA (2016) Association of industry
sponsorship with outcomes of nutrition studies: a systematic
review and meta-analysis. JAMA Intern Med 176, 1769–1777.
53. Sievenpiper JL, Carleton AJ, Chatha S, et al. (2009) Hetero-
geneous effects of fructose on blood lipids in individuals with
type 2 diabetes: systematic review and meta-analysis of
experimental trials in humans. Diabetes Care 32, 1930–1937.
54. Cozma AI, Sievenpiper JL, de Souza RJ, et al. (2012) Effect of
fructose on glycemic control in diabetes: a systematic review
and meta-analysis of controlled feeding trials. Diabetes Care
35, 1611–1620.
55. Ha V, Sievenpiper JL, de Souza RJ, et al. (2012) Effect of
fructose on blood pressure: a systematic review and
meta-analysis of controlled feeding trials. Hypertension 59,
787–795.
56. Chiu S, Sievenpiper JL, de Souza RJ, et al. (2014) Effect of
fructose on markers of non-alcoholic fatty liver disease
(NAFLD): a systematic review and meta-analysis of controlled
feeding trials. Eur J Clin Nutr 68, 416–423.
57. Sievenpiper JL, de Souza RJ, Mirrahimi A, et al. (2012) Effect of
fructose on body weight in controlled feeding trials: a systematic
review and meta-analysis. AnnInternMed156, 291–304.
58. Wang DD, Sievenpiper JL, de Souza RJ, et al. (2012) The
effects of fructose intake on serum uric acid vary among
controlled dietary trials. J Nutr 142, 916–923.
59. Stanhope KL, Griffen SC, Bair BR, et al. (2008) Twenty-
four-hour endocrine and metabolic profiles following
consumption of high-fructose corn syrup-, sucrose-, fructose-,
and glucose-sweetened beverages with meals. Am J Clin Nutr
87,1194–1203.
60. Sluik D, Engelen AI & Feskens EJ (2015) Fructose consump-
tion in the Netherlands: the Dutch National Food Consump-
tion Survey 2007–2010. Eur J Clin Nutr 69, 475–481.
61. Tappy L & Le KA (2010) Metabolic effects of fructose and the
worldwide increase in obesity. Physiol Rev 90,23–46.
62. Choo VL & Sievenpiper JL (2015) The ecologic validity of
fructose feeding trials: supraphysiological feeding of fructose
in human trials requires careful consideration when drawing
conclusions on cardiometabolic risk. Front Nutr 2, 12.
Fructose and postprandial TAG 9
https://doi.org/10.1017/S0007114518001538
Downloaded from https://www.cambridge.org/core. IP address: 201.21.63.95, on 03 Jul 2018 at 11:55:24, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.
