Liraglutide vs. Exercise for Weight Maintenance

After weight loss, hunger often returns before the celebratory jeans have finished their first laundry cycle. The practical question is whether a GLP-1 drug changes a person's eating behavior or suppresses appetite only while the drug remains in circulation. A 52-week randomized trial offers unusually direct, but still preliminary, evidence: liraglutide prevented the appetite rebound seen after diet-induced weight loss, while exercise alone did not show the same advantage.

The ceiling matters. The eating-behavior and activity findings came from exploratory secondary analyses with small groups and unadjusted p-values. They generate useful hypotheses; they do not prove that liraglutide permanently rewires anyone's relationship with food.

A rare head-to-head test after substantial weight loss

In the S-LiTE trial, adults first completed an eight-week low-calorie diet and lost a mean 13.7 kg. Those who had lost at least 5% of body weight were then assigned to 52 weeks of placebo, structured exercise, liraglutide 3.0 mg daily, or the combination. The randomized intention-to-treat sample included 195 people; the exploratory per-protocol behavioral analyses included 39 in placebo, 26 in exercise, 36 in liraglutide, and 29 in the combination group.

This design gives the comparison real value. It tests the interventions against the same post-diet biology, rather than comparing a medication study with an unrelated exercise study in a different population. But the trial was designed around weight maintenance, not to deliver definitive answers on every appetite questionnaire and accelerometer measure.

Liraglutide held back the post-diet appetite rebound

In the placebo group, post-meal appetite worsened over 52 weeks. Hunger rose about 21.9% (P=0.002), prospective food consumption rose 19.2% (P=0.01), fullness fell 13.4% (P=0.01), and satiety fell 13.6% (P=0.008). The composite Overall Appetite Suppression score declined 14.0% (P=0.001).

Liraglutide prevented that decline, producing an approximately 14% advantage over placebo on the composite score at week 52. It also maintained more weight loss: the mean difference versus placebo was -8.0 kg (95% CI, -11.7 to -4.2). Because that confidence interval excludes zero, the weight difference was statistically significant. The appetite result is less secure than the weight result because it was exploratory and one of many unadjusted behavioral comparisons.

The direction fits the drug's known pharmacology. Liraglutide is a GLP-1 receptor agonist that slows gastric emptying and affects central and peripheral signals involved in satiety, appetite, and food intake, as summarized in a 2021 liraglutide review. Mechanistic plausibility helps explain a trial result; it does not turn a 52-week questionnaire outcome into evidence of lasting behavioral retraining.

Exercise maintained weight, but its behavioral signal was smaller

Exercise alone supported weight maintenance compared with placebo, but S-LiTE did not find a statistically detectable appetite or eating-restraint advantage over placebo in the exercise group. It also did not detect compensatory hyperphagia. That result should not become "exercise causes no compensation." With only 26 people in that group, the analysis could easily miss a meaningful behavioral effect.

Exercise compensation also varies by setting. A 2024 review by Flack and colleagues describes ways people may offset exercise energy expenditure, including changes in non-exercise activity and motivation to move. In a separate 2024 pilot randomized trial in midlife and older Latino adults, the authors note that exercise interventions can increase sedentary time through compensatory rest. Neither paper studied S-LiTE's post-low-calorie-diet population, so neither overturns its finding. They do explain why a small exercise arm cannot settle the question.

The combination group had the most interesting, least settled result

Combination treatment produced the largest behavioral changes in S-LiTE. Compared with placebo, sedentary time fell by about 41 minutes per day, while cognitive restraint improved about 13%; it worsened about 9% in the placebo group. The combination also had the largest weight-maintenance difference versus placebo: -12.1 kg (95% CI, -16.1 to -8.1).

Those numbers make the combination worth studying further. They do not establish that exercise and liraglutide have proven additive behavioral mechanisms. The group included 29 people, and the trial cannot separate a true interaction between treatments from chance variation or other explanations. Exercise still has benefits a GLP-1 drug does not replace, including training adaptations and cardiovascular fitness, but S-LiTE did not test whether those benefits explain the behavioral findings.

The one-year appetite evidence already conflicts

The strongest reason to avoid grand claims comes from the S-LiTE paper itself. It cites another randomized trial of liraglutide 3.0 mg daily plus intensive behavioral therapy that found less hunger and greater fullness at 24 weeks, but no appetite benefit at 12 months despite continued treatment and better weight loss than placebo. Same drug dose, same one-year time point, different behavioral result.

That discrepancy could reflect differences in population, behavioral program, measurement, or ordinary statistical uncertainty. The available evidence does not identify the cause. A 2024 review of incretin durability similarly raises questions about how sustainable incretin-agonist effects are. It does not answer whether appetite suppression fades in every user; it reinforces that the answer is unsettled.

What happens after stopping remains unknown

No trial in this evidence base reports eating-behavior outcomes after GLP-1 discontinuation. S-LiTE kept participants on their assigned treatment throughout its 52 weeks. Weight regain after pharmacologic or dietary weight loss is common, but the relative roles of appetite, metabolism, reward, and behavior remain unresolved.

A functional MRI study by Salem and colleagues offers a related caution, not direct evidence about liraglutide. In people with diabetes, low-calorie-diet weight loss and gastric bypass produced different brain responses to food cues. Diet-induced loss did not reproduce the surgery-associated pattern. That finding cannot predict what liraglutide cessation will do, but it weakens the casual assumption that losing weight automatically resets food-reward circuitry for good.

For now, liraglutide's appetite effect looks best described as treatment-associated. Evidence does not show that it persists after treatment stops.

The practical verdict

Among the interventions tested in S-LiTE, liraglutide had the better one-year signal for preventing post-diet appetite rebound, and the combination group had the best weight and exploratory behavioral outcomes. Exercise alone did not appear to worsen eating behavior in this trial, but its small sample leaves substantial uncertainty about compensatory hunger and sedentary time.

If the goal is durable behavior change after stopping a GLP-1 drug, the evidence has not caught up with the marketing. Treat appetite suppression during active liraglutide use as a plausible and trial-supported drug effect. Treat permanent reprogramming, and a proven additive combination protocol, as questions for larger trials with corrected analyses and post-treatment follow-up.