In this review, we discuss the role of inactivity and exercise on appetite regulation, both in the short and long term, and the potential mechanisms involved. A better short-term appetite control has been described in active compared to sedentary men, and an exercise intervention was shown to improve appetite control in previously sedentary individuals. The mechanisms whereby exercise improves short-term appetite control remain obscure and although the changes in the postprandial release of satiety peptides are attractive hypotheses, it remains unproven. The impact of exercise on habitual food intake is also controversial and likely to be dependent on restraint level and body weight. We hypothesize that the beneficial impact of exercise on appetite regulation can contribute to its well-established efficacy in the prevention of weight regain in obese individuals. However, more studies are needed in the obese population to clearly establish the role of exercise on appetite control in this group.
The prevalence of obesity has reached epidemic proportions worldwide, with more than 1.6 billion adults overweight (body mass index >25 kg/m2) and at least 400 million clinically obese (body mass index >30 kg/m2).1 It is generally accepted that this steady increase in obesity prevalence has paralleled, on one hand, an increase in the consumption of energy-dense food and, on the other hand, a reduction in physical activity (PA) levels.2 However, somewhat surprisingly, the tendency, at a population level, seems to be to an overall reduction, as opposed to increase, in energy intake (EI) over the last two decades.3 It has been suggested that a larger reduction in PA levels, together with an inability of the organism to downregulate EI to a similar extent to match the reduced energy expenditure (EE), are the dominant factors in promoting obesity.4 This uncoupling between EI and EE is likely to result from a breakdown in the normal mechanisms regulating appetite and eating behaviour. It has been systematically shown that the adoption of a sedentary lifestyle inevitably produces a state of positive energy balance (EB), as the physiological system is unable, at least in the short to medium term, to compensate adequately by decreasing EI.5, 6 Although it is undeniable that in most populations, the quantity and intensity of PA has decreased over the past few decades,3 and that only a small proportion of the population in industrialized countries follow an active lifestyle,7, 8 physical inactivity alone is unlikely to explain the obesity epidemic and diet is, without doubt, also involved in the aetiology of obesity.9, 10
Appetite control—an overview
Eating behaviour in general and appetite in particular, as a subjective concept used to explain it, are complex phenomena that can be conceptualized by the selection of particular food items and the size and frequency of eating episodes, which together determine not only the total energy but also macronutrient intake.11 Although both physiological and environmental factors contribute to appetite and eating behaviour, it is widely accepted that strong social and environmental influences can easily overcome normal physiology.9
At the physiological level, it is well established that the hypothalamic region of the brain plays a key role in the central regulation of eating behaviour in humans. The hypothalamus, in particular the arcuate nucleus, is constantly receiving and processing neural, metabolic and endocrine signals from the periphery, enabling it to maintain energy homoeostasis by adjusting not only EI but also EE. Even though most of the peripheral signals are generated by the gastrointestinal tract, other organs are also actively involved such as the pancreas and the adipose tissue.12
Two types of peripheral signals can be described: firstly, episodic signals, which are periodically released, mainly from the gastrointestinal tract, in response to feeding or fasting, therefore signalling acute nutritional state; secondly, tonic signals, which are constantly released, mainly by the adipose tissue, in proportion to the amount of lipid stores, therefore signalling chronic nutritional state.13 To the first category of peripheral signals belongs ghrelin, an orexigenic hormone released in response to fasting, and several satiety hormones such as glucagon-like peptide-1 (GLP-1), peptide YY (PYY) and pancreatic polypeptide (PP), which are released in response to feeding. The actions of these hormones on eating behaviour can either be achieved directly, by modulating gastric empting, or indirectly, by modulating the activity of orexigenic and anorexigenic neuropeptides within the arcuate nucleus.14, 15 The second category of peripheral signals includes leptin, released mainly by the adipose tissue, and insulin, released by the pancreas, whose secretion is directly proportional to the amount of body fat. Leptin has been shown to reduce food intake and body weight and increase EE.14 Leptin and insulin play an active role in body weight regulation not only directly by acting as adiposity signals, stimulating anorexigenic neurons expressing propiomelanocortin and inhibiting orexigenic neurons expressing neuropeptide Y and agouti-related peptide, but also indirectly by modulating the sensitivity of the brain to satiety signals and, in doing so, determining EI at an individual meal.16 For a more detailed description of the physiology of appetite, see the review by Wynne et al.17
In addition to the physiological control of appetite, the presence of external stimuli arising from food and the surrounding environment play a key role in food intake. Environmental, psychological, social, and cultural stimuli have been shown to exert powerful effects on food selection.18 In this context, two cognitive variables have been shown to exert an important role on appetite and food intake, restraint and disinhibition. Dietary restraint is the extent to which individuals are concerned with their body weight and attempt to control it by restrictive eating behaviours, and disinhibition is the tendency towards overeating motivated by the external or emotional stimuli.19, 20 Moreover, food intake is also under hedonic control,21 activated by the sensory pleasure of eating palatable food.22 The above stimuli parallel to the homoeostatic system previously described, in which tonic and episodic signals provide a dynamic ‘picture’ of the state of energy stores and the periodic flux of nutrients derived from feeding.
