Semaglutide: what the big trials actually showed

Semaglutide is the best-studied drug on this site by a wide margin. In people without diabetes, a weekly 2.4 mg injection produced an average 14.9% weight loss over 68 weeks against 2.4% on placebo [1], and separate outcome trials involving more than 20,000 people found fewer heart attacks, strokes and kidney failures [5–7]. The costs are real too: most people get stomach side effects, gallbladder problems are more common, and most of the lost weight comes back within a year of stopping [4].

What it is

Semaglutide is a modified copy of GLP-1 (glucagon-like peptide-1), a gut hormone released after meals that nudges insulin up, glucagon down and appetite down. Natural GLP-1 lasts only minutes in the blood. Novo Nordisk's chemists swapped one amino acid to stop the enzyme that chews it up, and attached a fatty side chain that lets the molecule ride on albumin in the bloodstream, which stretches its half-life to about a week [10].

It is sold as Ozempic (weekly injection for type 2 diabetes), Rybelsus (daily tablet for type 2 diabetes) and Wegovy (weekly injection, and now a daily pill, for weight management). The drug in each is the same molecule; the doses and licensed uses differ.

How it cuts weight is only partly settled. The accepted explanation is that it acts on GLP-1 receptors in the brainstem and hypothalamus to reduce hunger and increase fullness, and it slows stomach emptying, particularly early in treatment [10]. People eat less. The trials measured weight and events, not the mechanism, so treat the brain-signalling story as well-supported theory rather than something each trial proved.

The human evidence

This is where semaglutide differs from almost everything else HPR covers. The evidence comes from multi-thousand-person, double-blind, placebo-controlled trials published in the leading medical journals, funded by the manufacturer, and designed to satisfy regulators.

Weight loss: the STEP trials

STEP 1 enrolled 1,961 adults with a BMI of 30 or more (or 27 with a weight-related condition) and no diabetes, and randomised them 2:1 to weekly semaglutide 2.4 mg or placebo, both with lifestyle counselling [1].

Half of the semaglutide group lost at least 15% of their starting weight [1]. That figure uses an analysis that counts people who stopped the drug, so it's closer to what a real clinic would see than a best-case number.

People with type 2 diabetes lose less. In STEP 2 (n = 1,210), the same 2.4 mg dose produced 9.6% weight loss against 3.4% on placebo over 68 weeks [2]. Nobody knows exactly why diabetes blunts the response, but it has shown up with every drug in this class.

What happens when you stop

Two trials answer this directly. STEP 4 gave 803 people 20 weeks of semaglutide, then randomly switched a third of them to placebo. Over the next 48 weeks, those who continued lost a further 7.9% while those switched to placebo regained 6.9% [3]. In the STEP 1 extension, 327 people were followed for a year after treatment ended: they regained about two-thirds of the weight they had lost, and blood pressure, blood sugar and cholesterol improvements drifted back towards baseline [4]. The authors' conclusion was that obesity behaves like a chronic condition and the drug works only while you take it.

Heart and kidney outcomes

SUSTAIN-6 was the first. In 3,297 people with type 2 diabetes at high cardiovascular risk, semaglutide (0.5 or 1.0 mg weekly) cut the combined rate of cardiovascular death, heart attack and stroke from 8.9% to 6.6% over two years (hazard ratio 0.74) [5]. It was designed to show the drug was not harmful, and it did more than that.

SELECT then asked the question for people without diabetes. It enrolled 17,604 adults aged 45 or older with established cardiovascular disease and a BMI of at least 27 [6].

A 20% relative reduction is meaningful, though the absolute gap is 1.5 percentage points over about three years. Also notice the dropout line: twice as many people stopped semaglutide because of side effects as stopped placebo [6].

FLOW studied 3,533 people with type 2 diabetes and chronic kidney disease on semaglutide 1 mg. The trial was stopped early for benefit. Major kidney disease events, which included kidney failure, a large fall in kidney function and kidney or cardiovascular death, were 24% lower on semaglutide (hazard ratio 0.76), and death from any cause was 20% lower [7].

