Skip to content
GLP Loss
← Research
Evidence

How GLP-1 Drugs Work, and What the Mechanism Leaves Out

The delayed-stomach explanation is measurably temporary: gone by week 20 on semaglutide 2.4 mg, in the same trial where treated adults ate 35% less at a free lunch and had lost 9.9% of body weight.

Owen Castellanos9 min read
The stomach effect fades. The eating effect does not.Three measurements of gastric emptying, taken at three exposuresOne day of infusion, nine healthy volunteersEmptying slowed after meal one, much less after meal two four hours laterTwelve weeks, semaglutide 1.0 mg, 30 adultsFirst hour delayed, ratio 0.73. Across five hours, no difference.Twenty weeks, semaglutide 2.4 mg, 72 adultsNo delay at either window, corrected for the weight already lostAt that same week-20 visitFree lunch: 1736 kJ against 2676 kJ. Weight: down 9.9% against 0.4%.The part of the mechanism people are told is the part that leaves first.

The explanation that travels with these drugs is that they slow the stomach, so a smaller meal feels like a larger one. That claim has been measured on the drug itself, and at twenty weeks it is gone. In 72 adults with obesity randomized to once-weekly semaglutide 2.4 mg or placebo, gastric emptying at week 20 showed no effect on the first-hour measure, and the five-hour measure lost significance once week-20 body weight was accounted for (P = 0.12).[1] At that same visit, the treated group ate 35% less at a free lunch — 1736 against 2676 kJ — and had lost 9.9% of body weight against 0.4% on placebo.[1] The stomach effect had faded. The eating effect had not.

A hormone found doing something else entirely

GLP-1 was not discovered as an appetite hormone. It was isolated inside the incretin effect: the observation that glucose taken by mouth provokes far more insulin than the same blood glucose delivered intravenously, because the gut releases hormones that the vein does not. Fourteen adults with type 2 diabetes and eight matched healthy controls were given a 50 g oral glucose load and, separately, an isoglycemic intravenous infusion. Gut hormones accounted for 72.8 ± 6.9% of the insulin response in the healthy group and 36.0 ± 8.8% in the diabetic group (P ≤ 0.05).[2] That halving is why the class exists, and it is a glucose finding. Weight came later, and the two indications still diverge, which is the subject of the two-approvals article.

The order matters for reading a sales page. A drug developed to restore an insulin signal turned out to reduce food intake, and the appetite effect was characterized after the molecule already existed, not before. That is the opposite of the usual story, in which a mechanism is worked out and a drug is then designed against it. It also means the appetite pathway was mapped in animals afterward, to explain a human result that had already been observed — which is why the mechanistic literature below is largely rodent and the efficacy literature is entirely human.

The hormone and the drug are not the same object

Native GLP-1 is destroyed within minutes of release, so the only way to study it in people is to infuse it continuously. Pooled subject-level data from 115 adults across intravenous infusion studies found energy intake at the following meal reduced by 727 kJ (95% CI, 548 to 908), or 11.7%. The absolute reduction was larger in lean participants (863 kJ) than in overweight ones (487 kJ), though the relative reduction was similar (13.2% against 9.3%), and infusion rate was the only independent predictor of the effect (r = 0.4, P < 0.001).[3]

One detail in that pooled analysis is worth more than the headline. The rise in plasma GLP-1 correlated with how full and how unwilling to eat more people said they felt — prospective consumption r = 0.40, fullness r = 0.38 — and did not correlate with how much they actually ate.[3] An eleven-percent cut in a single meal from an infusion of the natural hormone sits beside a 14.9% reduction in body weight over 68 weeks from a weekly injection of an engineered analog.[4] The drug is not the hormone turned up. It is a molecule redesigned to persist for a week, at exposures the gut never produces.

Gastric emptying, and how fast it stops mattering

The deceleration is real, and it is measurably temporary. Nine healthy volunteers received intravenous GLP-1 at 0.8 pmol/kg/min or placebo over 8.5 hours, with two liquid mixed meals four hours apart. Emptying was slowed significantly more after the first meal than after the second (P = 0.01). Pancreatic polypeptide, a marker of vagal activation, fell after the first meal and not the second, and glucose declined after meal one but rose after meal two.[5] Tachyphylaxis inside a single day, at the level of the nerve.

