Exercise and appetite: when hunger becomes compensatory eating

Does exercise make you eat back the calories you burn? Research separates short-term hunger, longer-term food intake and snacks chosen as a reward.

日本語で読む（Japanese version）

You finish exercising and head toward the station. A sweet snack in a shop window catches your eye. You worked hard today, and dinner still feels far away. Is the urge to eat a sign that your body needs food, or does the snack feel like something you have earned? Both explanations can travel home in the same shopping bag. By the time you sit down to eat, it may be difficult to tell them apart.

If exercise always made us eat back everything we burned, moving more for weight loss could seem pointless. Yet the overall research does not support the claim that a single exercise session inevitably increases appetite or food intake that day. There is also a longer-term trial in which participants ate more and lost less weight than expected. Keeping that unfavorable finding in view matters just as much as keeping the reassuring one. [1][4][10]

The apparent disagreement becomes clearer when we ask what was measured, and when. Hunger immediately after a workout, the amount eaten at dinner, and average food intake over an exercise program are different outcomes. This article follows those distinctions: where compensatory eating has been observed, where it should not be assumed, and how a reward snack can become part of the picture without being the same thing as physical hunger. [2][4][10]

Feeling hungry and eating more are different outcomes

Compensatory eating means increasing food intake after exercise to make up for energy spent during it. The response varies: some people eat back that energy, while others do not. Researchers may examine total food intake, but they also use a measure called relative energy intake. This subtracts exercise energy expenditure from the energy consumed. It addresses a different question from simply asking how much food was eaten. [12]

Think of the distinction as a way to keep the accounting clear. A meal can contain the same amount of energy after exercise as after rest, while relative intake differs because the exercise session used energy. Conversely, a change in the amount eaten needs to be considered alongside the expenditure being compensated for. Neither measure can be replaced by the statement, “I felt hungry.” The sensation is relevant, but it is not itself a measurement of food intake. [12]

Appetite is usually something participants report using a rating scale. Food intake concerns what they actually consume. Appetite-related hormones provide another set of measurements, and those changes do not necessarily line up with either the subjective ratings or the meal. Research on exercise and eating needs these outcomes to remain separate. A hormone result may help researchers investigate a mechanism without providing a complete description of what happens at the table. [9][20]

A comparison study involving 47 healthy men and women illustrates the problem. Participants included people of normal weight and people with overweight or obesity. They completed an exercise condition involving 60 minutes on a treadmill and a condition involving rest. The trials lasted 8 hours. Exercise temporarily suppressed subjective appetite, but energy intake at a meal where participants could eat freely did not differ significantly between conditions. [9]

There is no contradiction in finding a difference on an appetite scale and no significant difference in food intake. The study was measuring different things. It would be too strong to conclude that a lower appetite rating necessarily means a smaller meal. It would also be wrong to conclude that unchanged meal intake means there was no appetite response. Both findings belong in the account of what happened. [9]

The 95% confidence interval for the difference in freely chosen energy intake was −391 to 346 kJ. For this post-treadmill meal comparison, the interval included both lower and higher intake after exercise. The study did not establish that exercise increased intake. “No clear evidence of an increase” is more accurate than treating the two conditions as proven identical. The uncertainty is part of the result, rather than a detail to discard once the headline has been chosen. [9]

In everyday life, hunger on an exercise evening can easily be attributed to the workout. But the time until dinner and what you ate earlier also belong in your description of the day. Noting these circumstances is a practical way to avoid mixing the feeling with the behavior. It is not a claim that a particular diary method produces weight loss. The useful lesson from the research is to ask separate questions about appetite and what was eaten. [9][12]

A single workout does not inevitably trigger eating back

Research on acute exercise examines a single bout and its aftermath. It does not, by itself, describe what happens after exercise becomes part of someone’s routine. A 2023 review found that moderate-to-vigorous land-based exercise generally suppresses subjective appetite for a time. Across the evidence, there was little support for compensatory increases in appetite or energy intake on the same day. That is a narrower and more useful conclusion than saying exercise never affects eating. [10]

