nature metabolism

Perspective https://doi.org/10.1038/s42255-023-00864-1

Differential mechanisms affecting weight

loss and weight loss maintenance

Received: 23 September 2022 Michael Rosenbaum 1

& Gary Foster 2

Accepted: 13 July 2023

Published online: 23 August 2023 In most lifestyle, pharmacological and surgical interventions, weight loss
occurs over an approximately 6- to 9-month period and is followed by a
Check for updates
weight plateau and then weight regain. Overall, only about 15% of individuals
can sustain a 10% or greater non-surgical, non-pharmacological, weight
loss. A key question is the degree to which the genotypes, phenotypes
and environmental correlates of success in weight loss and weight loss
maintenance are continuous or dichotomous. This Perspective is a
comparison of the interactions of weight loss and maintenance with
genetic, behavioural, physiological and environmental homeostatic
systems and a discussion of the implications of these findings for research
in, and treatment of, obesity. Data suggest that weight loss and weight loss
maintenance are physiologically and psychologically different in many ways.
Consequently, individuals may require different interventions designed for
temporarily sustaining a negative energy balance during weight loss versus
permanently maintaining energy balance after weight loss.

Weight loss is defined as the change in body weight or adiposity from average, follow a typical pattern of an initial weight loss period of about
baseline to a future time point. Weight regain is defined as the change 6–9 months followed by a variable period of relative weight stability and
in body weight over time from the peak of weight loss1. Non-surgical then weight regain7–11. The degree of weight loss within this time period
weight loss and prolonged maintenance of non-surgically reduced varies substantially between interventions with greater weight loss
body weight are problematic2–4 largely owing to the predictable bio- due to pharmacotherapy12 and bariatric surgery11. However, the slope
logical consequences of central nervous system-mediated processes of weight regain after weight loss is similar between interventions7,13
that vary substantially between individuals. Negative energy balance (although the most recent generation of anti-obesity medications show
and decreases in stored energy (predominantly fat) are delimited by greater weight loss and longer weight loss maintenance, the studies
changes in metabolic, neuroendocrine, autonomic and behavioural are small in number and limited to 2 years in duration14). Mathematical
(increased hunger and delayed satiation) energy homeostatic systems, models have been developed to quantify the energy balance dynamics
which coordinately favour the regain of lost weight5. This regulation that account for this early weight loss followed by weight regain in
reflects evolutionary forces favouring defence of energy stores (body response to lifestyle15,16 and pharmacological17 interventions for obesity.
fat and fat-free mass) in service of survival and reproductive integrity Current interventions for obesity tend to treat weight loss and
under circumstances of restricted food access, coupled with environ- maintenance of reduced weight similarly. A key question in design-
mental availability of calorically dense foods and temporal reduction ing and prescribing weight management therapies is whether the
in external support of prolonged weight reduction. The result is that metabolic and behavioural factors that oppose weight loss and those
only about 15–25% of individuals can achieve and maintain a 10% or responsible for weight regain are all part of a continuum or are dis-
greater weight loss with ongoing lifestyle intervention6. crete entities that may require different therapeutic approaches.
The responses to any surgical or non-surgical (lifestyle or pharmaco­ The ‘continuum view’ would suggest that there are similar genotypic,
logical) interventions are highly heterogeneous, but most, on the phenotypic, behavioural and environmental drivers or determinants

Columbia University Irving Medical Center, Departments of Pediatrics and Medicine, Division of Molecular Genetics and the Irving Center for Clinical
1

and Translational Research (MR), New York, NY, USA. 2WW International, Perelman School of Medicine at the University of Pennsylvania, Department of
Psychiatry, Weight and Eating Disorders Program (GF), New York, NY, USA. e-mail: [email protected]

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1266

Perspective https://doi.org/10.1038/s42255-023-00864-1

of weight loss and weight loss maintenance, and that pharmacotherapy Table 1 | Comparison of dynamic weight loss and static
for weight loss would be equally effective if added to prevent weight reduced-weight maintenance
gain (that is, to prevent obesity) or regain (that is, to maintain reduced
body weight). The ‘dichotomous’ view would suggest different drivers During weight loss During maintenance of
reduced body weight
or determinants of success and different responses to therapies to
promote versus maintain weight loss. Prior metabolic Usual weight (energy Weight loss (negative
state balance) energy balance)