Exercise and appetite control
In the face of the current obesity epidemic and the widespread levels of physical inactivity, it is extremely important to determine how exercise (both in the short and long term) impacts on different aspects of appetite, from motivation to eat and food intake to, in a broader perspective, EB and body weight. This will allow a better understanding of the aetiology and/or potential treatments of obesity and constitutes the scope of this review.
The impact of exercise on motivation to eat and food intake
There have been a multitude of studies published in the past two decades exploring the association between exercise and food intake. The majority of them have shown that acute exercise does not increase hunger or EI,23, 24, 25, 26, 27, 28 even at high intensities,29 and that exercise is therefore able to induce a short-term negative EB. Somewhat surprisingly, vigorous exercise (high-intensity cycling or running) has been found to significantly reduce hunger, a phenomenon that became known as ‘exercise-induced anorexia’,25, 30 although this is temporary and unlikely to have any significant impact on subsequent EI.29, 30, 31 However, there is some controversy regarding the effects of acute exercise on objective and subjective measures of appetite, with a few studies showing an increase in hunger32 and subsequent EI33, 34, 35 in response to acute exercise or even a decrease in EI.26 Methodological differences among these studies, particularly concerning exercise intensity,23 nutritional state,36 gender,37 macronutrient composition of the test meal38 and the time interval between exercise and eating33 are likely to be involved in these inconsistencies.
Motivation to eat and food intake in response to acute exercise seem to be modulated by gender, body weight and eating behaviour, among other factors. In general, acute exercise has no effect on subsequent EI in men, whereas in women an increase in EI is usually observed, either decreasing or abolishing the effects of exercise on EB.23, 34, 37 Unrestrained normal weight women, unlike men, report an increased palatability of foods with exercise and do not experience the transient suppression of hunger observed immediately after exercise.24, 25 However, exercise-induced anorexia has been reported in restrained normal weight women27 and in obese women (restraint level not reported).39 Obesity can also effect the response to exercise and while it has been shown that normal weight individuals (mainly women) increase their EI in response to a 3-day exercise intervention, compared with a similar period of rest, obese subjects fail to show such a compensatory response.36 Similar results were reported after a moderate or intense bout of exercise, with only normal weight women showing a compensatory increase in EI, regardless of the exercise intensity.36, 40 However, in a study involving male volunteers, a temporary reduction in hunger and a decrease in EI was reported after moderate exercise in both obese and non-obese individuals,26 suggesting a clear interaction between body weight and gender in determining postexercise appetite.
Hill et al.41 proposed a model, more than 10 years ago, whereby body weight and eating behaviour would modulate the coupling between EI and EE. This model proposes that individuals with particular eating behaviours—unrestrained, restrained and disinhibited—would be either more or less responsive to physiologic cues of hunger and fullness and, therefore, present different patterns of energy compensation in response to increased PA. Unrestrained eaters, with a normal body weight, would be expected to fully compensate for increased PA. According to Hill et al.,41 unrestrained lean individuals would respond to increased PA by increasing their EI, whereas unrestrained overweight individuals would only partially compensate in terms of EI, therefore resulting in a negative EB. In restrained eaters, EI in response to exercise would be dependent on their levels of disinhibition. Restrained eaters with low levels of disinhibition (‘successful dieters’) would self-impose a tight cognitive control over their EI and, therefore would not increase their EI in response to increased PA, independently of their body weight. However, restrained eaters with high levels of disinhibition (‘unsuccessful dieters’) would respond to exercise depending on their current dieting status. During a period of restrained eating, physiological appetite cues would be ignored and EI maintained or even decreased, while in a period of uncontrolled eating brought on by disinhibition, an increase in EI would be observed.41 In fact, overweight women with high levels of disinhibition were shown to increase their EI in response to 2 weeks of exercise (5 days a week).42 However, the evidence is conflicting with at least two acute studies showing that exercise may have positive effects on appetite control in normal weight women characterized by high levels of disinhibition. In these studies, normal weight women, scoring high on disinhibition, reported an increase in the preference for low-fat foods43 and a reduction in motivation to eat after an acute bout of exercise, despite exhibiting no significant differences in food intake at a test meal (compared with a control condition).44 These findings are important to the obese population as their eating behaviour interacts with their response to exercise. Overall, exercise has been shown to be more effective in creating a negative EB in restrained compared with unrestrained eaters,45 suggesting, according to King,46 that exercise may act as a controlling mechanism over eating in restrained individuals, and not as a disinhibitor as initially thought. However, the interaction between different traits of eating behaviour and exercise in determining postexercise EI is complex and little is currently known in this area. An important area for future research will be to assess the ability of chronic exercise to modulate eating behaviour, particularly in overweight/obese individuals, and how changes in disinhibition and/or restraint scores in response to exercise can predict weight loss in the long term.
