It is estimated that only about 20% of individuals who experience significant weight loss are able to maintain their lost weight [7
]. Data from participants in the Biggest Loser Television program demonstrate the difficulty in maintaining lost weight over time. The average weight loss of the 14 participants during the 30-week intervention was 58 kg, but six years later, the contestants had regained an average of 70% of their lost weight (41 kg) [25
]. The lack of successful weight loss maintenance for many dieting individuals could largely stem from behavioral issues—the inability of the individual to permanently adopt long-term lifestyle habits that promote a reduced body weight in the face of an obesogenic environment. The availability of a highly palatable, relatively inexpensive food supply coupled with a living environment that requires little physical work to carry out daily tasks could stand in the way of permanent changes in dietary and physical activity patterns. Lack of successful treatment then would stem from the inability to permanently alter behavioral responses to environmental conditioning and pressures. However, there is evidence that metabolic factors can contribute substantially to poor treatment outcomes—in response to weight loss, regulatory physiological responses are invoked that can effectively work to re-establish positive energy balance leading to weight regain toward a pre-established body weight set point [26
]. While the voluntary behavior (conscious choices) versus biological determinism (pre-programmed set-point with tight control) debate is of keen interest to researchers in the field, the two are not mutually exclusive—behavioral and metabolic factors are inextricably intertwined, and both pose significant obstacles to long-term weight loss maintenance.
3.2. How Does Weight Loss Affect Hunger and Satiety?
Many metabolic factors contribute to the energy gap following dietary restriction. Observed changes in adiposity-related signals (leptin and insulin), hypothalamic neuronal activity and neuropeptide expression, and gut peptide expression are thought to play a role in the increased hunger in response to weight loss. An exhaustive review of these factors is beyond the scope of this paper. The interested reader is referred to comprehensive reviews on this topic [14
The hypothalamus integrates many signals from the periphery, including liver, gut, and adipose tissue to regulate energy expenditure and the initiation, termination, and frequency of eating. Homeostatic regulation of food intake occurs such that severe caloric energy restriction leading to weight loss results in a strong internal drive to eat, whereas an overabundance of food intake and weight gain may be followed by a reduction in food intake. Insulin and leptin are putative players in this regard [31
]. They circulate in proportion to body fat mass and bind receptors in hypothalamic neurons, promoting expression of the anorexigenic peptides, pro-opiomelanocortin (POMC) and cocaine-amphetamine related transcript (CART), and inhibiting expression of the orexigens, neuropeptide Y (NPY) and Agouti-related peptide (AgRP). Acting through second-order neurons, these leptin- and insulin-stimulated neuropeptide changes result in reduced food intake and increased sympathetic nervous system activity and energy expenditure [31
]. Because increasing fat mass results in incremental increases in circulating leptin, the majority of obese children and adults present with hyperleptinemia [35
]. However, this elevation in blood leptin concentrations occurs without an ensuing decrease in food intake, indicating the presence of leptin resistance among individuals who exhibit obesity. While the higher circulating leptin and insulin coincident with increasing adiposity should theoretically limit weight gain, cellular resistance to these hormones occurs suggesting that protection against weight gain may be less robust than protection against weight loss [32
] has argued that human survival against risk of starvation (i.e., body energy stores insufficient for reproduction and life) and risk of predation (i.e., excess body mass impairs the ability to escape predators) is delimited by dual (upper and lower) intervention points and environmental and behavioral pressures exert the primary influences on body weight within the range between them. In accord with this view, for most individuals who exhibit body weight between the upper and lower intervention points, the availability of food would be a primary determinant of food intake. However, as weight falls outside this range, genetically-driven (and/or epigenetic) physiologic changes occur that promote restoration of body energy stores that support survival. Therefore, when inadequate food availability (whether due to intentional energy restriction, as with dieting, or involuntary severe energy deficit, as with famine) results in weight loss that reaches the lower intervention point, homeostatic metabolic changes are invoked that promote weight regain. Conversely, as body weight increases (in today’s society, this would be largely due to the obesogenic environment) and reaches the upper intervention point, homeostatic adjustments should theoretically come into play to decrease energy intake and increase energy expenditure, thus limiting weight gain and risk of predation. However, Speakman [36
