GLP-1 & metabolic peptides

Start with the observation the whole field grew out of: the incretin effect. Give someone glucose by mouth and they release far more insulin than if the identical amount of glucose is delivered straight into a vein. The gut is telling the pancreas that food is coming, and in healthy people that gut signal accounts for a large share of the insulin released after a meal. GLP-1 is one of the two hormones responsible; the other is GIP ^[3].

The insulin release is glucose-dependent, and that detail is the reason this class became a safe drug rather than a dangerous one. GLP-1 amplifies insulin secretion only when glucose is already elevated, so on its own it lowers blood sugar with very little risk of driving it too low. That is not true of insulin or sulfonylureas, and it is why those combinations, and not GLP-1 alone, are where hypoglycaemia appears.

Slowed gastric emptying contributes to fullness after a meal and blunts the post-meal glucose spike. This effect attenuates somewhat with continued exposure to a long-acting agonist, which is part of why nausea usually eases over weeks, and it is also the mechanism behind two of the class's practical problems: the gastrointestinal side effects and the anaesthetic question further down this page.

The appetite effect is central and worth understanding properly. GLP-1 receptors are present in the hypothalamic arcuate nucleus and in the brainstem, and these areas sit outside or adjacent to the blood brain barrier, so a circulating agonist can reach them ^[1]. Activation reduces food intake, and the subjective description people give is less about willpower than about the constant background pull of food quietening down. Whether that quietening is specific to food is one of the genuinely interesting open questions in the field.

The problem with the natural hormone is that it is gone almost immediately. The enzyme DPP-4 cleaves two residues from GLP-1's N terminus within about two minutes, and the kidney filters what is left. Nothing about the biology needed fixing; the entire pharmaceutical project was making the message last ^[2].

Three pieces of chemistry solved the durability problem, and every drug in the table uses some combination of them. First, change the residue DPP-4 cuts, so the enzyme has nothing to act on. Second, hang a fatty diacid off a side chain, so the molecule binds albumin in circulation; albumin is far too large for the kidney to filter, so the drug is protected and released slowly. Third, stiffen the backbone with unnatural residues so that circulating peptidases cannot get purchase. Together those take a two-minute hormone to a once-weekly injection ^[2].

Then the field moved from copying one hormone to copying several. Tirzepatide activates both the GIP and GLP-1 receptors with a single molecule. GIP is the other incretin, and its contribution is still being worked out, but the combination clearly outperforms GLP-1 alone on weight ^[7]. Retatrutide adds a third: the glucagon receptor. Glucagon raises energy expenditure and drives hepatic fat oxidation, which is why adding an agonist of it to a drug that suppresses appetite is less contradictory than it first sounds. Its phase 2 results were the largest weight reductions reported for a drug at that stage ^[8].

Oral semaglutide deserves its own note because it is not what people assume. The tablet contains an absorption enhancer, SNAC, which raises the pH locally against the stomach wall and helps a small fraction of the peptide cross there rather than in the intestine. Bioavailability is on the order of one percent, absorption is variable, and the tablet must be taken fasted with very little water; this is why the oral strengths look nothing like the injectable ones ^[4].

The table sets out the main compounds. Read the weight-loss column as trial averages under trial conditions, with lifestyle support and steady dose escalation, not as a promise.

The reason this class is treated as more than a cosmetic drug is that the benefits have been demonstrated on outcomes that matter, in trials designed to find them.

Cardiovascular. In people with type 2 diabetes at high cardiovascular risk, semaglutide reduced the rate of cardiovascular death, non-fatal heart attack and non-fatal stroke ^[9]. The more striking result came later: in more than 17,000 people with established cardiovascular disease and overweight or obesity but without diabetes, semaglutide reduced the same composite endpoint by about 20 percent over roughly three years ^[10]. That trial is what moved the class from a metabolic drug to a cardiovascular one.

Kidney. In people with type 2 diabetes and chronic kidney disease, semaglutide reduced a composite of kidney failure, substantial loss of kidney function and death from kidney or cardiovascular causes by about 24 percent ^[11].

Liver. Semaglutide produced significantly higher rates of resolution of steatohepatitis than placebo in a randomised trial in non-alcoholic steatohepatitis, though without a significant improvement in fibrosis stage at that point ^[13].

Sleep apnoea. Tirzepatide reduced the apnoea-hypopnoea index substantially against placebo in adults with moderate to severe obstructive sleep apnoea and obesity, in both treated and untreated groups ^[12]. Sleep apnoea is a good illustration of the general shape here: much of the benefit across these conditions plausibly runs through the weight loss itself rather than through a distinct drug effect, and the trials mostly cannot separate the two.