There are methodological limitations within the literature to date and the inability of some studies to show a beneficial effect of acute exercise on EB may derive from the fact that they do not account for the energy cost associated with the exercise itself. To overcome this limitation, the concept of relative EI (REI) has been created.29 In fact, a short-term negative EB in response to acute exercise can be achieved even in the presence of a compensatory increase in absolute EI, if the energy expended during exercise is greater than the increase in EI (in comparison with a control condition). The evidence suggests that while in the short- to medium-term (up to 16 days), exercise is able to produce a negative EB, with no substantial compensatory responses in EI being observed, in the long term (more than 16 days), an increase in EI is likely to be observed. However, compensation is usually partial and incomplete, generally accounting for only 30% of the energy cost associated with exercise, therefore allowing the attainment of a negative EB and some degree of weight loss.47 In the short term, exercise may, in fact, be more effective than dieting in producing a negative EB. This is evidenced by the finding that while an acute energy deficit created by dietary restriction (low-energy versus high-energy breakfast) induces a significant increase in subjective hunger, subsequent EI at test meal and food cravings during the day, a similar energy deficit created by exercise failed to induce any significant change on these variables, thereby allowing the attainment of a short-term negative EB.48
In the long term, very few studies have looked at the impact of chronic exercise, as the sole intervention, on motivation to eat. In normal weight individuals (both men and women), a 7-day exercise programme (80 or 120 min per day), with ad libitum food intake, did not induce any significant changes in subjective feelings of hunger or fullness (assessed hourly during waking hours) compared with a control condition.49, 50 A more recent study in normal and overweight male adolescents also failed to show any significant changes in the sensations of hunger, desire to eat or prospective food consumption, either in fasting or postprandially, after 5 days of supervised aerobic exercise, despite a tendency towards a decrease in fullness (P=0.055), in both the normal weight and overweight groups.51 Another recently published study consisting of a 6-week intervention of fixed, reduced dietary intake and 6 h day−1 of skill-based PA in obese children resident in a weight loss camp showed a significant increase in diurnal profiles of hunger and a reduction in fullness, together with a significantly lower suppression of hunger in response to a test meal at the end of the intervention, parallel to weight loss.52 However, it needs to be recognized that this intervention resulted in a substantial negative EB, given the low baseline levels of PA in these children and the dietary restriction that they underwent. Although it is possible that the long-term exercise does not elicit any increase in motivation to eat until body weight starts to decrease, many questions remain unanswered: does chronic exercise, in the absence of dietary restriction, has a positive or negative impact on subjective measures of appetite? Is that impact dependent or independent of weight loss? Is the impact different in normal weight and overweight/obese individuals? Can changes in subjective feelings of hunger and fullness in response to exercise predict changes in body weight?
The impact of exercise on weight loss/prevention of weight regain—is there a role for appetite control?