] argues that owing to a substantial reduction in predatory risk within the human population today, genetic shifts have occurred such that there is a decidedly less robust defense against weight gain than weight loss. Weight gain that exceeds the upper intervention point produces, at best, only modest reductions in hunger and increases in energy expenditure, in part due to the aforementioned leptin resistance. On the other hand, when body weight falls below the lower intervention point, leptin and insulin rapidly decrease causing increased food intake and reduced energy expenditure. This is not unexpected given the decrease in fat mass. However, the magnitude of the decrement in circulating leptin is much greater than the magnitude of fat loss [39
], a phenomenon that may be one of the primary drivers of weight regain. During weight loss maintenance, leptin concentrations slightly increase relative to the dynamic weight loss state [35
]; however, these levels remain significantly reduced even when adjusted for changes in fat mass after one [39
] and two years of weight maintenance [41
A host of other anorexigenic peptides originate in the gut and typically increase in circulation in response to feeding, which then communicate with the hypothalamus to terminate food intake, increase satiety, and increase satiation between meals [42
]. These peptides include cholecystokinin, peptide YY, amylin, pancreatic polypeptide, and glucagon-like peptide-1 (GLP-1). There is increasing evidence of a sustained, long-term decrease in anorexigenic signals in response to diet-induced weight loss, with the decrement being greater than the decline in body weight [40
]. Such physiological changes could result in a metabolic milieu that readily promotes weight regain following weight loss.
Ghrelin is an orexigenic hormone, primarily produced by oxyntic cells of the stomach and is the endogenous ligand for the growth hormone secretagogues receptor type 1a (GHS-R1a) [45
]. The GHS-R1a is located throughout the body including the hypothalamus, pituitary, neuroendocrine tissues, pancreas, stomach, and vagus nerve [47
]. The presence of intact vagal afferents is essential for the centrally mediated effects of ghrelin on hunger and satiety. In lean individuals, plasma ghrelin concentrations rise during fasting and drop with meal ingestion proportional to the calorie content of the meals [48
]. Obese individuals may not display the same suppression of ghrelin in response to calorie ingestion [49
]. Weight loss leads to an elevation of plasma ghrelin in obese adolescents [51
] and adults [52
], which is thought to be a compensatory adjustment designed to increase energy intake in an attempt to return body fat stores to their initial levels. The available data suggest that the increase in circulating ghrelin that accompanies weight loss and persists into the weight maintenance phase could contribute to increased hunger and the energy gap.
Additionally, compared with normal-weight individuals with no history of obesity, individuals who are overweight or obese, but have lost weight, have a different neural response to overfeeding [53
]. In a randomized crossover study involving a two-day eucaloric feeding condition and a two-day 30% overfeeding condition, Cornier and colleagues [53
] used Functional Magnetic Resonance Imaging (fMRI) to compare the neuronal responses to viewing images of food among “thin” participants—normal-weight individuals (BMI 19–23 kg/m2
) with no history of obesity—and “reduced-obese” participants, who were overweight or obese (BMI 27–32 kg/m2
) but recently lost weight in a weight-loss program. The study was designed to analyze brain responses to food images in the overfed state versus eucaloric state. Among thin individuals, overfeeding attenuated neural activation compared to that observed during the eucaloric state. This response to overfeeding did not occur among reduced-weight overweight/obese individuals. In the baseline fasting state, thin individuals had a much more robust neuronal response to food-related visual cues than reduced-obese individuals. Overfeeding resulted in significant attenuation of the response to visual foods cues in thin but not reduced-obese individuals.