Longer exposures repeat the pattern. Thirty adults with obesity took semaglutide escalated to 1.0 mg for 12 weeks in a crossover trial: first-hour emptying was delayed, at an estimated treatment ratio of 0.73 (95% CI, 0.61 to 0.87), while overall emptying across five hours was not statistically different from placebo.[6] At 2.4 mg and twenty weeks, neither window showed an effect.[1] The honest summary is that delayed emptying is an early, first-meal phenomenon that attenuates, that it explains postprandial glucose better than it explains a year of weight loss, and that it explains most of the nausea — which is why nausea follows the same timeline and why the motility questions covered in the gastroparesis article are about a minority, not the mechanism.

Brain or gut: the evidence is animal, and it says both

Where the appetite effect is generated has been mapped in rodents, not in people. In mice and rats, semaglutide did not cross the blood-brain barrier. It reached the brainstem, the septal nucleus and the hypothalamus through the circumventricular organs and a handful of sites adjacent to the ventricles, and produced c-Fos activation in 10 brain areas — including regions such as the lateral parabrachial nucleus, where it has no direct receptor contact at all.[7] The drug does not bathe the brain. It touches a small number of doors and the signal propagates inward.

The gut arm has been isolated the same way. Using intersectional genetics in mice, gut-innervating vagal afferents that express the GLP-1 receptor were shown to relay meal-terminating signals to parabrachial neurons; activating them also improved glucose tolerance, and silencing them raised blood glucose independently of food intake. A separate afferent population marked by GPR65 drove hepatic glucose production and was dispensable for feeding altogether.[8] Two routes, one destination, both demonstrated in mice. No human experiment has separated them, and any consumer page that assigns a percentage to gut versus brain is assigning it to nothing.

The effect is on intake, not on burn

This is the part most often reversed in marketing copy. In the same 30-person crossover trial, semaglutide at 1.0 mg cut total ad libitum energy intake across lunch, dinner and snacks by 24% (−3036 kJ, P < 0.0001), lowered hunger and food cravings, and reduced the relative preference for high-fat foods — while resting metabolic rate adjusted for lean body mass did not differ between treatments. Weight fell 5.0 kg, predominantly as fat.[9] The rodent work reports the same direction: food intake and body weight fell without a decrease in energy expenditure.[7]

Nothing here speeds up a metabolism. The entire measured effect sits on the intake side of the equation, which is exactly why what comes off depends so heavily on what is eaten and whether anything is being trained — the concern in the body-composition article.

Four things the mechanism does not explain

Who responds. In STEP 1, 86.4% of participants on semaglutide reached a 5% reduction, 69.1% reached 10% and 50.5% reached 15%, against 31.5%, 12.0% and 4.9% on placebo.[4] Read the first figure the other way: roughly one in seven people on the full dose did not reach 5%, on the same receptor and the same titration. No feature of the pharmacology identifies them beforehand, and no test sold with a subscription does either. The placebo column is worth reading alongside it: 31.5% of people taking nothing reached 5% too, so a single person’s result is never clean evidence that the drug did or did not work for them.

Why it stops. Receptor agonism does not switch off at a particular week, yet the weight curve flattens — a separate problem covered in the plateau article.

Why it reaches things that are not food. The reward circuitry implicated in the rodent maps is not food-specific, and the clinical signals in alcohol and nicotine are much weaker than the mechanistic story predicts, which is the whole argument in the alcohol-use article.

Why the dose ladder works. Escalation exists because tolerability demands it, not because the receptor requires a run-up. The schedules are in the titration article, and the expected-loss arithmetic is in the expected weight loss tool.

What the mechanism does and does not certify at checkout

Every measurement above was made on FDA-approved product at labeled doses under trial conditions. Most sellers covered here dispense compounded semaglutide or tirzepatide, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed. A shared mechanism is a claim about a molecule, not about a vial: it says nothing about what concentration arrived, and nothing about whether the dose on the label is the dose in the syringe.

The practical reading of all of this is narrow and useful. The drug works by making less food feel like enough, the stomach contributes early and then recedes, nothing about it raises the rate at which calories are burned, and about one person in seven at full dose will not get a meaningful result from it. A page that promises a metabolic reset is describing a mechanism that has not been measured, on a product whose contents have not been verified.

