nutrients

Review
Dietary Strategies for Weight Loss Maintenance
Marlene A. van Baak * and Edwin C. M. Mariman
Department of Human Biology, NUTRIM School for Nutrition and Translation Research in Metabolism,
Maastricht University, 6229 ER Maastricht, The Netherlands
* Correspondence: m.vanbaak@maastrichtuniversity.nl




Received: 25 June 2019; Accepted: 12 August 2019; Published: 15 August 2019


Abstract: Weight regain after a successful weight loss intervention is very common. Most studies
show that, on average, the weight loss attained during a weight loss intervention period is not
or is not fully maintained during follow-up. We review what is currently known about dietary
strategies for weight loss maintenance, focusing on nutrient composition by means of a systematic
review and meta-analysis of studies and discuss other potential strategies that have not been studied
so far. Twenty-one studies with 2875 participants who were overweight or obese are included in
this systematic review and meta-analysis. Studies investigate increased protein intake (12 studies),
lower dietary glycemic index (four studies), green tea (three studies), conjugated linoleic acid (three
studies), higher fibre intake (three studies), and other miscellaneous interventions (six studies).
The meta-analysis shows a significant beneficial effect of higher protein intake on the prevention of
weight regain (SMD (standardized mean difference) −0.17 (95% CI −0.29, −0.05), z = 2.80, p = 0.005),
without evidence for heterogeneity among the included studies. No significant effect of the other
strategies is detected. Diets that combine higher protein intake with different other potentially
beneficial strategies, such as anti-inflammatory or anti-insulinemic diets, may have more robust
effects, but these have not been tested in randomized clinical trials yet.

Keywords: weight regain; obesity; diet composition; protein intake; systematic review and
meta-analysis

1. Introduction
Over the past decades, many studies have been performed to gain insight into the best dietary
strategy to lose weight. It turns out that there is probably not one best strategy and that individual
preferences can be taken into account as long as the individual adheres to the diet and energy restriction
is attained [1,2].
A much less studied topic is what the best dietary strategy is to prevent weight regain after
successful weight loss. Weight regain after a successful weight loss intervention is very common.
Indeed, most studies show that, on average, the weight loss attained during a weight loss intervention
period is not or is not fully maintained during follow-up [3,4]. However, the amount of weight regain
differs among studies, and there are also large inter-individual differences in weight regain within
studies (e.g., see [5]). Although it is often stated that weight regain is inevitable, still, a considerable
number of individuals are able to maintain some weight loss over longer periods of time. For instance,
Coughlin et al. (2016) [6] reported that after an initial weight loss of 8 kg over a 6-month period,
participants of the Weight Loss Maintenance trial maintained a 4 kg weight loss after 2 years of
follow-up and maintained a 3 kg weight loss after 5 years. Pekkarinen et al. (2015) [7] found that
35% of the participants of a 4-month weight loss intervention maintained a weight loss ≥ 5% after
2 years follow-up. In the Diabetes Prevention Outcome Study, it was found that 40% of the participants
who had lost ≥ 5% of their initial body weight after 1 year (in the intensive lifestyle group) had

Nutrients 2019, 11, 1916; doi:10.3390/nu11081916 www.mdpi.com/journal/nutrients

Nutrients 2019, 11, 1916 2 of 12

maintained at least 5% weight loss after 15 years of follow-up [8]. On average, no weight regain during
a follow-up period of 12 months was reported after a weight loss program of 3 months in a Japanese
worksite intervention [9]. In the Look Ahead Study, comparing an intensive lifestyle intervention
(ILI) with regular diabetes control, weight regain occurred after 1 year, but at 4 years weight loss was
still significantly more pronounced in the ILI group, as was the case after 8 years (ILI −4.7%, control
−2.1%) [10,11]. Thus, although weight regain is a common phenomenon, a considerable proportion of
the participants in structured weight loss interventions are able to maintain a certain amount of weight
loss over a prolonged period of time. Potential mechanisms that explain the tendency for weight loss
regain and the inter-individual differences are currently being studied in the hope to develop improved
(and more personalized) strategies for the prevention of weight regain.
Here, we will review what is currently known about dietary strategies for weight loss maintenance,
focusing on nutrient composition, by means of a systematic review and meta-analysis of studies.
Furthermore, we will discuss other dietary strategies that have not been studied yet, but that may be
worth investigating based on underlying mechanisms.