Much of the research on body weight/composition regulation has been adipocentric—that is the control of energy intake and expenditure has focused on adiposity-related signals as discussed above. However, recent studies have demonstrated that fat-free mass (FFM) and resting metabolic rate (RMR) are positively associated with energy intake [54
]. FFM is the strongest predictor of RMR and this relation suggests the possibility of a link between FFM-driven energy requirements and the homeostatic control of energy intake. In other words, the large amount of metabolically active lean tissue found in most obese individuals could provide signals to drive the high energy intake necessary to sustain the obese state. However, there is also evidence that the FFM depletion (along with loss of body fat) resulting from caloric restriction fails to dampen appetite, but instead contributes to hyperphagia. Dulloo et al. [56
] have suggested that the relation between FFM and energy intake is, in fact, U-shaped, such that the large FFM associated with obesity ‘passively’ drives high energy intake, but the decrement in FFM associated with diet-induced weight loss enhances the drive to eat. Further, they suggest that the reduced FFM that occurs with weight loss stimulates increased energy intake in order to restore the FFM, but also causes increased fat deposition, a phenomenon they describe as ‘collateral fattening’. If during weight regain following weight loss, the restoration of FFM lags behind fat restoration, hyperphagia could persist beyond the fat mass “catch up” and result in greater body fat storage than existed prior to dieting. Thus, weight loss-induced reductions in FFM could contribute to both aspects of the energy gap—a reduction in energy expenditure and increased hunger, both of which could contribute to weight regain.
Taken together, diet-induced weight loss appears to impact appetite via changes to both fat mass and FFM by altering known and unknown peripheral factors that communicate to the brain a state of nutrient deprivation. These changes result in increased hunger [58
], which is further exacerbated by increased food cravings in response to diet-induced weight loss [59
], especially from foods of higher energy density [59
], and a lower level of satiation in response to overfeeding when compared to never obese individuals [53
]. The above discussion is not to say that body weight and body composition are necessarily tightly controlled in all humans and that all weight regain is entirely a biologically-driven process. Indeed, the rapid population increase in the prevalence of obesity associated with the increasingly obesogenic environment would argue against exquisitely tight control, especially in regard to protection against weight gain.
3.3. How Does Weight Loss Affect Energy Expenditure?
Total daily energy expenditure (TDEE) is a function of four components: resting metabolic rate (RMR: the energy expenditure required for cellular processes necessary for life as measured when an individual is lying quietly and awake in a post-absorptive state), the thermic effect of food (TEF: the increase in energy expenditure above RMR in response to food ingestion), non-exercise activity thermogenesis (NEAT: energy expenditure above RMR required to support the activities of daily living as well as fidgeting), and exercise energy expenditure (ExEE: energy expended above RMR necessary for performing exercise). NEAT and ExEE together make up physical activity energy expenditure (PAEE). Diet-induced weight loss almost always causes significant decreases in TDEE, which can negatively impact the maintenance of lost weight. As far back as 1984, Leibel and Hirsch reported lower TDEE in post-obese compared to never-obese individuals—on the order of 25% lower than predicted by metabolic body size [60
]. In a later study, they determined that obese subjects undergoing 10–20% weight loss experienced a significant decrement in TDEE which could not be entirely explained by the loss of respiring body mass [61
]. They reported a mean reduction in TDEE of 8 kcal/kg fat-free mass per day in those obese subjects who lost at least 10% body weight. These decreases in TDEE reflect adaptive thermogenesis (AT)—the change in energy expenditure independent of changes in FFM and the composition of FFM. AT may persist long term [25
The TEF is reduced with dieting because there is a reduction in the total caloric load that requires obligatory digestion and absorption. A portion of the normal TEF results from increased sympathetic nervous system activity that accompanies food ingestion [62
], which also decreases with weight loss and lower quantities of food intake.
Numerous studies have shown that weight loss-induced declines in RMR contribute to the reduced TDEE. The reduction in RMR is partially the result of the loss of respiring body mass, but many studies [25
], but not all [66
], report the magnitude of the decline to be greater than can be explained by the reduction in respiring mass. In the case of diet-induced negative energy balance, a high AT characterized by a reduction in energy expenditure disproportionate to the reduction in FFM (increased resting energy efficiency) attenuates the weight loss drive. The AT associated decrement in RMR could be due to changes in FFM composition, decreases in sympathetic nervous system activity, and lower circulating tri-iodothyronine, leptin, and insulin. Note, however, the magnitude of AT and its contributors appear to vary significantly between individuals and also according to the phase of weight loss. Muller et al. [10
] point to the importance of characterizing differences in adaptive thermogenesis during on-going (active) weight loss versus fixed weight loss (maintenance). They suggest that the reduced insulin concentration and changes in the composition of FFM (reduced glycogen and intracellular water) are the primary drivers of AT in the RMR during the first several days of active weight loss (phase 1), but in the second phase as the velocity of ongoing weight loss slows there is little to no AT and the continuing decline in RMR is largely a function of decreased FFM.
There is substantial evidence of AT in regard to RMR with weight loss. For example, Leibel et al. found that a 10–20% weight loss in obese patients caused a decrease in RMR of 3–4 Kcal/kg fat-free mass per day [67
]. In the study of Biggest Losers participants, despite the mean weight regain of 41 kg during the six years following the competition, mean RMR of the participants remained 700 kcal/day lower compared to baseline and no different compared to the end of the intervention at 30 weeks [25
]. Notwithstanding the sizable inter-individual variation in the magnitude of RMR responses and use of different metabolic carts to measure RMR at six years compared to baseline and 30 weeks, these data provide evidence of significant long-term metabolic adaptation (increased metabolic efficiency) that accompanies weight loss. Even among athletes undergoing significant exercise training, weight loss may result in a decrement in RMR that is disproportionate to the loss of body mass [63
]. Note that during weight loss maintenance (fixed weight loss) AT is present, with TDEE lower than predicted based on metabolic body size, with this decrement due primarily to the non-resting component as explained below.
The effect of diet-induced weight loss on ExEE and NEAT is quite variable. From a thermodynamic perspective, without changes in movement efficiency and economy, the loss of body mass will result in fewer calories expended to perform the same weight bearing movements. When considering the energy expenditure incurred from activities of daily living at the lower body weight, this could conceivably contribute substantially to the reduction in total daily energy expenditure. Indeed, following a 10% weight loss, individuals were found to exhibit greater skeletal muscle work efficiency at lower intensities, which Rosenbaum et al. estimated could account for one-third of the reduction in PAEE [69
]. Others have also found weight loss to result in increased energy efficiency at low exercise workloads [70
], possibly resulting from the decrease in circulating leptin [70
]. There is substantial inter-individual variability regarding changes in energy expenditure and efficiency (i.e., adaptive thermogenesis) with weight loss, both of which are associated with the aforementioned reductions in circulating insulin and leptin. However, two recent studies suggest that changes in the magnitude of circulating leptin and insulin sensitivity alone are not sufficient to explain the magnitude of weight regain [72
To summarize the changes in energy expenditure, diet-induced weight loss can result in significant reductions in RMR, TEF, ExEE, and NEAT. These metabolic changes that result in lower energy expenditure would not obligatorily contribute to weight regain if there were a proportional decrease in food intake. However, as discussed earlier, body composition and hormonal changes occur with weight loss that are associated with increased rather than decreased appetite. This scenario coupled with the environmental pressures of modern society that favor excessive calorie consumption and minimize physical activity, could lead to a sense of futility among both patients and practitioners regarding lifestyle obesity treatment. However, as MacLean et al. [27
] suggest, the energy gap “should not be misconstrued into a conciliatory surrender to the inevitability of weight regain”. They instead indicate that the “biological drive to regain lost weight can be countered with environmental, behavioral, and pharmaceutical interventions”. However, they stop short of providing such information in their review.
Thus, the purpose of this second section is to examine lifestyle approaches to attenuate the increase in hunger and the decrease in energy expenditure that accompany weight loss and to identify important areas for further research.