Frequently asked

Do GLP-1 drugs work by slowing down your stomach?
Only partly, and mostly early on. An intravenous infusion of GLP-1 slowed gastric emptying markedly after a first meal and much less after a second meal four hours later, and 12 weeks of semaglutide 1.0 mg left a first-hour delay with no difference across five hours. At 2.4 mg and 20 weeks there was no measurable delay at either window, in a trial where the same participants still ate 35% less at a test meal.
What is the incretin effect?
It is the extra insulin released when glucose arrives through the gut rather than through a vein, because the intestine secretes hormones that an infusion does not. In the study that defined it, gut hormones accounted for 72.8% of the insulin response in healthy adults and 36.0% in adults with type 2 diabetes. GLP-1 is one of the two hormones responsible, and the drug class was built to replace a signal that had weakened.
Does a GLP-1 speed up your metabolism?
No measurement supports that. In a crossover trial of semaglutide 1.0 mg, resting metabolic rate adjusted for lean body mass did not differ from placebo while total food eaten across the day fell 24%. Rodent work reports the same direction: intake and weight fall without a decrease in energy expenditure. The entire measured effect is on the intake side.
Does the drug act on the brain or on the gut?
The mapping evidence is animal and it points to both. In mice and rats, semaglutide did not cross the blood-brain barrier and instead reached the brainstem and hypothalamus through the circumventricular organs, activating ten brain areas including regions it never touches directly. Separately, in mice, gut vagal nerve fibers carrying the GLP-1 receptor were shown to signal meal termination to the same brainstem region. No human experiment has separated the two routes.
Why do some people lose almost nothing on a GLP-1?
Nobody has established why. In STEP 1, 86.4% of participants on semaglutide reached a 5% reduction, which means roughly one in seven did not, at the same full dose. Nothing in the known mechanism identifies non-responders in advance, and no test offered with a telehealth subscription predicts it.
Does a shared mechanism mean a compounded vial works the same?
It does not. Compounded semaglutide and tirzepatide are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before dispensing, and every mechanistic measurement cited here was made on approved product at labeled doses. A mechanism is a fact about a molecule, not evidence about the concentration or purity of what arrived in a particular vial.

Sources

  1. [1] Friedrichsen M, Breitschaft A, Tadayon S, et al. (2021). The effect of semaglutide 2.4 mg once weekly on energy intake, appetite, control of eating, and gastric emptying in adults with obesity. Diabetes Obes Metab. PMID 33269530
  2. [2] Nauck M, Stöckmann F, Ebert R, Creutzfeldt W (1986). Reduced incretin effect in type 2 (non-insulin-dependent) diabetes. Diabetologia. PMID 3514343
  3. [3] Verdich C, Flint A, Gutzwiller JP, et al. (2001). A meta-analysis of the effect of glucagon-like peptide-1 (7-36) amide on ad libitum energy intake in humans. J Clin Endocrinol Metab. PMID 11549680
  4. [4] Wilding JPH, Batterham RL, Calanna S, et al. (2021). Once-Weekly Semaglutide in Adults with Overweight or Obesity. N Engl J Med. PMID 33567185
  5. [5] Nauck MA, Kemmeries G, Holst JJ, Meier JJ (2011). Rapid tachyphylaxis of the glucagon-like peptide 1-induced deceleration of gastric emptying in humans. Diabetes. PMID 21430088
  6. [6] Hjerpsted JB, Flint A, Brooks A, et al. (2018). Semaglutide improves postprandial glucose and lipid metabolism, and delays first-hour gastric emptying in subjects with obesity. Diabetes Obes Metab. PMID 28941314
  7. [7] Gabery S, Salinas CG, Paulsen SJ, et al. (2020). Semaglutide lowers body weight in rodents via distributed neural pathways. JCI Insight. PMID 32213703
  8. [8] Borgmann D, Ciglieri E, Biglari N, et al. (2021). Gut-brain communication by distinct sensory neurons differently controls feeding and glucose metabolism. Cell Metab. PMID 34043943
  9. [9] Blundell J, Finlayson G, Axelsen M, et al. (2017). Effects of once-weekly semaglutide on appetite, energy intake, control of eating, food preference and body weight in subjects with obesity. Diabetes Obes Metab. PMID 28266779

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence