How GLP-1s Produce Weight Loss, According to Four Trials

GLP-1 medications produce weight loss mainly by making people eat less, not by burning more calories at rest. Four randomized trials measured that mechanism directly, tracking appetite, actual food intake, and how fast food leaves the stomach rather than only weight on a scale.

If you are starting a GLP-1 medication, or deciding whether to, you have probably been told it "makes you less hungry." That is true, but it does not say what changes, when, or how much of the weight effect it explains. Researchers brought participants into a research unit, put real food in front of them, and measured how much they actually ate, what they said about their own hunger, and how fast a meal left the stomach. The mechanism was measured directly, not left as a slogan.

The main effect is on how much people eat

In a 12-week randomized, double-blind crossover trial in 30 adults with obesity (Diabetes, Obesity and Metabolism, 2017), each participant spent one period on once-weekly semaglutide escalated to 1.0 mg and one period on placebo. On test days they were given free access to food and told to eat as much as they wanted. Total energy intake across all meals was about 24% lower on semaglutide than on placebo, a difference of 3036 kJ (P < 0.0001). Lunch alone was 35% lower. Mean body weight fell 5.0 kg during the semaglutide period and rose 1.0 kg during placebo.

A separate 20-week double-blind trial in 72 adults with obesity (Diabetes, Obesity and Metabolism, 2021) tested semaglutide 2.4 mg once weekly against placebo. At week 20, energy intake at the free-access test meal was 1736 kJ on semaglutide versus 2676 kJ on placebo, roughly 35% lower, with an estimated treatment difference of 940 kJ (P < 0.0001).

Both trials were small, ran under research conditions, and were funded by Novo Nordisk, which makes semaglutide. These are averages that were observed in those participants, not a forecast of what any one person will eat.

What "less hunger" looked like on the actual scales

Both trials also asked participants to rate their own appetite. In the 20-week trial, semaglutide reduced hunger and prospective food consumption and increased fullness and satiety, all at P < 0.02, with a higher overall appetite suppression score (estimated treatment difference 13 mm, P = 0.001). Participants reported better control of eating and fewer, weaker food cravings, including less desire for sweet and savory foods.

The 2017 crossover trial found the same pattern and added something more specific about food choice. Participants showed lower explicit liking for high-fat, non-sweet foods on semaglutide than on placebo (P = 0.0016), and lower implicit wanting for them (P = 0.0203). So the shift was not only in quantity. What people were drawn to changed as well.

That trial also checked whether the drug was burning more energy at rest. Once resting metabolic rate was adjusted for lean body mass, the difference between semaglutide and placebo was not statistically significant (P = 0.0704). The measured route to weight loss ran through intake, not through a faster metabolism.

The gastric emptying story is narrower than you have heard

"It slows your stomach" is the most common one-line explanation, and the evidence for it is more specific than that.

In a 2018 analysis of the same 30-person crossover trial, gastric emptying was measured by paracetamol absorption after a meal. In the first hour, emptying was slower on semaglutide 1.0 mg than on placebo (estimated treatment ratio 0.73, 95% CI 0.61 to 0.87, P = 0.0012), about 27% slower. Across the full five hours, the difference was not statistically significant (ratio 0.94, 95% CI 0.88 to 1.01). The effect sat in the window right after eating, which is also when it would most plausibly affect how much of a meal you finish.

The 20-week trial at 2.4 mg did not reproduce that. Its primary gastric emptying endpoint, paracetamol AUC over five hours, was 8% higher on semaglutide (P = 0.005), pointing toward faster rather than slower emptying, and that difference was no longer statistically significant after adjusting for body weight at week 20 (P = 0.1218). The first-hour measure showed no difference between groups. The authors noted that paracetamol absorption is an indirect method with limits, particularly for solid food, and may miss short-term delays.

Taken together, the delay these trials captured is early, was clearest in the shorter 1.0 mg trial, and was not the dominant finding in either study. Reduced intake was.

Appetite does not explain all of it

A 28-week randomized, double-blind trial in people with type 2 diabetes (Diabetes Care, 2023) compared tirzepatide 15 mg (n = 45), semaglutide 1 mg (n = 44), and placebo (n = 28). Both drugs significantly reduced fasting appetite and free-access lunch energy intake versus placebo. Tirzepatide produced greater weight loss than semaglutide (difference 4.3 kg, P < 0.001), driven mostly by fat mass (difference 3.8 kg, P = 0.002).

But the lunch intake difference between the two drugs was 64.3 kcal and not statistically significant (P = 0.187). The authors state plainly that differences in energy intake at that lunch were not sufficient to explain the different weight outcomes. Something beyond a single measured meal is contributing, and this trial did not identify what.

This was a head-to-head comparison, which makes it worth citing, but note the boundaries: it enrolled people with type 2 diabetes, used semaglutide at 1 mg rather than the higher dose studied for weight management, measured intake at one lunch rather than across a full day, and was funded by Eli Lilly, which makes tirzepatide.

What this means if you are weighing one

The honest summary is that these medications were shown, in randomized trials, to lower how much people ate, lower hunger and cravings, raise fullness, and shift preference away from high-fat foods, with a stomach-emptying delay concentrated in the first hour after a meal at the lower dose studied. That is a real mechanism, measured against placebo, not a marketing story.

The limits are equally real, and so are the risks. In the 20-week trial, gastrointestinal side effects were reported by 69.4% of participants on semaglutide versus 38.9% on placebo, with nausea and diarrhoea most commonly reported; the trial recorded these as mild or moderate and generally short-lived, and none of these four studies was designed to characterise safety. These were small trials, 12 to 28 weeks long, run in research units where test meals are not how anyone normally eats, all funded by the companies that make the drugs, and all in adults with obesity or type 2 diabetes. Appetite measures did not fully account for the weight differences even between two drugs in the same trial. And none of these studies were designed to answer what happens over years, how you feel day to day, or whether the trade is right for your health history. Those are questions for a licensed provider who can look at your labs and your medical history.