The changes reviewed were temporary. Appetite-related hormone concentrations commonly returned to resting values in the hours after exercise. If you do not feel like eating immediately afterward, that feeling need not last until dinner. Equally, the return of appetite is not evidence that your body is demanding repayment of every calorie used. Returning toward the resting state and increasing intake enough to compensate for expenditure are distinct events. [10]

This timing matters when people compare experiences. Someone who rates hunger soon after exercising may give a different answer from someone describing how they felt later at home. Neither account automatically tells us what happened to the total food eaten that day. A statement about a short-lived suppression of appetite should retain its time window. Otherwise, a limited finding is quietly turned into a promise about an entire evening. [10]

The review also distinguished an energy deficit produced by exercise from one produced by reducing food intake. Those deficits may look similar when expressed as an energy shortfall, but the appetite responses are not necessarily alike. The evidence described an immediate appetite and intake response to diet-induced deficits that contrasted with the response to acute exercise-induced deficits. Predicting identical hunger from identical energy arithmetic misses that difference. [10]

That finding is not a prescription to exercise harder in order to erase appetite. The review emphasized substantial differences between individuals in subjective and hormonal responses. Some people experienced more appetite suppression than others. It also remained uncertain how far the short-term effects were relevant to longer-term appetite regulation and body weight. A result from a short experiment does not establish an exercise intensity that reliably prevents overeating in daily life. [10]

A 2018 review reached a related conclusion: individual bouts can create a short-term energy deficit without necessarily provoking a compensatory appetite response. On balance, the evidence suggested that adiposity and sex did not modify appetite or energy intake responses to acute or chronic exercise interventions. People with higher habitual activity might adjust intake more effectively when energy balance changes, but the authors called for further well-controlled research. That possibility should remain a possibility. [1]

If you remember eating more after exercise, the average research finding does not cancel your experience. The question is whether that experience supports the broader conclusion you are drawing. “I ate extra food after this session” describes an event. “Exercise always makes me eat back everything” adds a claim about a recurring pattern and its size. Looking at the particular meal or addition gives you something concrete to examine before deciding that exercise has no place in your routine. [1][10]

Longer-term training brings small average changes in hunger

A 2021 systematic review and meta-analysis examined exercise training in adults with overweight or obesity. It included 48 articles, with a median intervention duration of 12 weeks. The authors evaluated the training studies systematically and pooled comparable results to examine appetite across continued exercise. This evidence addresses a different timescale from the laboratory studies following an individual workout. It asks about changes across a period of continued exercise. [2]

Across the included research, fasting hunger and dietary restraint showed negligible-to-small increases. Dietary restraint concerns efforts to limit eating. Disinhibition, the tendency for restraint over eating to break down, decreased. There were also some favorable changes in satiety and food reward or preferences. These outcomes do not all move together: a small increase in hunger can coexist with changes that suggest less susceptibility to overconsumption. [2]

It is tempting to choose the outcome that fits an existing view of exercise. A person worried about appetite might focus on the hunger increase. Someone eager to defend exercise might focus on the favorable changes in satiety. The review includes both. Its findings do not justify reducing the response to “exercise stimulates appetite” or “exercise suppresses appetite.” Continued training can involve different changes in different parts of appetite control. [2]

Among studies rated fair or good in quality, there was no significant change in daily energy intake from before to after training. Yet a comparison with no-exercise control groups at the end of the interventions showed an average difference of 102 kcal, with a confidence interval of 1 to 203 kcal. The authors described the standardized effect as small but negligible. The difference should be reported, along with that assessment of its size. [2]

These statements can both be true because they use different comparisons. A before-and-after analysis asks whether the exercise group changed from its own starting point. A comparison at the end asks whether that group differed from a group that did not exercise. They are not interchangeable questions. Reading only “no significant change” or only “higher intake” removes the comparison that gives the result its meaning. [2]