2. Systematic Review and Meta-Analysis of Dietary Strategies for Weight Loss Maintenance

2.1. Methods

2.1.1. Search Strategy and In-/Exclusion Criteria

We conducted a systematic literature search on dietary strategies that have been studied in the
context of weight loss maintenance in adults who were overweight or obese. Search terms in PubMed
were ((weight regain) or (weight maintenance) or (weight loss maintenance)) and diet, with filters:
clinical trial, humans. The search yielded 1062 articles. Based on title and abstract and the in- and
exclusion criteria below, 46 studies were selected for further scrutiny of the text. Studies were included
if participants were adults, were overweight or obese at baseline, and had been randomized to different
diets after following the same energy-restricted diet to attain weight loss, and if the randomized diets
were ad libitum and the study design was a randomized controlled trial (RCT). Outcome was the
weight change over the experimental period. Seven studies were excluded because they were not ad
libitum; six were no RCT, eight had no weight loss phase before randomization, two reported no or
incomplete outcomes, one study had no appropriate control, and one study was a review. Therefore,
21 studies were included in this systematic review and meta-analysis.

2.1.2. Data Extraction

Data extracted from each study included the following items: first author, year of publication, initial
BMI, initial weight loss, sample size, type of intervention, study duration, pre- and post-intervention
body weight, or weight change plus measure of variation standard deviation (SD), standard error of
the mean (SE), or confidence interval (CI) in the experimental and the control group.

2.1.3. Quality Assessment

Quality of the selected studies and risk of bias was assessed with the risk of bias checklist (RoB 2.0)
from the Cochrane Collaboration using Review Manager 5.3 software (The Nordic Cochrane Centre,
The Cochrane Collaboration, Copenhagen, Denmark, 2014).

2.1.4. Data Synthesis and Statistical Analysis

To calculate the effect size of each study, we used the mean change and SD of body weight over
the experimental period in the control and intervention groups. If these values were not reported, we
calculated the mean difference as the difference in mean body weight pre and post treatment and its
SD using the formula: SD = square root ((SDpretreatment )2 + (SDposttreatment )2 ) − (2r × SDpretreatment ×
Nutrients 2019, 11, 1916 3 of 12

SDposttreatment ). Because the pretest–posttest correlation coefficients (r) were not reported in the studies,
an r value of 0.5 was assumed throughout.
The meta-analysis of included studies was conducted using the publicly available Cochrane Review
Manager 5.3 software (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen, Denmark,
2014). If a study included more than two experimental groups, which were compared with one control
group, the number of subjects in the control group was divided by the number of comparisons. The effect
size was expressed as standardized mean difference (SMD), because not all studies reported the outcome
in kg; in some %, weight loss was reported. For comparability, data from the completers analysis were
used for the meta-analysis. Dropout was often high (>20%), and methods used to statistically correct for
dropout were diverse and often arbitrary (e.g., last-observation-carried-forward, return-to-baseline).
Random-effects models were used for the statistical analysis. Heterogeneity was assessed using
the I2 index. The effect size is reported as the standardized mean difference with its 95% confidence
interval (CI). A p-value < 0.05 is considered statistically significant.

2.1.5. Publication Bias

Publication bias was assessed by visual inspection of the funnel plots.