Even though exercise is able to produce a negative EB in the short to medium term, there is good evidence that, in the absence of energy restriction, long-term exercise only results in modest weight loss.53, 54, 55 This ‘relative’ inefficacy of exercise in inducing weight loss does not mean, nevertheless, that exercise confers no additional benefits in obesity treatment. Regular exercise, without weight loss or energy restriction, is associated with a substantial reduction in total and visceral fat and in skeletal muscle lipid, together with a significant improvement in fitness levels in obese individuals.56 Body weight, in isolation, may, therefore, not be the best outcome measure to evaluate the health benefits of exercise interventions. Moreover, although dieting alone is usually seen as a more effective way of losing weight in the short term, compared with exercise alone, its efficacy in the long term is questionable. In a recently published meta-analysis on the long-term outcomes of energy restriction on weight loss, it was reported that, despite an initial 5–10% weight loss in the first 6 months of dieting, most of the participants regained all that weight in the long term, and one- to two-thirds of the subjects regained more weight than they had originally lost within 4–5 years.57
Several physiological/metabolic mechanisms may account for the relative lack of efficacy of exercise as a weight loss strategy.46, 55 Firstly, exercise induced changes in body composition with loss of fat mass but preservation of fat-free mass,58 of greater density, underestimates the net impact of exercise on weight loss when compared with diet-only interventions associated with the loss of fat and fat-free mass.59 Secondly, exercise is associated with a reduction in energy expended in basal metabolic rate as a result of weight loss.60 Thirdly, an increase in maximal oxygen uptake occurs with exercise training, with less energy being expended for the same volume of exercise, and finally, a reduction in net exercise-induced EE occurs, especially for weight bearing exercise,46 also as a result of weight loss. However, behavioural mechanisms are probably more important in preventing the likelihood of weight loss in response to exercise.61 As previously discussed, the relative ineffectiveness of exercise on weight loss is likely to originate from the energy deficit created by exercise being partially compensated by an increase in EI.28 Moreover, a compensatory reduction in EE, probably driven by a downregulation in spontaneous PA (non-exercise PA), or a combination of both behavioural mechanisms, may also play a role.41, 62 In summary, both behavioural and metabolic compensatory mechanisms are activated in response to exercise-induced energy deficits in an attempt to defend the body against a negative EB. The extent and degree of these compensatory mechanisms are likely to vary among individuals and explain their resistance or susceptibility to exercise-induced weight loss. For an update on compensatory responses to exercise see the review by King et al.61
Even though exercise alone is usually seen as an ineffective strategy to lose weight, a great variability of outcomes, from weight loss to weight gain, has been found in published studies.63 Methodological differences among studies concerning the macronutrient composition of the diet and the characteristics of the exercise intervention (type, duration, frequency and intensity) may contribute to some of the variation.63 It can be suggested that individual differences in the coupling between EI and EE are likely to be more important than the previously discussed factors in explaining the large interindividual variation in weight loss in response to exercise.4 A major problem is that most of the studies looking at the impact of chronic exercise on EB do not control for, or accurately measure EI or EE. In free-living conditions, most individuals upregulate their EI in response to exercise, even if instructed otherwise, and often compliance with the exercise programme is poor.59, 64, 65 Moreover, it cannot be assumed that every exercise intervention results in increased total EE, unless it can be ensured that normal activity throughout the rest of the day remains unchanged or increases; what has been shown in some,63 but not all studies.66, 67
In a recent study by King et al.68 where EE and EI were carefully monitored, it was demonstrated, for the first time, that the large interindividual variability in weight change (−14.7 to +1.7 kg) following a supervised exercise intervention (12 weeks) was, in fact, due to the differences in behavioural compensatory responses. When they compared those individuals who lost less weight than expected (1.5±2.5 kg) with those who lost more weight than predicted (6.3±3.2 kg), the only significant difference was on EI; while the first group compensated for the increased EE induced by exercise by increasing their EI, the second reduced it. Some important questions remain unanswered. Why do some individuals increase their habitual EI in response to exercise and others do not? Are the mechanisms involved in such differences physiological, psychological or a combination of both? A better understanding of the mechanisms behind such differences would allow more individually tailored and appropriate strategies for weight management programmes to treat obesity effectively.
It needs to be emphasized, nevertheless, that substantial weight loss can be achieved with exercise alone, if EI remains unchanged over time and compliance with exercise is good.60, 69 However, the volume of exercise needed to attain this weight loss is usually very high (between 1 to 2 h per day in 6 to 7 days of the week) representing a huge challenge and commitment, particularly in the obese population. Interestingly, significant reductions in body weight can also be achieved under ad libitum food intake if the compliance with the exercise programme is good, as shown in a military environment.