The lean-mass question

When anyone loses a lot of weight, some of it is lean tissue, and the worry with GLP-1 drugs is that the share is unusually high. A 2024 review of the body-composition data found the numbers all over the place: in some studies lean mass accounted for 40–60% of total weight lost, in others around 15% or less [8]. The reviewers pointed out that "lean mass" on a DXA scan includes organs, bone and water as well as muscle, so a drop in lean mass is not the same thing as a drop in muscle, let alone strength. Whether the lean tissue lost on semaglutide matters for function, especially in older people, is still an open question [8].

What the animal and lab work suggests

Because the human data are so extensive, animal studies matter less here than for most peptides. The one animal finding that still shapes the label is from rodents: semaglutide caused thyroid C-cell tumours in rats and mice, which is why US labelling carries a boxed warning and why the drug is not used in people with a personal or family history of medullary thyroid cancer [10]. Whether the finding applies to humans has not been established [10].

Safety and side effects

Stomach and gut. Nausea and diarrhoea were the most common side effects in STEP 1, usually mild to moderate and fading with time [1]. In STEP 2, gastrointestinal events affected 63.5% on 2.4 mg against 34.3% on placebo [2].

Gallbladder. In the Wegovy trials, gallstones were reported in 1.6% on semaglutide against 0.7% on placebo, and gallbladder inflammation in 0.6% against 0.2% [10]. A meta-analysis of 76 randomised trials covering the whole GLP-1 class found a 37% higher relative risk of gallbladder or bile-duct disease, rising to more than double in trials using weight-loss doses [9]. Rapid weight loss raises gallstone risk by itself, but the FDA label notes the excess persisted after allowing for the amount of weight lost [10].

Pancreatitis. The signal is small but on the label. Across the Wegovy trials, acute pancreatitis was confirmed in 4 people on semaglutide (0.2 cases per 100 patient-years) against 1 on placebo [10]. The label says to stop the drug if pancreatitis is suspected and not restart it if confirmed.

Eyes. SUSTAIN-6 found more diabetic retinopathy complications on semaglutide (hazard ratio 1.76), including vitreous haemorrhage and cases needing laser or injection treatment [5]. The leading explanation is that a rapid fall in blood sugar can temporarily worsen existing retinopathy, which is why people with diabetic eye disease are monitored. Separately, in June 2025 the European Medicines Agency concluded that NAION, a type of sudden optic-nerve damage, is a very rare side effect affecting up to 1 in 10,000 people [11].

Stopping. Across SELECT, 16.6% of people stopped the drug because of adverse events [6]. That is a large number for a trial lasting three years, and worth knowing before anyone calls semaglutide easy to take.

Legal and regulatory status

Semaglutide is a licensed prescription medicine in both the UK and US.

NICE also advises considering stopping if less than 5% of starting weight has been lost after 6 months [15]. Ozempic and Rybelsus are separately licensed for type 2 diabetes. Anything sold online as "research semaglutide" is not the licensed medicine.

What we still don’t know

• Whether long-term use (beyond four to five years) keeps its benefits and safety profile, since few people have been followed that long in controlled trials.
• How much of the lean mass lost is functional muscle, and whether resistance training or protein intake changes the picture meaningfully [8].
• The best way to come off the drug, if at all, given how much weight returns after stopping [3,4].
• Why people with type 2 diabetes lose roughly a third less weight than people without it [1,2].
• How the rare eye signals translate to individual risk, particularly for people without diabetes [5,11].

Educational content only — not medical advice. Many peptides discussed on HPR are not approved for human use. Talk to a qualified clinician before making any decision about your health.

Free live webinar · Thursday 1 October · 7pm UK time

Join the next webinar: Introduction to Peptides

with Mitch O’Callaghan, ISSCA Peptide Therapist · Founder, Proper Protocols

Save my free placeHPR×Proper Protocols