2.2. Results
An overview of the characteristics of the 21 included studies is given in Table 1. Of the 21
included studies, eight studies (with a total of 12 intervention arms) (intervention n = 664, control
n = 504) examined the effect of increasing protein intake on weight loss maintenance [12–19]. The other
studies reported on other dietary strategies (lowering of glycemic index, three studies (intervention
n = 254, control n = 223) [12,17,20]; green tea or its component epigallocatechin-3-gallate (EGCG)
supplementation, three studies (intervention n = 93, control n = 93) [15,21,22]; whole grain enriched diet
or fibre supplementation, two studies (intervention n = 101, control n = 99) [23,24]; conjugated linoleic
acid (CLA) supplementation, two studies (intervention n = 78, control n = 77) [25,26]. The remainder of
the studies reported on miscellaneous interventions: mono-unsaturated, fat-enriched diet, one study with
short or long-term follow-up [27,28]; low-fat diet, one study with short or long-term follow-up [27,28];
acarbose supplementation, one study [29]; capsaicin supplementation, one study [30]; gamma-linoleic
acid (GLA) supplementation, one study [31]; and CHO supplementation without or with a mixture of
chromium picolinate, soluble fibre, and caffeine, one study [32] (intervention n = 274, control n = 177).
Figure 1 shows the Forest plots for the included studies per type of dietary intervention. For the
protein studies, the SMD was −0.17 (95% CI −0.29, −0.05). The test for the overall effect was significant
(p = 0.005), with no evidence for heterogeneity among the included studies (I2 = 0%), suggesting that
an increased protein content of an ad libitum diet has a beneficial effect on weight loss maintenance.
The effect size was modest: the mean difference in weight change across studies was −1.02 kg (95% CI
−1.77, −0.28). However, approximately half of the weight regain was, on average, prevented in these
studies, with durations between 3 and 12 months. Since the study by Aller et al. (2014) [12] was a substudy
(12 month results in part of the study centres) of the Diogenes trial reported in Larsen et al. (2010) [17]
(6 months results in all centres), the outcomes are not completely independent. However, leaving out the
study by Aller et al. [12] does not change the result of the analysis (SMD −0.14 (95% CI −0.27, −0.02) (p
= 0.03). In three of the studies, participants were given dietary advice to increase protein intake by 10
to 15 energy % without specific attention to the protein source (Aller, Debridge, Larsen). In three other
studies, protein intake was increased by supplementation (two studies with casein (30 and 48 g/day)
(Lejeune, Westerterp-Plantenga) and one study with whey or soy (45 g/day) (Kjølbaek). In the remaining
two studies (Claessens, Hursel), a combination of casein supplementation (50 g/day) and dietary advice
was used. Potential mechanisms that have been suggested for the beneficial effect of a higher protein
intake are the satiating properties of protein, their thermogenic effect, and their effect on fat free mass
maintenance/increase [33]. Whether the changes in gut microbiota that are seen with higher protein intake
also play a role or whether they are potentially harmful remains to be established [34].
Nutrients 2019, 11, 1916 4 of 12

Table 1. Characteristics of the randomized clinical trials evaluating the effect of dietary strategies on weight loss maintenance that have been included in the
systematic review.

Author Year Country Participants Intervention Characteristics Main Conclusion Authors