Overview: body weight regulation Current metabolic Negative energy balance Energy balance
The relative long-term constancy of body weight in adult humans (both state
with and without obesity) at an average of approximately 0.3–0.8 kg Changes compared to weight maintenance at usual body weight
per year1 (3,000 kcal a year of stored energy in the setting of approxi- ↓↓ REE ↓ REEa
mately 800,000 kcal ingested per year)18 is compatible with—but does
↓ NREE ↓ NREEa
not demonstrate—the operation of homeostatic mechanisms regulating Energy expenditure
body weight within the contexts of socio-environmental, psychological, ↑ Muscle contraction ↑ Muscle contraction
efficiency efficiencya
behavioural and other environments in which they operate. This so-called
‘norm-of-reaction’, which describes the powerful effect of environment ↓↓ T3, ↓↓T4, ↓↓TSH, ↑rT3 ↓ T3a, ↓T4a, ↓TSH, ↑rT3
on the salience (penetrance) of gene-mediated phenotypes, is commonly ↓↓ Leptin/fat mass a
↓ Leptin/fat massa
recognized in genetics19. What is most regulated is the rank order of Neuroendocrine
axes ↑ Cortisol Cortisol within normal
body fatness in a given population in a specific environment rather than range
absolute body weight. Thus, the prevalence of overweight and obesity
↑ GH No change or small ↑ GH
are increasing owing to the interactions of a relatively stable genotype
with a rapidly changing environment and, by association, lifestyle20,21. Autonomics ↑↑ PNS tone and ↓↓ SNS ↑ PNS tone and ↓SNS
tone tonea
Basically, the overall lack of substantial weight change in the
setting of highly variable caloric intake suggests that, at ‘usual’ levels ↓↓ Satiation ↓ Satiationa
of body energy stores in a given environment (sometimes called a ‘set Energy intake ↑↑ Hunger a
↑ Hungera
point’), energy intake and output are coupled and co-vary (if we eat
↑ Restraint ↓ Restraint
more we burn more, and if we exercise more we eat more) to maintain a
At least partially reversed by the administration of recombinant human leptin. NREE,
energy stores. If this co-variance persisted after weight loss then there non-resting energy expenditure; T3, triiodothyronine; T4, thyroxine; TSH, thyroid stimulating
would be a natural tendency to maintain a reduced weight and little hormone; rT3, reverse T3 (bio-inactive enantiomer of T3); GH, growth hormone. Based on
additional work would be required beyond the weight reduction phase. Rosenbaum et al.26,27, Sumithran et al.81,82 and Salem et al.83.
However, almost anyone who has ever lost weight can attest that it is
harder to sustain weight loss than to lose it. This is due, at least in part,
to the combined increase in the desire to eat in an obesogenic environ- weight loss done entirely in the controlled environment of a clinical
ment and a decrease in energy expenditure, which favour a return to research resource, it was possible to quantify the relative contributions
usual weight rather than maintenance of reduced weight. of some of these components to the maintenance of lost weight (Fig. 1).
Periods of active weight loss (negative energy balance) and sus- The degree to which these systems act coordinately to oppose down-
tained reduced body weight (energy balance) are both associated with ward perturbations from usual body weight are highly heterogeneous
hypometabolism (energy expenditure below what is predicted on the and reflect gene–environment interactions that ‘provoke’ physiological
basis of changes in body mass and composition) and hyperphagia and behavioural resistance.
(increased appetite compared to energy expenditure), which collec- Of note, some, but not all, studies of individuals following bari-
tively facilitate weight regain22–24. atric surgery show a ‘blunting’ of this disproportionate decline in
During weight loss (energy intake < energy expenditure) and energy expenditure resulting—in humans—from relative increases in
during attempts to maintain weight loss (energy intake = energy resting energy expenditure (REE) or the thermic effect of feeding as
expenditure), these systems are uncoupled, resulting in an increased well as a smaller effect on energy intake than seen following a lifestyle
desire to eat in the setting of a disproportionate decline in energy intervention28,29. In mice and rats, compensatory decreases in energy
expenditure. Following weight loss, there are changes in skeletal mus- expenditure following weight loss caused by bariatric surgery are
cle (increased chemomechanical work efficiency) and in autonomic minimal, perhaps owing to a lack of effect of surgical weight loss on
(decreased sympathetic nervous system (SNS) and increased parasym- SNS tone and the importance of SNS tone in regulating brown adipose
pathetic nervous system (PNS) tone) and neuroendocrine (decreased tissue thermogenesis in rodents.
leptin and bioactive thyroid hormones) functions, as well as increased Overall, while the metabolic and behavioural opposition to losing
hunger, delayed satiation and changes in food-related neuronal activa- weight and keeping it off are similar (Table 1), this does not mean that
tion, all of which favour weight regain. These findings are consistent the same is true for the factors that determine how those systems are
with the adaptations (increased exercise and dietary restraint) that engaged or what should be done to disengage them to assist in weight
are reported by those successful at long-term maintenance of reduced loss or keeping it off. If the various factors that oppose weight loss and
body weight, that is, little or no weight regain25. maintenance of reduced body weight are all part of a physiological
The simultaneous declines in expenditure and satiety, coupled continuum, then the genotypic and phenotypic predictors of these
with increased hunger and food reward, during and following weight should be identical. As discussed below, there is only a partial overlap
loss are synergistic in creating the optimal biological circumstances between genotypes and phenotypes that are premonitory of the degree
for weight regain. As shown in Table 1, the metabolic and behavioural of weight loss and of weight regain.
factors that oppose weight loss and maintenance of reduced body
weight represent a coordinate multisystem process predisposing Correlates of weight loss and maintenance of
towards regain of lost weight that is similar in content, if not magnitude, reduced body weight
between these two states. Genotypes
In a series of long-term inpatient studies26,27, in which participants Loos et al.30 reported that the overall heritability (meaning the herit-
with and without obesity were studied before and after a 10% or greater ability attributable to all genetic variants in a given individual) was 0.28

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1267

Perspective https://doi.org/10.1038/s42255-023-00864-1

Table 2 | Summary of data regarding the correlation of

d fo
od
d genotypes with weight loss and weight regain in the DPP,
r an ty, foo
un g e
tie n Look AHEAD and DiOGenes84–87
↑ H d, ↓ sa rceptio
ar pe n
usc
le rew traint, nt eate
d res amou SNP Systems likely to Weight loss Weight regain
↑ M cy an
e icie
n e of be affected
ton , eff daptiv esis
NS a gen
↑ P S tone 4, o BDNF rs6265 Energy intake Not significant 0.55 (0.21) kg per
N rm
↓ S y, T3, T in the
et pt (ref. 84) allele per year,
sati and le Autonomic and
TSH P = 0.011
neuroendocrine
Energy expenditure FTO Energy intake Not significant 1.56 (0.55) kg per
rs3751812 allele per 3 years,
Energy intake (ref. 85) P < 0.005
rs9939609 Energy intake Not significant 1.03 (0.52) kg per
Fig. 1 | Schematic of the multiple systems that favour weight regain after (ref. 85) allele per 3 years,
successful weight reduction based on long-term inpatient studies4,27,100–104. P < 0.049
Any one of these systems would, in and of itself, predispose towards weight
rs9922708 Energy intake Not significant 1.38 (0.55) kg per
regain. However, at usual weight they would tend to counterbalance each other, (ref. 85) allele per 3 years,
resulting in weight stability over time. If energy expenditure decreased then P < 0.012
so would energy intake (and, over time, leptin and thyroid hormones as well).
KTCD15 Energy intake 0.50 (0.24) kg per Not significant
Following weight loss, these systems act coordinately such that the decline in rs29941 (food reward) allele per year,
energy expenditure is now accompanied by an increase in energy intake and (ref. 84) P = 0.041
changes in autonomic and neuroendocrine systems that favour weight regain
MTIF3 Energy intake and −0.66 (0.20) kg Not significant
rather than the maintenance of a stable weight.
rs79884312 mitochondrial per allele per
(ref. 85) function year, P = 0.020
rs1885980 −0.70 (0.21) kg
(ref. 86) per allele per
year, P = 0.001
for thinness and 0.33 for severe obesity, somewhat lower than the 60%
heritability estimated from twin studies by Elder et al.31. Obesity is a rs1885988 −0.70 (0.21) kg
(ref. 86) per allele per
polygenic disease with additive effects of obesity-related allelic variants 4 years, P = 0.002
in assessing genetic risk of obesity32. This type of risk score has not, to
NEGR1 Energy intake and −0.79 (0.29) kg −0.35 (0.16) kg
our knowledge, been calculated for weight loss or regain, although it is
rs2815752 expenditure per allele per per allele per
a logical assumption that with sufficient data this will be forthcoming. (ref. 84) year, P = 0.006 year, P = 0.034
Table 2 shows the association of specific single-nucleotide polymor-
PPARG Adipogenesis −0.63 (0.22) −0.79 (0.27) kg
phisms (SNPs) with weight loss and regain from large clinical trials such rs1801282 Food choice kg per allele per allele per
as the Diabetes Prevention Program (DPP), Look AHEAD and DiOGenes, (ref. 84) per 0.5 years, year, P = 0.004
which generally studied candidate genes rather than carrying out a full P = 0.005
genome-wide association study analysis. While the effects of individual TFAP2B Relationship of −0.64 (0.31) kg Not significant
SNPs are relatively small, the aggregate effects of multiple obesity- rs2272903 dietary fat to per allele per
risk variants cannot yet be determined. Comparing SNPs that are (ref. 85) weight loss year, P = 0.037
rs987237
correlated with weight loss and weight regain, it is apparent that (ref. 87)
there is a greater dissimilarity than identity in these clinical trials.
TMEM18 Energy intake 0.62 (0.31) kg per
rs6548238 allele per year,
Phenotypes (ref. 84) P = 0.044
If the same systems opposed both the dynamic process of losing weight Many of the genes that are associated with weight loss and/or regain are also associated
and then weight maintenance, then there should be a significant nega- with risk for having obesity. The distinctive genotypes that are premonitory of losing weight
tive correlation of the degree of weight loss and the degree of weight and keeping it off suggests that different genetic profiles would be necessary for future
applications to a precision medicine approach to identify the best method of weight loss
regain within any intervention group. Patterns of weight loss (the slope
versus that of maintenance of reduced body weight for a given individual. For example,
and duration of the weight loss phase) are not significantly correlated it might be more likely that someone with the relevant TFAP2B SNP would be particularly
with patterns of weight regain. Weight loss in the initial phases of a life- responsive to a low-fat diet during weight loss but not maintenance of reduced body weight,
style intervention for obesity is predictive of subsequent weight loss, while the various FTO SNPs would suggest that dietary adherence requires more intensive
attention to prevent weight regain. Numbers in parentheses denote s.e.m.
final weight and response to pharmacotherapy33,34. However, it is not
significantly correlated with the absolute amount of weight regained
after weight loss or the slope of weight regain over time following
successful weight loss33. subsequent weight loss are not significantly correlated with subse-
As shown in Fig. 2, the degree of weight loss (defined as change quent weight regain. Similarly, studies of changes in appetite, ratings
in weight from baseline to lowest point) is not associated with the of fullness, hunger, desire to eat and prospective consumption during
degree of weight maintenance (defined as change in weight from caloric restriction on a very low-calorie diet are not premonitory of
the lowest point to the end of the study). Despite the dissimilarities weight regain35 and the degree of adaptive thermogenesis in REE during
in the slopes of weight loss between interventions, there are similar weight caloric restriction is significantly correlated with the degree of
changes in weight over time following successful weight reduction, weight loss, not the success in minimizing weight regain36,37.
which are largely independent of the degree of weight lost.
The clinical, biochemical and energetic phenotypes associated Environmental and demographic factors
with weight loss and weight regain are summarized in Table 3. Lifestyle and environment interact with genes that affect the level
Biochemical and clinical phenotypes associated with weight loss of body energy stored (body fatness) and the activity of the systems
and regain are distinct, much like the genotypes in Table 2. It is clear outlined during attempts to lose weight and keep it off. For example,
that multiple baseline measures that are significantly correlated with climate change and the coronavirus disease 2019 pandemic have been

