Study quantifies how much semaglutide cuts food intake

Adults with obesity taking once-weekly semaglutide 2.4 mg ate about 35% less food at an unrestricted lunch than people on placebo, according to a phase 1 trial published in Diabetes, Obesity and Metabolism. ^[1] The same study found no clear evidence that the drug delayed stomach emptying after 20 weeks of treatment, which shifts attention toward appetite signaling rather than a "full stomach" effect as the main reason people eat less. ^[1]

The trial enrolled 72 adults with obesity at a single center in Germany and randomly assigned them, double-blind, to either semaglutide escalated to 2.4 mg once weekly or placebo for 20 weeks, followed by a 7-week follow-up period. ^[1] It was a parallel-group design, meaning each participant received only one treatment. ^[1] The study was funded by Novo Nordisk, and several authors are Novo Nordisk employees. ^[1]

What the numbers showed

At an "ad libitum" lunch, where participants could eat as much as they wanted, energy intake was 1736 kilojoules on semaglutide versus 2676 kilojoules on placebo, an estimated treatment difference of −940 kJ (P < 0.0001). ^[1] In more familiar units, that is roughly 415 calories versus 640 calories at that single meal, a gap of about 225 calories.

Participants on semaglutide also reported less hunger and lower prospective food consumption, plus greater fullness and satiety, compared with placebo (all P < 0.02). ^[1] On the Control of Eating Questionnaire, they reported better control of eating and fewer or weaker food cravings (P < 0.05). ^[1]

Body weight fell 9.9% with semaglutide and 0.4% with placebo over the 20 weeks. ^[1] Safety findings were consistent with what is already known about semaglutide, the authors wrote. ^[1]

The stomach-emptying question

The trial's primary endpoint was gastric emptying, measured indirectly by how quickly the body absorbed paracetamol (acetaminophen) after a standardized breakfast. ^[1] The area under the concentration-time curve over 0 to 5 hours was 8% higher with semaglutide than placebo (P = 0.005) — but that difference was no longer statistically significant after adjusting for participants' week 20 body weight (P = 0.12). ^[1] There was no effect on absorption in the first hour, on peak paracetamol concentration, or on the time to reach that peak. ^[1]

The authors concluded there was no evidence of delayed gastric emptying at week 20. ^[1] That matters because an earlier 12-week study of semaglutide 1.0 mg in people with obesity had suggested a delay in first-hour gastric emptying, and because slowed emptying can in theory affect how other oral medicines are absorbed. ^[1] This study did not include brain imaging or other direct measures of central appetite pathways, so it does not itself prove where the appetite effect originates.

Why it matters for patients

The findings put a number on something many people on GLP-1 medications describe: eating less without consciously trying. A roughly one-third drop in intake at a single test meal, along with lower hunger scores and fewer cravings, lines up with the weight loss seen in the larger STEP program for semaglutide 2.4 mg, which is sold as Wegovy for weight management. ^[1]

The gastric-emptying result is relevant to a common assumption — that GLP-1 drugs work mainly by keeping food in the stomach longer. After 20 weeks at 2.4 mg, this trial found essentially no measurable delay using the paracetamol method. ^[1] That said, paracetamol absorption is an indirect measure, the trial tested emptying only at week 20 rather than during dose escalation, and 72 participants is a small sample. ^[1] Whether stomach emptying is slowed earlier in treatment, when nausea is often most common, was not answered here.

A few limits are worth keeping in mind. The lunch measurement captures one meal in a lab setting, not daily eating over months. ^[1] The trial was in adults with obesity and used the 2.4 mg dose, so it does not speak to lower doses or to people with type 2 diabetes. And because the study was industry-funded and industry-authored, independent replication carries weight. ^[1]

The paper appeared online on December 2, 2020, and in print in Volume 23, Issue 3, pages 754 to 762. ^[1] The trial is registered as NCT03842202. ^[1]