Total Diagnostic Dropout n
Sex Age Diet Duration
Number Criteria (%)
five diet groups: lower
protein/lower glycemic index higher protein content
Netherlands 40% 27–59 BMI ≥ 27
Aller 2014 256 (GI); higher protein/lower GI; 12 months 117 (45%) improves weight loss
and Denmark male year kg/m2
lower protein/higher GI; higher maintenance
protein/higher GI; control
green tea consumption did not
men and 49–65 BMI 30–50 rye bread with green tea extract
Bajerska 2015 Poland 55 12 weeks 0 (0%) improve weight loss
women # year kg/m2 vs. rye bread
maintenance
low-fat, high-casein, or
High-protein plus whey or
high-whey protein diets are
35% 30–60 BMI ≥ 27 casein supplements vs.
Claessens 2009 Netherlands 5 12 weeks 6 (11%) more effective for weight loss
male year kg/m2 high-carbohydrate plus
maintenance than low-fat,
maltodextrin supplements
high-carbohydrate diets
the protein or carbohydrate
50% 18–75 BMI ≥ 27 higher protein vs. lower
Delbridge 2009 Australia 141 12 months 59 (42%) content of the diet had no effect
male year kg/m2 protein diet
on weight loss maintenance
diet composition had no major
42% 18–35 BMI ≥ 28 high-MUFA vs. low-fat vs.
Due 2008 Denmark 131 6 months 25 (19%) effect on preventing weight
male year kg/m2 control diet
regain
42% 18–35 BMI ≥ 28 high-MUFA vs. low-fat vs. weight regain did not differ
Due 2017 Denmark 131 18 months 58 (44%)
male year kg/m2 control diet among the diets
20% 21–66 BMI 32–38 no benefit of acarbose to
Hauner 2001 Germany 110 acarbose vs. placebo capsules 26 weeks 35 (32%)
male year kg/m2 stabilise weight reduction
high-protein diet plus
EGCG/caffeine vs. high-protein
both EGCG/caffeine and higher
men and 18–60 BMI 25–35 diet plus placebo vs.
Hursel 2009 Netherlands 80 13 weeks 0 (0%) protein improved weight
women # year kg/m2 normal-protein diet plus
maintenance independently
EGCG/caffeine vs. normal
-protein diet plus placebo
conjugated linoleic acid (CLA)
43% 20–50 BMI 25–30 CLA did not result in improved
Kamphuis 2003 Netherlands 60 low dose vs. placebo and CLA 13 weeks 6 (10%)
male year kg/m2 weight loss maintenance
high dose vs. placebo
Nutrients 2019, 11, 1916 5 of 12

Table 1. Cont.

Author Year Country Participants Intervention Characteristics Main Conclusion Authors

whey supplement vs. whey
protein supplementation did
men and 18–60 BMI 28–40 plus calcium supplement vs.
Kjølbaek 2017 Denmark 189 24 weeks 38 (20%) not result in improved weight
women # year kg/m2 soy supplement vs. placebo
maintenance
(maltodextrin)
men and 18–60 BMI 25–35 weight maintenance not
Kovacs 2004 Netherlands 104 green tea vs. placebo capsules 13 weeks 0 (0%)
women # year kg/m2 affected by green tea
no effect of whole grain on
20–50 BMI 27–34 whole grain vs. refined grain
Kristensen 2017 France 178 women 12 weeks 9 (5%) weight maintenance, but
year kg/m2 foods
dietary adherence was low
five diet groups: lower
protein/lower GI; higher higher protein content and
Eight European men and 18–65 BMI 27–45
Larsen 2010 773 protein/lower GI; lower 26 weeks 225 (29%) lower GI improve maintenance
countries women # year kg/m2
protein/higher GI; higher of weight loss
protein/higher GI; control
men and 18–65 BMI 28–35 conjugated linoleic acid (CLA) CLA supplementation does not
Larsen 2006 Denmark 101 52 weeks 24 (24%)
women # year kg/m2 or placebo capsules prevent weight regain
men and 18–60 BMI 25–35 capsaicin had no limiting effect
Lejeune 2003 Netherlands 91 capsaicin vs. placebo capsules 12 weeks 0 (0%)
women # year kg/m2 on weight regain
men and 18–60 BMI 25–35 protein supplement vs. control a higher protein intake
Lejeune 2005 Netherlands 103 6 months 0 (0%)
women # year kg/m2 (no placebo) improved weight maintenance
no effect of fibre
41 ± 7 fibre supplement vs. control
Pasman 1997 Netherlands 39 women obese 14 months 8 (20%) supplementation on weight
year (no placebo)
maintenance
CHO plus CHO supplementation
chromium-picolinate plus fibre beneficial for weight
35 ± 7
Pasman 1997 Netherlands 39 women obese plus caffeine supplementvs. 14 months 6 (15%) maintenance, no additional
year
CHO supplement vs. control effect of
(no placebo) chromium-picolinate/fibre/caffeine
18–65 BMI 27–45 changing the diet GI does not
Philippou 2008 UK 42 * ? higher GI vs. lower GI diet 4 months ?
year kg/m2 affect weight maintenance
BMI after
mean gamma-linolenate (GLA) vs. GLA reduced weight regain
Schirmer 2007 USA 50 9% male weight loss 12 months 20 (40%) **
~50 year placebo capsules after major weight loss
<34 kg/m2
men and 44 ± 10 BMI 25–35 protein supplement vs. control higher protein intake resulted
Westerterp-Plantenga
2004 Netherlands 148 13 weeks 0 (0%)
women # year kg/m2 (no placebo) in lower body weight regain
# did not report numbers of men and women separately; * number refers to completers; ** due to early termination.
Nutrients 2019,
Nutrients 11, x
2019, 11, 1916
FOR PEER REVIEW 16of
of 12
12