Despite the previously described relative inefficacy of exercise alone in achieving weight loss, researchers have consistently reported a critical role for exercise in obesity prevention70 and in the long-term maintenance of weight loss,53, 55, 71, 72 reinforcing the need to incorporate exercise into every strategy intended to tackle the problem of obesity in the longer term. Large prospective studies show an inverse longitudinal relationship between changes in fitness and changes in body weight73, 74 and an improvement in fitness levels has been shown to help to attenuate age-related weight gain.73 Although cohort studies tend to show that individuals with high levels of PA are less likely to gain weight, the evidence regarding the predictive effect of baseline PA on subsequent weight gain is inconsistent, probably due to confounding factors, reverse causality and measurement error.70
In terms of long-term maintenance of weight loss, the role of exercise appears stronger. The US National Weight Control Registry, published in 2008, reports that those who are successful at maintaining weight loss (individuals maintaining a 13.6 kg weight loss for more than 1 year) are an extremely physically active group, despite a large variance in individual levels of PA.72 The reasons for this association between high levels of PA and successful maintenance of weight loss in the long term are not fully understood. Catenacci and Wyatt55 have recently proposed three hypotheses for this association. Firstly, exercise might increase resting metabolic rate (RMR), compensating for the expected reduction in RMR usually observed during weight loss, which would help in weight maintenance by increasing total EE. Secondly, high levels of PA are associated with a better coupling between EI and EE, therefore, facilitating the maintenance of EB. Thirdly, high levels of PA are associated with better adherence to energy-restricted diets. In agreement with Catenacci and Wyatt's (2007) second hypothesis, we propose that the critical role of exercise in weight loss maintenance may be, in fact, attributable to an improved appetite regulation. We suggest that: (1) not only does exercise improves the coupling between EI and EE in the long term, so that changes in EE are quickly followed by proportional changes in EI in the opposite direction, thus ensuring the maintenance of a constant body weight over time, but (2) also leads to a better energy compensation, so that the changes in EI at a particular meal are detected and compensated for at the following meal, leading to an improved appetite control in the short term. Although there is evidence supporting these mechanisms, the available data are confined to the normal weight population (see below) and, therefore, studies are urgently needed in obese individuals for these hypotheses to be tested.
The impact of exercise on the coupling between EI/EE and energy compensation
The relationship between EI and EE at different levels of PA was studied as early as 1956 by Mayer et al.75 in a male population of mill workers in West Bengal (India). They were able to show a good correlation between EI and energy requirements, with an increase in EE being followed by a proportional increase in EI, but only for moderate-to-high levels of PA. In sedentary individuals, however, EI and EE were uncoupled, so that a reduction in EE was followed by an increase (instead of a reduction) in EI, leading to positive EB and weight gain.75 The coupling between EI and EE, and in a broader sense, appetite control in general, seems, therefore, to be disrupted at low levels of PA. This would help to explain why it is so hard to prevent weight regain, after a weight loss intervention, in sedentary individuals, due to the difficulty in reducing EI to match the low levels of EE. On the other hand, a tight coupling between EI and EE has been described at high levels of PA,75, 76 suggesting that exercise may, in fact, ‘fine tune’ the physiological mechanisms that regulate appetite.
Evidence that exercise improves energy compensation in response to covert preload energy manipulation also exists. Active men are able to adequately compensate in response to covertly manipulated drinks, following a bout of exercise, by adjusting EI at a subsequent meal, denoting a good short-term appetite control.77 A later study published in 2002 by Long et al.78 strengthened the support for a beneficial role of exercise on appetite control by showing a better compensatory response to covert preload energy manipulation in active compared to sedentary men.78 Healthy normal weight men were randomly assigned a low- or high-energy preload on 2 different days and energy compensation assessed as the ability to distinguish between preloads by adequately adjusting EI at an ad libitum buffet lunch presented 1 h later. Sedentary men showed a deficient homoeostatic feedback control of hunger and satiety; they were unable to distinguish between the two preloads and had a similar EI at the buffet lunch on both occasions. In contrast, active men showed very good energy compensation, eating significantly less at the buffet meal after the high-energy preload compared to after the low-energy preload. Similar results were described in a recently published study, with a more accurate energy compensation in active versus sedentary individuals, but this time over the course of a day instead of acutely, at a test meal.79 Van Walleghen et al.79 measured energy compensation, in normal weight unrestrained young and older adults (both genders), using a no-preload versus preload condition and measuring EI 30 min later at a buffet meal and for the rest of the day. They reported a significantly better energy compensation (%), over the course of the day, in active compared to sedentary individuals, independently of the age group. However, the above studies are cross-sectional in design. They provide indirect evidence, but not causal evidence, for the beneficial role of exercise in appetite regulation, as the observed effects may be due to lifestyle or other factors in the two groups, which were unrelated to their PA levels. Intervention studies where sedentary people engage in an exercise programme, shifting them into being ‘active’, are needed to clearly examine the effect of exercise on appetite control.
To overcome this limitation, we have recently evaluated the long-term effects of exercise on energy compensation in response to covertly manipulated preloads, using a longitudinal design.80 We used a methodology similar to the one previously described by Long et al.78; the ‘preload-test meal’ paradigm, to measure short-term appetite control at baseline and after a 6-week moderate intensity exercise programme in unrestrained normal weight sedentary individuals (both genders). Participants were given a high- or low-energy preload, on two separate occasions, and EI at an ad libitum buffet meal was measured 1 h later. Participants were also asked to record in a food diary all they consumed after the buffet lunch until (and including) breakfast of the following day to estimate 24 h cumulative EI. The exercise intervention resulted in improved appetite regulation, with a more sensitive eating behaviour in response to previous EI, not only acutely, at lunchtime, but also for the next 24 h. Before the start of the exercise programme, participants were unable to adjust buffet EI in response to preload energy manipulation; however, after the exercise intervention, they adequately compensated by eating significantly more at the buffet test meal after the low-energy preload compared to after the high-energy preload.80 These findings expand and confirm previous cross-sectional data78, 79 and propose a role for exercise on both sides of the EB equation, not only by increasing EE, but also by adjusting food intake accordingly in response to previous EI. However, all the previous investigations were performed in normal weight individuals and, studies are now needed in the overweight/obese population to clearly establish the role of exercise on appetite control in this group.
It was proposed, more than a decade ago, that changes in appetite regulation may result from modifications at three different levels: long-term signals such as leptin and insulin; intermediate postabsorptive signals associated with macronutrient oxidation such as glucose and free-fatty acids levels and short-term satiety signals released by the gastrointestinal tract in response to feeding, such as cholecystokinin, GLP-1 and PYY.11, 14 These appetite regulatory factors may be modified by exercise and this may help to explain the observed improvement in short-term appetite control following an exercise intervention. Surprisingly, we found no significant changes in fasting insulin, glucose, triacylglycerol or non esterified fatty acids plasma levels, or in insulin sensitivity, following a 6-week exercise intervention in previously sedentary volunteers80 and leptin levels, although not measured, were unlikely to have changed, as no significant changes in body weight or composition were observed over time. Short-term satiety signals arising from the gastrointestinal tract in response to food intake are, therefore, the most likely to have been targeted by exercise and to account for the improvement in appetite regulation observed with increased levels of PA.
The impact of exercise on the plasma levels of appetite-related hormones
Investigation of the impact of both, acute and chronic exercise, on circulating levels of the hormones and metabolites involved in appetite regulation is an area that has received considerable attention over the last decade, with a special emphasis on leptin and ghrelin. A recent review on leptin and exercise concluded that, in the absence of weight loss, no significant change in the fasting plasma levels of this hormone is observed in response to exercise.81 However, a significant reduction in plasma leptin levels was reported after an intense, prolonged exercise period (25 km swimming race) in highly trained individuals,82 as well as, after a 1-year exercise intervention, even after adjusting for body mass index and fat mass.83 However, in the first study,82 a huge negative EB was imposed and the subjects were not in a fasting state, therefore allowing the release of insulin, which is known to exert a negative feedback over leptin secretion. It has been proposed that an exercise intervention that causes improvements in insulin sensitivity would have the ability to alter leptin levels independently of any changes in fat mass, due to changes in insulin and cortisol levels, which are known to modulate leptin synthesis.84 Although acute exercise, when performed in the fed state, does not seem to impact on leptin plasma levels in normal weight individuals,85 extreme exercise (until exhaustion) in healthy normal weight men was reported to significantly increase plasma leptin concentrations.86 This increase is likely to originate, according to the authors, from the stomach rather than from the adipocytes.