Addiction and compulsion is the interesting frontier and the one to be most careful about. There are GLP-1 receptors in reward-related brain regions, the mechanism by which food noise quietens is not obviously food-specific, and observational data and early trials in alcohol use disorder have generated real signals. It is a legitimate hypothesis under active investigation. It is not, at the time of writing, an established indication, and the gap between those two statements is where most of the coverage goes wrong.

Two things about this class are consistently underplayed, and they are related.

The first is body composition. Losing weight quickly means losing some lean tissue along with fat, and this is not specific to these drugs; it happens with any large energy deficit, including surgery and severe dieting. The commonly quoted figure is that something like a quarter to 40 percent of the total mass lost is lean tissue, and the imaging in these trials cannot distinguish muscle from the water, connective tissue and organ mass that also sit in the lean compartment. The practical response is the boring one: adequate protein intake and resistance training during the loss phase, which is well established to preserve lean mass under an energy deficit. This matters most in older adults, where the lean-mass reserve is already thinner.

The second is that the effect depends on continued treatment. In the extension of the main semaglutide obesity trial, participants who stopped the drug and the lifestyle support regained about two thirds of the weight they had lost within a year, and the improvements in blood pressure, lipids and glycaemic markers reverted along with it ^[6]. That is not a failure of the drug; it is what treating a chronic condition looks like, and it is the same pattern seen when antihypertensives are stopped. But it means the honest framing is ongoing treatment rather than a course.

Those two facts interact badly if the drug is used as a short intervention: the lean mass lost during a rapid decline is not automatically the tissue regained afterwards, and repeated cycles of loss and regain can shift composition in the wrong direction over time. That is a mechanistic concern rather than a demonstrated harm of these drugs specifically, and it deserves to be described as such.

The most common problems are gastrointestinal: nausea, vomiting, diarrhoea, constipation. These follow directly from slowed gastric emptying, they are dose-related, and they are the reason every one of these drugs is started low and escalated slowly. In most people they fade. In a minority they do not, and they are the leading reason for discontinuation.

Less common but more serious effects have been quantified. In a population-based study of people using these drugs for weight loss rather than diabetes, the incidence of pancreatitis, bowel obstruction and gastroparesis was higher than with a non-GLP-1 comparator ^[14]. Gallbladder disease is also raised, largely because rapid weight loss of any cause raises it. These are uncommon events, and naming them is not the same as saying they are likely.

Anaesthesia has become a practical issue precisely because gastric emptying is slowed: residual stomach contents after the usual fasting period raise the risk of aspiration, and anaesthetic societies have issued guidance on how long to hold these drugs before a procedure. Anyone taking one should say so before any surgery or sedation.

On interactions, the important one is straightforward. A GLP-1 agonist on its own rarely causes hypoglycaemia, because its insulin effect is glucose-dependent. Combined with insulin or a sulfonylurea, which push insulin regardless of glucose, it very much can, and the doses of those agents usually need reducing. That is the flag the interactions and stacks engine raises. Slowed gastric emptying can also change the absorption rate of other oral drugs.

There are two formal contraindications worth knowing: a personal or family history of medullary thyroid carcinoma or multiple endocrine neoplasia type 2, based on C-cell tumours seen in rodents given these drugs. Whether that finding translates to humans has not been established, and the labelling is precautionary.

The table sets the effects out with the mechanism each one comes from, because almost every item on the list traces back to one of two things: slowed gastric emptying, or the rate of weight loss itself.

This is a young class carrying enormous expectations, so it is worth being explicit about the edges of the evidence.

Very long-term safety. The longest randomised exposure is a few years. These are drugs many people will take for decades. Nothing in the record suggests a problem at that horizon; equally, nothing in the record could yet detect one.

Whether the multi-agonists are better on outcomes, or only on weight. Tirzepatide and retatrutide clearly produce more weight loss than semaglutide ^[7][8]. Whether that converts into proportionally more cardiovascular or kidney protection is an open question with trials running.

The muscle question, properly measured. Nearly all of what is known about lean-mass loss comes from imaging that cannot see muscle specifically, and from trials not designed to answer it. Whether the lean-mass loss on these drugs is greater than, equal to, or smaller than the loss from an equivalent deficit achieved another way is not settled.

The addiction and compulsion signal. Plausible, actively studied, not established.

Compounded and grey-market copies. A great deal of what is sold online as semaglutide or tirzepatide is not a regulated product, and the failure modes there are the ordinary ones for the peptide market: wrong content, wrong strength, non-sterile preparation, and salt forms that are not the approved drug at all. The trial results on this page describe the regulated product, and nothing else.

The fair summary is that this is the best-evidenced class covered anywhere on this site, that its benefits extend well beyond the scale, and that it is a long-term treatment with a real side-effect profile rather than a shortcut. Not medical advice.