Figure 1. Forest plots of dietary strategies for the prevention of weight regain after weight loss. The box
Figure 1. Forest plots of dietary strategies for the prevention of weight regain after weight loss. The
size is proportional to the weight contributed by the study to the combined study mean. Horizontal
box size is proportional to the weight contributed by the study to the combined study mean.
lines span individual study 95% confidence intervals (CI). Diamonds represent the combined study
Horizontal lines span individual study 95% confidence intervals (CI). Diamonds represent the
standardized mean value and the corresponding 95% CI values.
combined study standardized mean value and the corresponding 95% CI values.
The meta-analysis of studies on the lowering of the glycemic index of the diet shows an SMD
The (95%
of −0.07 meta-analysis
CI −0.43,of studies
0.28) (p =on the but
0.68), lowering of theheterogeneity
with high glycemic index (I2of
= the dietThis
69%). shows an SMD of
heterogeneity
−0.07 (95%
is clearly dueCI to
−0.43, 0.28) (p
the study by =Aller
0.68),etbut with high
al. (2014) [12]. heterogeneity
When this study (I =is69%).
2 This
left out heterogeneity
of the analysis, theis
clearly due to the study by Aller et al. (2014) [12]. When this study is left out of the
SMD is −0.23 (95% CI −0.45, −0.02) (p = 0.03) and the mean difference is −1.09 kg (95% CI −2.06, −0.13). analysis, the SMD
is
The−0.23 (95%
study by CI −0.45,
Aller −0.02)
et al. had (pthe= 0.03)
longestand the mean(12
follow-up difference
months),iscompared
−1.09 kg (95% CI −2.06,
to 4 and −0.13).
6 months inThe
the
study by Aller
other two et al.
studies had the longest
(Philippou follow-up
et al. 2009 (12 Larsen
[20] and months), et compared
al. 2010 [17], to 4respectively),
and 6 monthsand in the other
changes
two studiesadherence
in dietary (Philippou et al.
over 2009may
time [20]play
and Larsen et al.
a role. As it 2010 [17],
is, the respectively),
picture is not yetandclearchanges
for thein dietary
potential
adherence
effectivenessoverof time
low may play adiets
glycemic role.for
As prevention
it is, the picture is not yet
of weight clearand
regain, for the
morepotential
studieseffectiveness
are needed.
of low glycemic diets for prevention of weight regain, and more studies
The interventions with green tea, conjugated linoleic acid, and fibre so far do not show are needed. The interventions
an effect on
with green tea,
prevention conjugated
of weight regain,linoleic
without acid, and fibre
evidence for so far do not show
heterogeneity. The ansameeffect on prevention
holds of weight
for the miscellaneous
regain,
studies,without
although evidence
these are,forofheterogeneity.
course, hard toThe same holds for the miscellaneous studies, although
compare.
these are, of course, hard to compare.
Nutrients 2019, 11, 1916 7 of 12