In a similar way to leptin, a recent review paper by Kraemer and Castracane87 also concluded that in the absence of weight loss exercise has no significant impact on the fasting plasma levels of ghrelin. Although acute exercise, specifically, has been shown not to induce significant changes in the fasting plasma levels of ghrelin, in either normal weight85 or overweight volunteers,88 the results may be dependent on the intensity of exercise, with low- rather than high-intensity exercise stimulating total plasma ghrelin levels independently of its duration.89 Surprisingly, even though only the acylated form of ghrelin—acylated ghrelin (AG), exhibits orexigenic properties and is, therefore, involved in appetite regulation,90 studies looking at the impact of exercise on ghrelin have measured only total ghrelin plasma levels.85, 88, 89, 91, 92, 93 To overcome this limitation, a recent study looked at the impact of 1 h of running followed by 8 h of rest (exercise condition), performed in the fasting state (or 9 h of rest—control condition) on the plasma levels of AG for a period of 9 h (with a meal being presented 3 h after the start of each trial) in normal weight trained men. The authors described a significant suppression in AG in the exercise compared with the control condition, both on the first 3 h and on the full 9 h of the study, as well as a reduction in subjective hunger during the first 3 h,94 suggesting that AG responds differently from total ghrelin to acute exercise.
Although desacyl ghrelin (DG) has been traditionally seen as devoid of any endocrine activities, a study published in 2005 challenges this by reporting that DG induces a negative EB by reducing food intake and delaying gastric emptying in an inverse manner to AG.95 Mackelvie et al.51 measured the impact of 5 days of supervised aerobic exercise on total ghrelin, AG and DG plasma levels, both in fasting and postprandially, in normal weight and overweight male adolescents. They reported no change in total ghrelin plasma levels, but a significant increase in AG, which were independent of changes in body weight or insulin sensitivity. Moreover, the increase in AG was higher in the normal weight group, both in fasting and postprandially, and was associated with an increase in hunger sensations and a reduction in fullness 30 min after a meal. Exercise was also associated with a decrease in postprandial DG in normal weight, but with an increase in overweight adolescents. Although this study may suggest a limited effectiveness of exercise in inducing a negative EB, particularly in the normal weight group, it is limited by its short duration and the fact that EI was not controlled. A recent paper examining the impact of a 12-week supervised exercise programme (without energy restriction) on total ghrelin, AG and DG in overweight children, reported no change in AG plasma levels and a significant increase in total and DG plasma levels throughout the 12-week exercise intervention. Even more interesting, the reduction in body weight and body fat observed was strongly associated with the increase in DG plasma levels, suggesting a favourable effect of exercise on energy metabolism in overweight/obese individuals.96 Overall, the available evidence seems to suggest that exercise, in the absence of weight loss, has no significant impact on the fasting plasma levels of total ghrelin. The impact of exercise on AG and DG plasma levels is likely to be time-dependent; as the duration of exercise increases, its impact on EB becomes more favourable by reducing AG/DG ratio, therefore promoting weight loss and helping in weight maintenance.
In contrast to the large number of studies investigating the impact of exercise on plasma leptin and ghrelin levels, only a few have examined the effects of exercise on the plasma levels of satiety hormones involved in appetite regulation and almost all have been performed in normal weight individuals. These studies are summarized in Table 1. The few available studies show an increase in the fasting plasma levels of GLP-197, 98 and fasting and postprandial levels of cholecystokinin86, 99 in response to acute exercise, although both studies concerning GLP-1 had been performed in athletes. Increases in PP plasma levels have also been reported after acute exercise, not only in fasting86, 97, 100, 101 but also postprandially,86, 102 although these are likely to be dependent on the intensity of exercise.103 Even less evidence is available concerning chronic exercise; with two single studies only, one showing no change in fasting cholecystokinin plasma levels in active men99 and the other reporting a slight increase in both fasting and peak postprandial PP plasma levels, in previously sedentary men,104 after an exercise intervention. A recent study, where normal weight and overweight male adolescents underwent a 5-day supervised aerobic exercise intervention (1 h day−1), showed no significant change in fasting GLP-1 plasma levels or in the integrated plasma response (for a 4-h period) following a test meal, but a significant increase in the magnitude of the GLP-1 response in the first 30 min postprandially, in both weight groups.105 However, the majority of the studies previously described are limited by the fact that only fasting plasma levels have been measured. Cholecystokinin, GLP-1 and PP are released in the postprandial state and, therefore, changes in plasma levels in the fasting state are unlikely to be physiologically relevant in satiety regulation. Moreover, in the investigations previously described, appetite was rarely a primary outcome and very few have tried to establish associations between changes in the plasma levels of satiety gut peptides, in response to both acute and chronic exercise, and alterations in subjective and objective measures of appetite and net EB.105
To overcome these limitations, we designed a study to investigate the effects of acute exercise on postprandial plasma levels of appetite-related hormones and metabolites and, simultaneously, on motivation to eat and prospective food intake at a subsequent meal. Total ghrelin, PYY, GLP-1 and PP were measured for a period of 3 h in healthy unrestrained, normal weight volunteers, using a randomized crossover design. Participants were given a 500 kcal breakfast and 1 h later, cycled for 60 min, at 65% of their maximal heart rate or rested. During the study period, subjective hunger and fullness were continuously measured using visual analogue scales and at the end (3 h postbreakfast; 1 h postexercise), participants were presented with a buffet test meal and EI assessed.106 Acute exercise significantly increased postprandial plasma levels of PYY, GLP-1 and PP, but had no impact on ghrelin, suggesting that exercise can trigger physiological changes in hormone secretion, which could help in appetite control. Moreover, a significant increase in absolute EI at the buffet meal was observed with acute exercise; despite a significant reduction in REI, once the energy expended during exercise had been accounted for. Somewhat unexpectedly, this occurred in the absence of any significant differences in hunger sensations, or circulating appetite-related hormones, between the exercise and the control leg, immediately before the buffet meal was presented.106 It can be hypothesized that the increase in absolute EI observed in response to acute exercise was the result of cognitive factors such as the belief that exercise increases hunger or the common behaviour of using food as reward for exercising.46
The transitory increase in the plasma levels of satiety hormones during acute exercise, reported in the previous study,106 may, in fact, help to explain the short-term suppression of hunger (exercise-induced anorexia) observed during the same period. Although we did not measure AG, one study reported a positive correlation between integrated values for plasma AG and hunger during the 3 h of an exercise trial, consisting of 1 h running plus 2 h rest.107 These findings provide support for the hypothesis that changes in the plasma levels of appetite-related hormones following exercise underpin changes in the subjective measures of appetite (motivation to eat) during the same period. However, a recent paper has also proposed that the acute appetite-suppressive effects of exercise might be mediated at the hypothalamic level. Acute exercise in rats has been shown to increase the phosphorylation activity of several proteins involved in leptin and insulin signal transduction in the hypothalamus, leading to increased sensitivity to these two hormones.108
The above studies provide us a plausible mechanistic approach with the potential to establish associations between changes in the plasma levels of appetite-related hormones, subjective feelings of appetite and objective measurements of food intake. Although one can argue that the invasive nature of such studies may contaminate the measurement of appetite and food intake, data from our laboratory have demonstrated that this is not the case.109, 110
Overall, the impact of exercise on the plasma levels of these hormones seems to be favourable in terms of appetite regulation (see Table 1) and may, therefore, contribute to the improved appetite control observed in response to exercise and the differences described between sedentary and physically active individuals.78, 79, 80 What is currently missing are studies designed to investigate the impact of long-term exercise on appetite-related hormones and how any changes observed relate to weight loss/prevention of weight regain in response to exercise in the obese population. The interindividual variability in weight loss following exercise has already been shown to be largely dependent on compensatory changes in habitual EI.68 However, it remains unknown whether the compensatory increase in EI seen in those less responsive to exercise is physiologically or psychologically driven (or both). A synthesis of the available evidence to date,96, 105 suggests that a hypothesis worthy of future testing, would be that if a sedentary person engages in activity of sufficient duration and intensity to be considered active', then beneficial effects would be measurable in their appetite regulatory system. These beneficial effects would include a neutral or negative effect (reduction) on hunger signals (with no changes in total ghrelin or an increase driven by DG in opposition to AG) and a positive effect (increase) on satiety signals (such as GLP-1 and PYY) in the postprandial period, therefore offering a valuable strategy in weight loss management.
There is a large body of evidence supporting a beneficial role of exercise on appetite regulation. Exercise has been shown to lead to a more sensitive eating behaviour in response to previous EI and not to induce any acute/chronic physiological adaptations that would lead to an increase in hunger and/or EI in the short term. However, most of these studies have been performed in normal weight individuals and are short term. More long-term studies are needed in the obese population, to clearly establish the role of chronic exercise on appetite control in this group. If positive effects are revealed, this will reinforce the critical role of exercise in body weight management in obesity.
The evidence to date reinforces the need to increase PA levels, particularly in the face of the high prevalence of obesity and apparent failure of public health messages to attenuate its increase worldwide. More research is necessary to clarify the mechanisms behind the improvement in short-term appetite control observed with exercise, and their long-term implications in terms of EB and weight change.
C Martins was supported by a PhD grant (SFRD/BD/16294/2004) from Fundação para a Ciência e Tecnologia (Portugal) under the third European Union community support programme. None of the authors have any conflict of interest.