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1268

Perspective https://doi.org/10.1038/s42255-023-00864-1

Non-surgical vs surgical

Non-surgical Surgical
0 0
–2

Change in weight (%)

Change in weight (%)
–4
–10
Control
–6
Banding
–8 Vertical-banded
–20 gastroplasty
–10
Gastric bypass
–12 < 3%
3–6% –30
–14
> 6%
–16
0 12 24 36 48 60 72 80 96 0 1 2 3 4 6 8 10 15
Month Year

Diet and exercise

Diet Exercise
0 0
–2 –2

Change in weight (%)

Change in weight (%)

Reduced lipids –4
–4
Reduced carbohydrates –6 < 150 min per week
–6 150–249 min per week
–8
–8 250–299 min per week
–10
–10 ≥ 300 min per week
–12
–12 –14
–14 –16
3 6 12 24 0 6 12 18 24
Month Month

Pharmacotherapy

0 0 0
–2
–2 –4
–4
Change in weight (%)

Change in weight (%)

–6
Weight loss (kg)

–8
–4 –8
–10 –12
–6 –12
–14 –16
–8 –16 –20
–18
–10 –20 –24
0 26 52 78 104 130 156 0 4 12 20 28 36 44 52 60 68 70 84 92 104104’ 0 4 8 12 20 24 36 48 60 72
Week Week Week

Orlistat Liraglutide NB Placebo Placebo Tirzepatide, 10 mg

Lorcaserin PT Semaglutide Tirzepatide, 5 mg Tirzepatide, 15 mg

Fig. 2 | Smoothed temporal changes in absolute weight or percentage of the change in weight over time between approximately 1 and 2 years after
weight in non-surgical, surgical, dietary, exercise and pharmacological the initiation of therapy. Data from tirzepatide reflects shorter time periods
interventions over time. Although there are substantial differences and cannot be extrapolated to determine what the pattern of weight change, if
in the amount of weight loss between treatments (surgical and any, will be. The weight loss period is clearly longer with tirzepatide than with
pharmacological treatments producing greater weight loss than lifestyle other interventions8,14,33,52,105–108. NB, naltrexone–bupropion; PT, phentermine–
changes) and within treatments (depending on type of surgery and topiramate.
pharmacological intervention), there are tremendous similarities in the slope

associated with increased food insecurity and other factors favour- treatments produce clinically meaningful weight losses across racial
ing weight gain and with an increasing prevalence of obesity and its and ethnic subgroups, the data are clear that weight loss is less among
comorbidities38,39. Similarly, against the backdrop of a relatively stable Black and Hispanic participants than in non-Hispanic White partici-
genotype, environmental factors such as stress, available portion size pants45–48, who then have the least success in avoiding weight regain49.
and socioeconomic status are associated with higher rates of obesity in In the Look AHEAD trial, weight losses among non-Hispanic White,
epidemiological assessments40,41. These same environmental variables African American and Hispanic participants, assessed by self-report,
and interventions to change them have different effects when it comes were 9.4%, 6.4% and 7.2%, respectively. At the 8-year assessments,
to weight loss (usually the change in weight achieved during the first non-Hispanic White participants had regained 50% of their initial
6–12 months following initiation of therapy) and weight regain. For weight loss compared to regains of only 22% and 33% in the African
example, the use of meal replacements has a strong effect on weight American and Hispanic groups, respectively48. The reasons for this
loss42,43 but no effects on weight regain44. are unclear, but greater research, especially using qualitative and
Perhaps the most notable difference between weight loss and mixed-method approaches, will help us to better understand the
weight regain is in the area of race and ethnicity. While behavioural factors that influence these differences49,50.

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1269

Perspective https://doi.org/10.1038/s42255-023-00864-1

Table 3 | Summary of comparisons of behavioural, weight is the same, then hormonal or other therapies should promote
biochemical and energetic phenotypes as correlates of and sustain weight loss. The adipocyte-derived hormone leptin is an
weight loss and weight regain in longitudinal studies88–95 excellent example of a therapeutic physiological intervention that is
uniquely efficacious in preventing weight regain but not in initiating
Weight loss Weight regain
or augmenting weight loss in humans.
Clinical phenotypes Leptin repletion in individuals who have congenital leptin defi-
Adiposity88–90,96 r = 0.27, P < 0.01 r = −0.24, P = 0.05 ciency (Lepob) results in weight loss and sustained maintenance of
reduced body weight55 suggesting that leptin may be effective as a
Hunger 88,89,97
r = −0.76, P < 0.001 Not significant
weight loss medication, or medication to prevent weight regain, in
Dietary restraint89,90,97,98 r = 0.35, P < 0.01 r = −0.47, P < 0.05 low-leptin states. Heymsfield et al. showed that leptin administration
Anxiety/neuroticism91 r = 0.50, P < 0.01 Not significant to participants with or without obesity at their usual weight (without
Depression92 Not significant Increased hazard ratio 1.31
hypoleptinaemia) had little effect on initiating weight loss, even in the
(P = 0.03) for depression setting of a lifestyle intervention56. Despite the fact that circulating
and 1.72 (P < 0.005) for leptin-to-fat mass ratios are severely reduced during weight loss, as
anti-depressant use. opposed to only mildly reduced during reduced-weight maintenance57,
Biochemical and energetic phenotypes leptin repletion in adults during caloric restriction has a small sup-
Leptin93,94 r = −0.31, P = 0.009 Not significant pressive effect on appetite and no effect on energy expenditure, the
hypothalamic–pituitary–thyroid hormone axis, or autonomic nervous
Ghrelin 93
r = 0.31, P = 0.014 Not significant
system function58–61. By contrast, short-term leptin repletion during
Angiotensin-converting Not significant r = 0.23, P = 0.012 reduced-weight stability reverses the hypometabolism, hyperphagia,
enzyme95
reduced concentrations of bioactive thyroid hormones and decreased
TSH99 r = 0.36, P < 0.001 Not significant SNS that characterize the weight-reduced state26,62–64 (Table 1), provid-
Free T3 (ref. 99) r = 0.33, P = 0.002 Not significant ing an example of selective pharmacotherapy to prevent weight regain.
REE90 r = −0.38, P = 0.01
Environmental and behavioural manipulations
When considering whether weight loss and the reduced maintenance
state are continuous or dichotomous, it is important to consider the
Treatments to promote and maintain weight loss behavioural and psychological contexts in which both of these pro-
Pharmacology cesses occur65. While the predictors of success from a behavioural
Most studies of pharmacological treatments for obesity also involve and psychological perspective are largely the same for both weight
(typically modest) lifestyle interventions, and the pharmacological loss and weight loss maintenance (increased restraint, lower disinhibi-
efficacy must be evaluated in the context of treatment versus control tion, augmented by self-monitoring and self-efficacy)44, the context
(usually lifestyle intervention plus placebo) groups. As noted above, in which these cognitive and behavioural skills must be practiced are
most pharmacotherapy for obesity generally results in weight loss for dramatically different.
a period of 6–9 months, followed by a plateau period and then weight First, weight loss occurs over a fixed interval (6–9 months for
regain. Many clinical trials do not report weight during pharmaco­ most), while maintenance requires lifelong attention. The mindset
therapy beyond the first year of treatment, which may not be long required to approach a goal with a clear beginning, middle and end
enough to reveal weight regain (Fig. 2). Without longer-term data, it (weight loss) is quite different from one that requires the persistence of
is not possible to determine the rate of weight regain, as exemplified newly acquired behaviours with no clear end (keeping weight off). The
by recent GLP-1 receptor agonists, in which such data are limited for lack of a fixed endpoint leads to behavioural fatigue in key behaviours
semaglutide14,51 and not yet available for tirzepatide52. The cessation such as self-monitoring66. Second, the goal of weight loss is to lose some
of weight loss after initiation of pharmacotherapy does not mean fixed amount of weight following a prolonged period of weight gain,
that these drugs have become ineffective after weight loss, because while successful weight loss maintenance requires reversing small
stopping therapy will generally result in rapid weight regain. What is amounts of weight regain as they occur67. The cognitive, behavioural
not known is whether different doses of pharmacotherapies or differ- and affective responses are also quite different in these two scenarios.
ent modalities (injection versus oral) would be sufficient to sustain Weight loss is characterized by experiences of success as each weight
weight loss. loss interval is met, while maintaining a reduced weight consists of
Numerous combinations of weight loss medications have been reversing some unwanted outcome and just trying to get back to where
shown to be more effective than either medication alone in promoting ‘I started’50. In addition, most have at least partially successful experi-
weight loss12,53. There are few studies examining whether the addi- ences of weight loss (for example, some weight is lost even if it is not
tion of other weight loss medications will result in additional weight the amount desired by the individual or their health care provider)68,
reduction once weight plateaus have been reached. In a 12-week study, whereas few have successful experiences keeping it off69, thereby
Tronieri et al.54, examined the effects of adding a submaximal dose of undermining self-efficacy.
phentermine to the pharmacological treatment for obesity in par- These issues are further exacerbated by differences in support and
ticipants who had already lost weight with liraglutide and found no reinforcement for losing weight and not regaining it. During weight
additional weight loss in either the liraglutide–placebo or liraglu- loss, patients receive assistance and encouragement from profession-
tide–phentermine group. While this would suggest that the addition als, family members and friends as well as like-minded people who form
of a complementary weight loss drug will not change the slope of the both digital and in-person communities who are going through the
weight versus time curve once weight loss has reached its maximum, same process. Family support may include trying different foods as
further clinical trials of other combinations and studies of longer dura- a family unit, storing problematic foods out of sight, frequenting res-
tion are clearly needed. taurants that serve healthier choices and adjusting family schedules to
accommodate physical activity70. During attempts to sustain a reduced
Physiological manipulations weight, the novelty of the behavioural changes and weight loss have
If the same physiological mechanisms underlie the coordinate multi- diminished; those around the patient begin to see the weight loss jour-
system oppositions to weight loss, and maintenance of a reduced body ney as ‘over’, and some desire a return to the ‘way it used to be’. Finally,

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1270

Perspective https://doi.org/10.1038/s42255-023-00864-1

while many structured programmes provide virtual and in-person 18 months (with over 90% retention) were trivial, ranging from gains of
support for weight loss, few are focused on weight loss maintenance71. 2.5 kg (±6.7 kg) to 4.9 kg (±6.5 kg) across conditions. There were similar
Reinforcement during and after weight loss are quite different. data when examining those who gained 2.3 kg or more over 18 months,
Consuming highly palatable foods and being sedentary, for most, are which ranged from 46% to 72% of participants across conditions.
quite reinforcing in the immediate short term. Altering these behav- Two other studies, both by Yancy et al., examined weight change
ioural patterns requires deferral of the immediate reinforcement and in programmes specifically focused on weight loss maintenance using
focusing on the benefits to be accrued over time, such as weight loss, a variety of financial incentive strategies. The first study78 randomized
improved health and a better quality of life. During the weight loss 191 participants who had recently lost approximately 11% of their body
phase, individuals experience a multitude of reinforcers, including a weight to different financial incentives or a control group. After 12
sense of control, positive feedback about changes in behaviour and months of the weight loss maintenance intervention (approximately
weight from those around them, changes in clothing sizes and the abil- 80% retention across groups), there were no differences among the
ity to perform tasks such as walking and playing, all of which counter groups, but the weight change across conditions was remarkably small,
the lack of immediate benefits associated with decreased intake and ranging from −0.3 to −1.8 kg, and 67% to 79% of participants maintained
increased activity. While these are not all immediate reinforcers, they weight loss (defined as a regain of ≤1.36 kg). In the second study79, 259
constitute frequent reminders that the changes that have been made participants who had lost approximately 12% of their body weight were
have positive effects. By contrast, few if any will compliment individu- randomized to different types of financial incentive. At 12 months (88%
als on their lack of regain, or even recognize that keeping weight off retention) after the initial weight loss period, mean weight changes
requires long-term maintenance of lifestyle changes that are difficult ranged from −0.3 to −3.0 kg, and between 59% and 66% of participants
and not the same lifestyles as individuals ‘naturally’ at the same weight. had sustained weight loss (defined as <1.36 kg regain). The results
Similarly, clothing sizes don’t change and there is not an ever-increasing across these trials are remarkable and stand in stark contrast to those
list of activities that people can engage in this new steady state of that follow people from the beginning of treatment and report that,
weight regain. This lack of reinforcement is even more striking because after dietary weight loss, only approximately 25% of participants are
it occurs at the same time that the same, or even greater, amount of successful long-term maintainers80. These data suggest, but do not
behavioural effort is required but with less support and reinforcement demonstrate, that when weight loss and weight loss maintenance
than in the weight loss phase65. are considered, evaluated or treated separately, even though largely
A final distinction between weight loss and weight regain centres the same behaviours are required, the outcomes are quite different
on physical activity. During weight loss, physical activity can increase in the long term. It is possible that viewing treatment for weight loss
the magnitude of weight loss by about 20%, but it is possible to lose a maintenance separately from that for weight loss provides partici-
substantial (>5%) amount of weight without engaging in physical activ- pants a sense of a new start, with a renewed focus on the behaviours
ity72. By contrast, physical activity is strongly correlated with success required for weight loss maintenance that is beneficial78,79. It is also pos-
in sustaining weight loss73. The mechanisms for this correlation that sible that those who enrolled in these trials were a select sample of the
‘allows’ increased energy expenditure without provoking an overriding most successful and their results may not generalize beyond the
increase in energy intake are not clear but might include the increased individuals sampled.
production of endorphins and overall anxiolytic effects of exercise, as
well as the obvious effects on energy expenditure74. There appear to be Discussion
differential effects of the quantity of physical activity on weight loss In summary, obesity is a complex phenotype reflecting the interactions
and reduced-weight maintenance. In the DiOGenes study75, the amount of numerous genes that have been ‘selected’ for in service of repro-
of physical activity was not significantly correlated with the degree ductive integrity with modern environments. Data regarding genetic,
of weight loss but was significantly correlated with the rate of weight phenotypic and environmental factors with the likelihood of losing
regain—suggesting that the uncoupling between energy intake and weight versus the likelihood of keeping it off suggest that there are
output discussed above and opposing reduction in energy stores may differences in the ‘inner machinations’ that lead to weight loss versus
be less potent after weight loss. Wang et al. found that the addition of the maintenance of a reduced body weight. This is further supported by
treadmill exercise during weight loss was significantly negatively cor- studies examining the differences in responses to therapies designed
related with weight regain after weight loss, even though the exercise specifically to maintain negative energy balance (weight loss) versus
and dietary regimen was not continued76. energy balance after weight loss. Overall, the biological and behavioural
Another factor to consider in the continuous versus dichotomous context or milieu of active weight loss and the maintenance of a reduced
distinction is the efficacy of programming specifically designed for body weight are quantitatively different. The molecular physiological
weight loss maintenance with behavioural interventions. It is possible bases of weight regain, and their interactions with environment and
that viewing treatment for weight loss maintenance, separately from lifestyle include some of these same factors and mechanisms, but there
that for weight loss, provides participants a sense of a new start with are likely to be different factors and mechanisms as well.
a renewed focus on the behaviours required for weight loss mainte- The genotypic, phenotypic and environmental predicates of suc-
nance, rather than being perceived as a continuation of behaviours that cess in weight loss and sustaining it are highly heterogeneous. These
have been practiced over the preceding months and years. As already data suggest that treatments to sustain weight loss are more likely to
reviewed, the bulk of the literature provides quite sobering data on be successful if they specifically consider weight regain as a distinct
long-term weight regain for most individuals, generally based on study process with different physiological and behavioural contexts.
designs that follow people from baseline through a weight loss phase Unlike the extensive investigations of weight loss interventions,
and then usually through 12-month or 24-month follow-up assessments there remains a relatively small body of research into therapeutic
with similar interventions. By contrast, Wing et al.77 followed 314 par- approaches to sustain lost weight, even though this is the most difficult
ticipants with a minimum weight loss of 10% over the previous 2 years problem for most individuals with obesity. There is a clear opportunity
who were recruited into an 18-month trial of self-regulation strategies for precision medicine approaches to target both of these processes
(either quarterly newsletters without daily weighing or regular internet with therapies that address the most salient risk factors in any given
or face-to-face contact with daily weighing) designed specifically for individual. Interventions to treat, or even prevent, obesity could be bet-
the maintenance of weight loss. Participants had, on average, achieved ter leveraged to yield algorithmic approaches to leverage current inter-
approximately an 18% weight loss before study entry. Despite this ventions and guide the development of new ones based on genotypic,
large amount of weight loss, the changes in body weight over the next phenotypic and environmental factors to maximize effectiveness.