Nutrients 2019, 11, x FOR PEER REVIEW 2 of 12

The risk of bias analysis of the individual studies is shown in Figure 2. In general, most studies
The risk of bias
were of acceptable analysis
quality, of the individual
although in many studies
studiesisinformation
shown in Figure 2. Inone
about general,
or moremostaspects
studies was
missing, which resulted in ‘unclear bias’ grading. In diet interventions, not using supplement was
were of acceptable quality, although in many studies information about one or more aspects blinding
missing, which
of participants resultedpersonnel
and study in ‘unclear bias’ grading.
is often In dietand
not possible, interventions,
these studies not were
usingtherefore
supplement graded
blinding of participants and study personnel is often not possible, and these studies were therefore
as ‘unclear bias’ with respect to performance and detection bias. In some of the supplementation
graded as ‘unclear bias’ with respect to performance and detection bias. In some of the
studies, no placebo supplement was included, also resulting in an unclear risk of performance and
supplementation studies, no placebo supplement was included, also resulting in an unclear risk of
detection bias qualification.
performance and detection Attrition rate was considerable
bias qualification. Attrition rate inwasmany studies, in
considerable and reasons
many for and
studies, dropout
and distribution across and
reasons for dropout intervention
distributiongroups
across were sometimes
intervention groupsnot reported.
were sometimesThis resulted
not reported.inThis
‘unclear
bias’ resulted
qualification withbias’
in ‘unclear respect to attrition
qualification with bias.
respectLack of trial
to attrition registration
bias. Lack of trialresulted in ‘unclear
registration resulted bias’
qualification with
in ‘unclear respect
bias’ to selective
qualification reporting
with respect bias. Areporting
to selective summary of the
bias. risk of bias
A summary analysis
of the risk of is shown
bias
analysis is shown in Figure 3. Visual inspection of the funnel plot of the protein
in Figure 3. Visual inspection of the funnel plot of the protein studies (Figure 4) showed no evidencestudies (Figure 4)
showed no evidence
for publication for publication
bias. Similar results werebias.found
Similar results
for were found
the other for the other
intervention types.intervention types.

.
Figure 2. Risk
Figure ofofbias
2. Risk bias analysis
analysis ofofthe
the individual
individual studies
studies included
included in the systematic
in the systematic review andreview
meta- and
meta-analysis. Green
analysis. Green dotdot = low
= low risk risk of bias;
of bias; yellowyellow dot = unclear
dot = unclear risk of bias.
risk of bias.
Nutrients 2019, 11, 1916 8 of 12
Nutrients
Nutrients2019,
2019,11,
11,xxFOR
FORPEER
PEERREVIEW
REVIEW 33 of
of 12
12

Figure
Figure 3.
3.Summary
Summary of
of the
the risk
risk of
of bias
bias analysis.
analysis.
Figure 3. Summary of the risk of bias analysis.