The quality of the evidence is a substantial limitation. Of the 48 articles, 39 were rated poor, 7 fair, and 2 good. Many studies relied on self-reported daily food intake. A large number of papers does not make every measurement equally dependable. The conclusion about small average changes must therefore travel with its qualification: the authors were working within limitations imposed by study quality and by how intake was assessed. [2]

This is particularly relevant when a small group difference is turned into a personal calorie rule. The 102 kcal figure describes a comparison in the reviewed research; it is not an amount each exerciser is expected to add to their meals. It does not tell you how much to remove from dinner either. The result helps characterize the average response under those study conditions, while leaving individual responses and measurement limitations visible. [2]

Someone can become a little hungrier and still stop eating when satisfied. That is a plausible way to think about the separation between hunger and intake, rather than a diagnosis of what is happening in your body. The reviews suggest that training may influence both hunger and meal-related satiety. They do not support treating the whole response as a simple upward push on appetite, nor do they guarantee that every person will avoid compensatory eating. [1][2]

A trial found eating back and less weight loss than expected

The E-MECHANIC randomized controlled trial provides an important unfavorable finding. It assigned 198 adults with overweight or obesity to a no-exercise control group or to supervised exercise at different doses. The program lasted 24 weeks. A per-protocol analysis, involving participants who followed the protocol, included 171 people, of whom 72.5% were women. The study examined why weight loss with continued exercise could fall short of what was expected. [4]

The exercise doses were 8 kcal per kilogram of body weight per week and 20 kcal per kilogram per week. These were conditions used to separate the trial groups. They are not ready-made instructions for readers to calculate their own exercise schedules. The study estimated intake using doubly labeled water, an energy assessment method, and separately examined physical activity and resting metabolic rate. It therefore looked beyond participants’ impressions of how hungry they had become. [4]

Using estimates adjusted for changes in resting metabolic rate, daily energy intake increased by 90.7 kcal in the lower-dose group and 123.6 kcal in the higher-dose group. Their respective 95% confidence intervals were 35.1 to 146.4 kcal and 64.5 to 182.7 kcal. The control group’s change was −2.3 kcal. The increases in the exercise groups were significant compared with the control group. This trial did observe additional intake during an exercise program. [4]

Weight compensation averaged 1.5 kg in the lower-dose group and 2.7 kg in the higher-dose group. The corresponding 95% confidence intervals were 0.9 to 2.2 kg and 2.0 to 3.5 kg. Here, compensation means the amount by which actual weight loss fell short of expected weight loss. It does not mean participants gained that much weight because they exercised. Losing less than predicted and gaining weight are different descriptions. [4]

Changes in resting metabolic rate and activity outside the prescribed exercise did not differ between groups. The authors attributed compensation to increased energy intake together with increased appetite. That finding makes eating more a relevant explanation in this trial. It does not establish that every disappointing weight outcome after exercise has the same cause. The participants, program, duration, and measurements define what the study can tell us. [4]

Participants with greater compensation reported increased appetite and stronger beliefs that healthy behaviors could compensate for unhealthy ones. They also reported increased cravings for sweet foods, more sleep disturbance, and worsening bodily pain. These findings give a fuller account than a calorie figure alone. However, an association between a participant’s beliefs and compensation does not prove that the belief caused the outcome, even though assignment to exercise groups was randomized. [4]

Numbers: exercise, appetite and compensation

48 articles: the review of training interventions in adults with overweight or obesity, with substantial differences in study quality. [2]

12 weeks: the median training duration in that review, a different timescale from appetite immediately after a workout. [2]

102 kcal: the average end-of-intervention difference versus no-exercise controls in fair or good studies; confidence interval 1 to 203 kcal, with a negligible standardized effect. [2]

24 weeks: the duration of the randomized trial in which increased intake and weight compensation were observed. [4]

23 studies: the psychology review’s evidence base; the population prevalence of these eating behaviors remained unknown. [5]