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1271

Perspective https://doi.org/10.1038/s42255-023-00864-1

Reporting summary 24. Thomas, J., Bond, D., Phelan, S., Hill, J. & Wing, R. Weight-loss
Further information on research design is available in the Nature Port- maintenance for 10 years in the National Weight Control Registry.
folio Reporting Summary linked to this article. Am. J. Prev. Med. 46, 17–23 (2014).
25. Phelan, S., Halfman, T., Pinto, A. & Foster, G. Behavioral and
References psychological strategies of long-term weight loss maintainers
1. Zheng, Y. et al. Associations of weight gain from early to middle in a widely available weight management program. Obesity 28,
adulthood with major health outcomes later in life. JAMA 318, 421–428 (2020).
255–259 (2017). 26. Rosenbaum, M. & Leibel, R. 20 years of leptin: role of leptin in
2. Wing, R. & Phelan, S. Long-term weight maintenance. Am. J. Clin. energy homeostasis in humans. J. Endocrinol. 223, T83–T96 (2014).
Nutr. 82, 222S–225S (2005). 27. Rosenbaum, M., Hirsch, J., Murphy, E. & Leibel, R. The effects
3. Phelan, S. & Wing, R. Prevalance of successful weight loss. of changes in body weight on carbohydrate metabolism,
Arch. Int. Med. 165, 2430 (2005). catecholamine excretion, and thyroid function. Am. J. Clin. Nutr.
4. Aronne, L. et al. Describing the weight-reduced state: physiology, 71, 1421–1432 (2000).
behavior, and interventions. Obesity 29, S9–S24 (2021). 28. Munzberg, H., Laque, A., Yu, S., rezai-Zadeh, K. & Berthoud, H.
5. Rosenbaum, M. & Leibel, R. in Treatment of the Obese Patient Appetite and body weight regulation after bariatric surgery.
2nd edn (eds. R. Kushner & D. Bessesen) Ch. 7 (Springer, 2014). Obes. Rev. 16, 77–90 (2015).
6. Unick, J. et al. Effectiveness of lifestyle interventions for 29. Das, S. et al. Long-term changes in energy expenditure and body
individuals with severe obesity and type 2 diabetes: results from composition after massive weight loss induced by gastric bypass
the Look AHEAD trial. Diabetes Care 34, 2152–2157 (2011). surgery. Am. J. Clin. Nutr. 78, 28–30 (2003).
7. Wadden, T. et al. The Look AHEAD study: a description of the 30. Loos, R., Burant, C. & Schur, E. Strategies to understand the
lifestyle intervention and the evidence supporting it. Obesity 14, weight-reduced state: genetics and brain imaging. Obesity 29,
737–752 (2006). S39–S50 (2021).
8. Sjostrom, L. et al. Effects of bariatric surgery on mortality in 31. Elder, S. et al. Effect of body composition methodology on
swedish obese subjects. N. Eng. J. Med. 357, 741–752 (2007). heritability estimation of body fatness. Open Nutr. J. 5, 48–58
9. Sacks, F. et al. Comparison of weight-loss diets with different (2012).
compositions of fat, protein, and carbohydrates. N. Eng. J. Med. 32. Khera, A. V. et al. Polygenic prediction of weight and obesity
360, 859–873 (2009). trajectories from birth to adulthood. Cell 177, 587–596 (2019).
10. Unick, J. et al. The long-term effectiveness of a lifestyle 33. Unick, J. et al. Weight change in the first 2 months of a lifestyle
intervention in severely obese individuals. Am. J. Med. 126, intervention predicts weight changes 8 years later. Obesity 23,
236–242 (2013). 1353–1356 (2015).
11. Stanford, F. et al. Comparison of short and long-term outcomes 34. Tronieri, S. et al. Early weight loss in behavioral treatment predicts
of metabolic and bariatric surgery in adolescents and adults. later rate of weight loss and response to pharmacotherapy.
Front. Endocrinol. 11, 157 (2020). Ann. Behav. Med. 53, 290–295 (2019).
12. Tchang, B. G., Aras, M., Kumar, R. B. & Aronne, L. J. in Endotext 35. Turicchi, J. et al. Associations between the proportion of
(ed. K. Feingold) (MDText.com, 2021). fat-free mass loss during weight loss, changes in appetite, and
13. The Look AHEAD Research Group. The Look AHEAD Study: subsequent weight change: results from a randomized 2-stage
a description of the lifestyle intervention and the evidence dietary intervention trial. Am. J. Clin. Nutr. 111, 536–544 (2020).
supporting it. Obesity 14, 737–752 (2006). 36. Martins, C., Gower, G., Hill, J. & Hunter, G. Metabolic adaptation is
14. Garvey, W. et al. Two-year effects of semaglutide in adults with not a major barrier to weight-loss maintenance. Am. J. Clin. Nutr.
overweight or obesity: the STEP 5 trial. Nat. Med. 28, 2083–2091 112, 558–565 (2020).
(2022). 37. Thom, G. et al. The role of appetite-related hormones, adaptive
15. Guo, J., Brager, D. & Hall, K. Simulating long-term human thermogenesis, perceived hunger and stress in long-term
weight-loss dynamics in response to caloric restriction. Am. J. weight-loss maintenance: a mixed-methods study. Eur. J. Clin.
Clin. Nutr. 107, 558–565 (2018). Nutr. 74, 622–632 (2020).
16. Hall, K. et al. Quantificaiton of the effect of energy imbalance on 38. Dietz, W. & Pryor, S. How can we act to mitigate the global
bodyweight. Lancet 378, 826–837 (2011). syndemic of obesity, undernutrition, and climate change?
17. Hall, K., Sanghvi, A. & Gobel, B. Proportional feedback control Curr. Obes. Rep. 11, 61–69 (2022).
of energy intake during obesity pharmacotherapy. Obesity 25, 39. Pryor, S. & Dietz, W. The COVID-19, obesity, and food insecurity
2088–2091 (2017). syndemic. Curr. Obes. Rep. 11, 70–79 (2022).
18. Ford, E. & Dietz, W. Trends in energy intake among adults in 40. Newton, S., Braithwaite, D. & Akinyemiju, T. Socio-economic
the United States: findings from NHANES. Am. J. Clin. Nutr. 97, status over the life course and obesity: systematic review and
848–853 (2013). meta-analysis. PLoS ONE 15, e0177151 (2017).
19. Lewontin, R. The analysis of variance and the analysis of causes. 41. Moore, C. & Cunningham, S. Social position, psychological stress,
Int. J. Epidemiol. 35, 520–525 (2006). and obesity: a systematic review. J. Acad. Nutr. Diet. 112, 418–426
20. Ogden, C. et al. Trends in obesity prevalence by race and hispanic (2012).
origin-1999–2000 to 2017–2018. JAMA 324, 1208–1210 (2020). 42. Heymsfield, S., van Mierlo, C., van der Knaap, H., Heo, M. &
21. Koethe, J. R. et al. Rising obesity prevalence and weight gain Frier, H. Weight management using a meal replacement strategy:
among adults starting antiretroviral therapy in the United States meta and pooling analysis from six studies. Int. J. Obes. Relat.
and Canada. AIDS Res. Hum. Retroviruses 32, 50–58 (2016). Metab. Disord. 27, 547–549 (2003).
22. Rosenbaum, M. & Leibel, R. in Novel Insights into Adipose Cell 43. Astbury, N. et al. A systematic review and meta-analysis of the
Functions, Research and Perspectives in Endocrine Interactions effectiveness of meal replacements for weight loss. Obes. Rev.
(eds. Y. Christen & K. Clément) 121–133 (Springer, 2010). 20, 569–587 (2019).
23. Hall, K. & Kahan, S. Maintenance of lost weight and long-term 44. Varkevisser, R., Van Stralen, M., Kroeze, W., Ket, J. & Stenhuis, I.
management of obesity. Med. Clin. North Am. 102, 183–197 Determinants of weight loss maintenance: a systematic review.
(2018). Obes. Rev. 20, 171–211 (2019).

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1272

Perspective https://doi.org/10.1038/s42255-023-00864-1

45. Fitzgibbon, M. et al. Weight loss and African-American women: 65. Wadden, T., Butryn, M. & Byrne, K. Efficacy of lifestyle modification
a systematic review of the behavioural weight loss intervention for long-term weight control. Obes. Res. 12, 151S–162S (2004).
literature. Obes. Rev. 13, 192–213 (2012). 66. Bryant, D. T. L. L. M. et al. Efficacy of a commercial weight
46. Wingo, B. & Ard, T. C. J. Weight loss differences by race. Obes. Rev. management program compared with a do-it-yourself approach:
15, 46–61 (2014). a randomized clinical trial. JAMA Netw. Open. 5, e2226561 (2022).
47. West, D. et al. Weight loss experiences of African American, 67. Hayes, J. et al. Recovery from weight regain among long-term
Hispanic, and Non-Hispanic white men and women with type 2 weight loss maintainers in WW. Obesity 30, 2404–2413 (2022).
diabetes: the Look AHEAD trial. Obes 27, 1275–1284 (2019). 68. Wadden, T. et al. Four-year weight losses in the Look AHEAD
48. The Look AHEAD Research Group. Eight-year weight losses with study: factors associated with long-term success. Obesity 19,
an intensive lifestyle intervention: the Look AHEAD study. Obesity 1987–1998 (2011).
22, 5–13 (2014). 69. Wadden, T., Tronieri, J. & Butrybn, M. Lifestyle modification
49. Kinsey, A. et al. Factors associated with weight loss maintenance approaches for the treatment of obesity in adults. Am. Psychol.
and weight regain among african american and white adults 75, 235–251 (2020).
initially successful at weight loss. J. Racial Ethn. Health Disparities 70. Cornelius, T. et al. How prescriptive support affects weight loss in
9, 546–565 (2022). weight-loss intervention participants and their untreated spouses.
50. Reyes, N. et al. Similarities and differences between weight loss Health Psychol. 37, 775–781 (2018).
maintainers and regainers: a qualitative analysis. J. Acad. Nutr. Diet. 71. Gudzune, K. et al. Efficacy of commercial weight-loss programs:
112, 449–505 (2012). an updated systematic review. Ann. Intern. Med. 162, 501–512
51. Bergman, N., Davies, M., Lingvay, I., Knop, F. Semaglutide for the (2015).
treatment of overweight and obesity: a review. Diab. Obes. Metab. 72. Curion, C. & Lourenc, P. Long-term weight loss after diet and
https://doi.org/10.1111/dom.14863. (2022). exercise: a systematic review. Int. J. Obes. 29, 1168–1174 (2005).
52. Jastrebogg, A. et al. Tirzepatide once weekly for the treatment of 73. Donnelly, J. et al. American College of Sports Medicine Position
obesity. N. Eng. J. Med. 387, 205–216 (2022). Stand. Appropriate physical activity intervention strategies for
53. Hollander, P. et al. Coadministration of canagliflozin and weight loss and prevention of weight regain for adults. Med. Sci.
phentermine for weight management in overweight and obese Sports Exerc. 41, 459–471 (2009).
individuals without diabetes: a randomized clinical trial. Diabetes 74. Ai, X., Yang, J., Lin, Z. & Wan, X. Mental health and the role of
Care 40, 632–639 (2017). physical activity during the COVID-19 pandemic. Front. Psychol.
54. Tronieri, J. et al. Effects of liraglutide plus phentermine in adults 12, 759987 (2021).
with obesity following 1 year of treatment by liraglutide alone: a 75. van Baak, M., Hul, G., Astrup, A. & Saris, W. Physical activity,
randomized placebo-controlled pilot trial. Metabolism 96, 83–91 weight loss, and weight maintenance in the DiOGenes
(2019). Multicenter Trial. Front Nutr. 8, 683369 (2021).
55. Farooqi, I. et al. Effects of recombinant leptin therapy in a child 76. Wang, X. et al. Weight regain is related to decreases in physical
with congenital leptin deficiency. N. Eng. J. Med. 341, 879–884 activity during weight loss. Med. Sci. Sports Exerc. 40, 1781–1788
(1999). (2008).
56. Heymsfield, S. B. et al. Recombinant leptin for weight loss in 77. Wing, R., Tate, D., Gorin, A., Raynor, H. & Fava, J. A self-regulation
obese and lean adults: a randomized, controlled, dose-escalation program for maintenance of weight loss. N. Eng. J. Med. 355,
trial. JAMA 292, 1568–1575 (1999). 1563–1571 (2006).
57. Rosenbaum, M. et al. Effects of weight change on plasma leptin 78. Yancy, W. Jr et al. Effect of escalating financial incentive rewards
concentrations and energy expenditure. J. Clin. Endocrinol. on maintenance of weight loss: a randomized clinical trial.
Metab. 82, 3647–3654 (1997). JAMA Netw. Open. 2, e1914393 (2019).
58. Hukshom, C. et al. The effect of pegylated recombinant human 79. Yancy, W. Jr et al. Financial incentive strategies for maintenance
leptin (PEG-OB) on weight loss and inflammatory status in obese of weight loss: results from an internet-based randomized
subjects. Int. J. Obes. 26, 504–509 (2002). controlled trial. Nutr. Diabetes 8, 33 (2021).
59. Hukshorn, C., Meneheere, P., Westerterp-Plantenga, M. & Saris, W. 80. Flore, G. et al. Weight maintenance after dietary weight loss:
The effect of pegylated human recombinant leptin (PEG-OB) on systematic review and meta-analysis on the effectiveness of
neuroendocrine adaptations to semi-starvation in overweight behavioural intensive intervention. Nutrients 14, 1259 (2022).
men. Eur. J. Endocrinol. 148, 649–655 (2003). 81. Sumithran, P. & Proietto, J. The defence of body weight: a
60. Lejeune, M., Hukshorn, C., Saris, W. & Westerterp-Plantenga, physiological basis for weight regain after weight loss. Clin. Sci.
M. Effect of dietary restraint during and following pegylated 124, 231–241 (2013).
recombinant leptin (PEG-OB) treatment of overweight men. Int. J. 82. Sumithran, P. et al. Long-term persistance of hormonal
Obes. 27, 1494–1499 (2003). adaptations to weight loss. N. Eng. J. Med. 365, 1597–1604 (2011).
61. Saris, C. H. W., Westerterp-Plantenga, M., Farid, A., Smith, F. & 83. Salem, V. et al. Weight loss by low-calorie diet versus gastric
Campfield, L. Weekly subcutaneous pegylated recombinant bypass surgery in people with diabetes results in divergent brain
native human leptin (PEG-OB) administration in obese men. activation patterns: a functional MRI Study. Diabetes Care. 44,
J. Clin. Endocrinol. Metab. 85, 4003–4009 (2000). 1842–1851 (2021).
62. Rosenbaum, M. et al. Low dose leptin reverses skeletal muscle, 84. Delahanty, L. et al. Genetic predictors of weight loss and weight
autonomic, and neuroendocrine adaptations to maintenance of regain after intensive lifestyle modification, metformin treatment,
reduced weight. J. Clin. Invest. 115, 3579–3586 (2005). or standard care in the Diabetes Prevention Program. Diabetes
63. Rosenbaum, M., Sy, M., Pavlovich, K., Leibel, R. & Hirsch, J. Leptin Care 35, 363–366 (2012).
reverses weight loss–induced changes in regional neural activity 85. McCaffery, J. et al. FTO predicts weight regain in the Look AHEAD
responses to visual food stimuli. J. Clin. Invest. 118, 2583–2591 clinical trial. Int. J. Obes. 37, 1545–1552 (2013).
(2008). 86. Papandonatos, G. et al. Genetic predisposition to weight loss
64. Rosenbaum, M. et al. Triiodothyronine and leptin repletion in and regain with lifestyle intervention: analyses from the Diabetes
humans similarly reverse weight-loss induced changes in skeletal Prevention Program and the Look AHEAD randomized controlled
muscle. Am. J. Physiol. Endocrinol. Metab. 315, E771–E779 (2018). trials. Diabetes 64, 4312–4321 (2015).

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1273

Perspective https://doi.org/10.1038/s42255-023-00864-1

87. Larsen, L. et al. Analyses of single nucleotide polymorphisms in 104. Rosenbaum, M. & Leibel, R. L. Models of energy homeostasis in
selected nutrient-sensitive genes in weight-regain prevention: the response to maintenance of reduced body weight. Obesity 24,
DiOGenes study. Am. J. Clin. Nutr. 95, 1254–1260 (2012). 1620–1629 (2016).
88. Batra, P. et al. Eating behaviors as predictors of weight loss in a 105. Unick, J., Pellegrini, C., Demose, K. & Dorfman, L. Initial weight
6-month weight loss intervention. Obesity 21, 2256–2263 (2013). loss response as an indicator for provding early rescue efforts to
89. Womble, L., Williamson, D., Greenway, F. & Redmann, S. improve long-term treatment outcomes. Curr. Diab Rep. 17, 69
Psychological and behavioral predictors of weight loss during (2018).
drug treatment for obesity. Int. J. Obes. 25, 340–345 (2001). 106. Dong, Z., Xu, L., Liu, H., Lv, Y. & Li, L. Comparative efficacy of
90. Vogels, N., Diepvens, K. & Westerterp-Plantenga, M. Predictors five long-term weight loss drugs: quantitative information for
of long-term weight maintenance. Obes. Res. 13, 2162–2168 medication guidelines. Obes. Rev. 18, 1377–1385 (2017).
(2005). 107. Wadden, T., Webb, V., Moran, C. & Bailer, B. Lifestyle modification
91. Munro, I., Bore, M., Nunro, D. & Garg, M. Using personality as a for obesity. Circulation 125, 1157–1170 (2012).
predictor of diet induced weight loss and weight management. 108. Smith, S. et al. Multicenter, placebo-controlled trial of lorcaserin
Int J. Behav. Nutr. Phys. Act. 8, 129 (2011). for weight management. N. Eng. J. Med. 363, 245–256 (2010).
92. Price, D. et al. Depression as a predictor of weight regain among
successful weight losers in the diabetes prevention program. Acknowledgements
Diabetes Care 36, 216–221 (2013). These studies were supported, in part, by grants from the National
93. Labayen, I. et al. Role of baseline leptin and ghrelin levels on Institutes of Health (UL1 TR00040 to M.R.), the National Institute of
body weight and fat mass changes after an energy-restricted diet Diabetes and Digestive and Kidney Diseases (R01 30583 and R01
intervention in obese women: effects on energy metabolism. 64773 to M.R., R01 DK130851 to G.F.) and the National Heart, Lung, and
J. Clin. Endocrinol. Metab. 96, E996–E1000 (2011). Blood Institute (1UG3 H163121 to G.F.).
94. Shih, L. et al. Leptin, superoxide dismutase, and weight loss: initial
leptin predicts weight loss. Obesity 14, 2184–2192 (2006). Author contributions
95. Wang, P. et al. Circulating ACE is a predictor of weight loss The authors contributed equally to the writing of this manuscript.
maintenance not only in overweight and obese women, but also
in men. Int. J. Obes. 36, 1545–1551 (2012). Competing interests
96. Hansen, D. et al. Predictors of weight loss and maintenance M.R. declares no competing interests. G.F. is an employee and
during 2 years of treatment by sibutramine in obesity. Results shareholder at WW International.
from the European multi-centre STORM trial. Sibutramine Trial
of Obesity Reduction and Maintenance. Int. J. Obes. 25, 496–501 Additional information
(2001). Supplementary information The online version
97. Foster, G. et al. The Eating Inventory in obese women: clinical contains supplementary material available at
correlates and relationship to weight loss. Int. J. Obes. 22, https://doi.org/10.1038/s42255-023-00864-1.
778–785 (1998).
98. Stunkard, A. et al. Weight change in depression: influence of Correspondence should be addressed to Michael Rosenbaum.
‘disinhibition’ is mediated by body mass and other variables.
Psychiatr. Res. 38, 197–200 (1991). Peer review information Nature Metabolism thanks Michael Lean and
99. Wolters, B., Lass, N. & Reinehr, T. TSH and free triiodothyronine Steven Heymsfield for their contribution to the peer review of this
concentrations are associated with weight loss in a lifestyle work. Primary Handling Editor: Ashley Castellanos-Jankiewicz,
intervention and weight regain afterwards in obese children. in collaboration with the Nature Metabolism team.
Eur. J. Endocrinol. 168, 323–329 (2013).
100. Leibel, R., Rosenbaum, M. & Hirsch, J. Changes in energy Reprints and permissions information is available at
expenditure resulting from altered body weight. N. Eng. J. Med. www.nature.com/reprints.
332, 621–628 (1995).
101. Rosenbaum, M., Hirsch, J., Gallagher, D. & Leibel, R. Long-term Publisher’s note Springer Nature remains neutral with regard to
persistence of adaptive thermogenesis in subjects who have jurisdictional claims in published maps and institutional affiliations.
maintained a reduced body weight. Am. J. Clin. Nutr. 88, 906–912
(2008). Springer Nature or its licensor (e.g. a society or other partner) holds
102. Goldsmith, R. et al. Effects of experimental weight perturbation exclusive rights to this article under a publishing agreement with
on skeletal muscle work efficiency, fuel utilization, and the author(s) or other rightsholder(s); author self-archiving of the
biochemistry in human subjects. Am. J. Physiol. 298, R79–R88 accepted manuscript version of this article is solely governed by the
(2010). terms of such publishing agreement and applicable law.
103. Kissileff, H. et al. Leptin reverses decline in satiation in weight-
reduced obese individuals. Am. J. Clin. Nutr. 95, 309–317 (2012). © Springer Nature Limited 2023

Nature Metabolism | Volume 5 | August 2023 | 1266–1274 1274