Figure
Figure 4.
Figure Funnel
4. Funnel plot
Funnel plot of
of the
plot of the protein
protein studies
studies included
included in
in the
the meta-analysis. SMD,
meta-analysis. SMD, standardized
SMD, standardized mean
standardized mean
mean
difference,
difference, SE (SMD), standard error of SMD.
difference, SE (SMD), standard error of SMD.
3. Other Potential Dietary Strategies for Weight Loss Maintenance
3.
3. Other
Other Potential
Potential Dietary
Dietary Strategies
Strategies forfor Weight
Weight Loss
Loss Maintenance
Maintenance
Overall, the meta-analysis of currently published studies on dietary strategies to prevent weight
Overall,
Overall,the themeta-analysis
meta-analysisof ofcurrently
currentlypublished
publishedstudies
studieson ondietary
dietarystrategies
strategiestotoprevent
preventweight
weight
regain above shows no evidence for a beneficial effect of a number of different dietary interventions
regain
regain above shows no evidence for a beneficial effect of a number of different dietary interventions
above shows no evidence for a beneficial effect of a number of different dietary interventions
related to glycemic index, fibre, green tea and EGCG, CLA, and miscellaneous other interventions
related
related toto glycemic
glycemic index,
index, fibre,
fibre, green
green teatea and
and EGCG,
EGCG, CLA,
CLA, andand miscellaneous
miscellaneous other
other interventions
interventions
(high MUFA, low fat, acarbose, capsaicin, a chromium picolinate/fibre/caffeine mixture, and GLA),
(high
(high MUFA, low fat, acarbose, capsaicin, a chromium picolinate/fibre/caffeine mixture, and
MUFA, low fat, acarbose, capsaicin, a chromium picolinate/fibre/caffeine mixture, and GLA),
GLA),
except for a beneficial effect of increased protein intake. When interpreting this outcome, it should be
except
except for a beneficial effect of increased protein intake. When interpreting this outcome, itshould
for a beneficial effect of increased protein intake. When interpreting this outcome, it shouldbebe
taken into account that the number of studies and the number of participants included addressing
taken
taken into
into account
account that
that the
the number
number of of studies
studies and
and the
the number
number of of participants
participants included
included addressing
addressing
each of these interventions is lower than the number of those related to protein intake. The question
each
each of
of these
these interventions
interventions is is lower
lower than
than the
the number
number ofof those
those related
related to
to protein
protein intake.
intake. The
The question
question
arises whether there are any other strategies that might also be effective but have not been tested so
arises
arises whether
whether therethere are
are any
any other
other strategies
strategies that
that might
might also
also be
be effective
effective but
but have
have not
not been
been tested
tested so
so
far with the paradigm of an initial weight loss phase followed by randomization to ad libitum diets
far with the paradigm of an initial weight loss phase followed by randomization
far with the paradigm of an initial weight loss phase followed by randomization to ad libitum diets to ad libitum diets
differing in nutrient composition.
differing
differing inin nutrient
nutrient composition.
composition.
We have recently reviewed the adipose-tissue-related mechanisms for weight regain and discussed
We
We havehave recently
recently reviewed
reviewed the the adipose-tissue-related
adipose-tissue-related mechanisms
mechanisms for for weight
weight regain
regain and
and
strategies for weight loss maintenance based on these mechanisms [35]. One of the suggested strategies
discussed
discussed strategies for weight loss maintenance based on these mechanisms [35]. One of the
strategies for weight loss maintenance based on these mechanisms [35]. One of the
was a diet enriched in omega3-PUFAs, either by omega3-PUFA containing foods, such as fatty fish,
suggested
suggested strategies
strategies waswas aa diet
diet enriched
enriched in in omega3-PUFAs,
omega3-PUFAs, either either by
by omega3-PUFA
omega3-PUFA containing
containing
nuts and seeds, and plant oils, or by omega3-PUFA fortified foods or supplementation. Omega3 fatty
foods,
foods, such
such as as fatty
fatty fish,
fish, nuts
nuts andand seeds,
seeds, andand plant
plant oils,
oils, or
or by
by omega3-PUFA
omega3-PUFA fortified
fortified foods
foods oror
acids have multiple potentially beneficial effects. They increase membrane fluidity, which may reduce
supplementation. Omega3 fatty acids have multiple potentially beneficial
supplementation. Omega3 fatty acids have multiple potentially beneficial effects. They increase effects. They increase
membrane
membrane fluidity,
fluidity, which
which maymay reduce
reduce cellular
cellular stress
stress [36].
[36]. Cellular
Cellular stress
stress in
in adipocytes
adipocytes may may develop
develop
during
during weight
weight lossloss when
when extracellular
extracellular matrix
matrix remodeling
remodeling is is not
not able
able to
to keep
keep track
track ofof adipocyte
adipocyte
Nutrients 2019, 11, 1916 9 of 12

cellular stress [36]. Cellular stress in adipocytes may develop during weight loss when extracellular
matrix remodeling is not able to keep track of adipocyte shrinking, and may be a signal for refilling of
the adipocyte and thus weight regain [35]. Additionally, omega3-PUFAs are important for a healthy
adipose tissue by improving the metabolic state and reducing the pro-inflammatory state [37,38]. No
effect of increased omega3 fatty acid intake on body weight in individuals with obesity has been found,
although a facilitating effect on weight loss in combination with other weight loss regimens has been
suggested [37]. However, this does not exclude a potential beneficial effect on weight regain in those
individuals that are not able to lower weight loss-induced cellular stress and inflammation in their
adipose tissue.
Polyphenols (e.g., resveratrol, curcumin, green tea components, and (isoflavones) occur naturally
in fruit and vegetables, green tea, black tea, red wine, coffee, chocolate, olives and olive oil, herbs and
spices, nuts, and algae. They may act as anti-oxidants, but also have anti-inflammatory actions and
metabolic effects (e.g., increased fat oxidation [39–41]). As can be concluded from the meta-analysis
above, there is currently no evidence that green tea or its component EGCG has a beneficial effect
on weight loss maintenance. Although there is some evidence for beneficial effects of polyphenols
on body weight gain in experimental animals, the effects on body weight in humans remain to be
determined [39,41,42]. A combination of omega3 fatty acids and polyphenols did not affect body
weight in healthy adolescents [43], but this has not been tested in the context of weight (re)gain.
Another mechanism suggested to play a role in the tendency for weight regain in humans is low
post-weight loss lipolysis [35,44]. Stimulation of lipolysis by capsaicin is not effective [30] (see above).
There also does not seem to be a beneficial effect of CLA supplementation on weight regain [25,26].
CLA is derived from the metabolism of linoleic acid, a poly-unsaturated essential fatty acid, in the
human gut, and is also present in animal products such as meat and dairy. CLA has been shown to
stimulate browning of adipose tissue.
Short chain fatty acids (SCFA) derived from fermentation of dietary fibres by the gut microbiota
have been shown to increase energy expenditure and fat oxidation, but their effect on long-term body
weight regulation in humans is unknown [45]. The microbiota composition, which may be influenced
by dietary means, may affect adipose tissue inflammation and thus potentially be of importance for
body weight regain [46]. A study by Sanchez et al. (2013) found that weight loss was better maintained
in women who consumed a probiotic supplement consisting of Lactobacillus rhamnosus, oligofructose,
and inulin, but no effect was seen in men [47].
Dietary strategies that combine multiple potentially beneficial mechanisms may be a more robust
way forward than single nutrient strategies. Moreover, some strategies may work better in some
individuals than others, depending on preference, which may be important for adherence, but also
on the individual proneness for weight regain and its underlying mechanism, which may differ from
one person to the next. Examples of whole-diet approaches are anti-inflammatory or anti-insulinemic
diets [48]. Both are associated with less weight (re)gain in observational cohort studies [48–50],
but randomized clinical trials are needed to assess their efficacy for prevention of weight (re)gain.

4. Conclusions
Although weight regain after successful weight loss is a major problem in many individuals,
relatively few randomized controlled trials have been performed exploring dietary strategies for more
successful weight maintenance. Our meta-analysis shows that, currently, only diets with increased
protein content have been shown to have a beneficial effect. Diets that combine a higher protein intake
with other potentially beneficial strategies, such as anti-inflammatory or anti-insulinemic diets, may
have more robust effects, but these have not been tested in randomized clinical trials yet.

Author Contributions: Conceptualization, M.A.v.B.; Analysis, M.A.v.B. and E.C.M.M.; Writing—Original Draft
Preparation, M.A.v.B.; Writing—Review & Editing, M.A.v.B. and E.C.M.M.
Funding: This research received no external funding.
To Have The Best Diet Products Choose Between This