A triple dose of semaglutide bought about three more percentage points

The obvious question about a drug that works is whether more of it works better. For once-weekly semaglutide, approved for weight management at 2.4 mg, two phase 3b trials published together on September 14 give an answer that is affirmative and smaller than the dose increase might suggest [s1][s2].

Tripling the maintenance dose to 7.2 mg produced roughly three additional percentage points of weight loss in adults with obesity and no diabetes, and it did so while gastrointestinal adverse events and a sensory side-effect called dysaesthesia became substantially more common [s1].

The two trials

STEP UP enrolled 1,407 adults with a BMI of 30 kg/m² or greater and no diabetes across 95 hospitals, specialist clinics, and medical centres in 11 countries, randomising them 5:1:1 to semaglutide 7.2 mg, semaglutide 2.4 mg, or placebo, each alongside a lifestyle intervention, for 72 weeks [s1]. Mean baseline bodyweight was 113.0 kg and mean BMI 39.9 kg/m²; 73.7% of participants were female and mean age was 47 [s1].

STEP UP T2D ran the same three-arm design in 512 adults who had both obesity and type 2 diabetes, with a BMI of 30.0 kg/m² or greater and HbA1c between 7.0% and 10.0%, at 68 sites in Bulgaria, Canada, Hungary, Poland, Portugal, Slovakia, South Africa, and the USA [s2]. Mean bodyweight was 110.1 kg, mean BMI 38.6 kg/m², mean HbA1c 8.1% [s2].

What the higher dose added

In STEP UP, mean bodyweight change at 72 weeks was −18.7% with 7.2 mg versus −15.6% with 2.4 mg, an estimated treatment difference of −3.1% (95% CI −4.7 to −1.6) [s1]. Against placebo, which produced −3.9%, the difference was −14.8% (95% CI −16.2 to −13.4) [s1].

The clearer signal is at the extremes of response. Compared with the 2.4 mg group, participants on 7.2 mg were more likely to lose 20% or more of bodyweight (odds ratio 1.8, 95% CI 1.3 to 2.4) and 25% or more (2.4, 95% CI 1.6 to 3.5) [s1]. Against placebo the odds ratio for reaching a 25% reduction was 127.4 (95% CI 36.8 to 441.4) — a number whose enormous confidence interval reflects how rarely placebo participants got there, not precision [s1]. Waist circumference fell by 11.7 cm more than placebo [s1].

In the diabetes trial the numbers are smaller across the board, as they usually are when GLP-1-based drugs are tested in people with type 2 diabetes. Mean bodyweight change was −13.2% with 7.2 mg versus −3.9% with placebo, an estimated treatment difference of −9.3% (95% CI −11.0 to −7.7) [s2]. Waist circumference fell 6.5 cm more than placebo [s2].

The cost side of the ledger

Gastrointestinal adverse events were reported by 711 of 1,004 participants (70.8%) on 7.2 mg, compared with 123 of 201 (61.2%) on 2.4 mg and 86 of 201 (42.8%) on placebo [s1].

Dysaesthesia — abnormal or unpleasant sensation — is the finding that stands out, because it scaled sharply with dose: 230 participants (22.9%) on 7.2 mg, 12 (6.0%) on 2.4 mg, and one (0.5%) on placebo [s1]. That is close to a fourfold increase between the two active doses.

Serious adverse events were reported in 68 of 1,004 (6.8%) on 7.2 mg, 22 of 201 (10.9%) on 2.4 mg, and 11 of 201 (5.5%) on placebo [s1]. The serious-event rate was therefore not higher on the larger dose in this trial, though with 201 participants in the 2.4 mg arm the comparison is imprecise. The trial investigators' own summary describes the 7.2 mg dose as retaining a favourable risk–benefit profile [s1].

How to read a three-point difference

Three percentage points on top of 15.6% is a real effect and a much smaller one than the headline dose change. Whether it matters clinically depends on who is being treated. For someone whose weight loss on the approved dose has stalled short of a therapeutic goal — the population the trials were explicitly designed around [s1][s2] — a further increment may change what the treatment achieves. For a person already responding well, the trade against a more than threefold rise in dysaesthesia and a roughly ten-point rise in gastrointestinal events looks different.

Both trials ran for 72 weeks, which is long by obesity-trial standards and short relative to how long these drugs are taken. Neither was designed to measure cardiovascular events, and neither reports what happens after the drug stops. Both were funded by Novo Nordisk, the manufacturer [s1].

The trials also do not answer the question a clinician actually faces, which is how a higher dose of semaglutide compares with switching to a different agent. STEP UP had no active comparator other than the lower dose of the same drug [s1].

What to watch

Whether regulators accept a 7.2 mg maintenance dose, and on what label; whether the dysaesthesia signal persists and is characterised further in longer follow-up; and whether the extra weight loss translates into differences in the outcomes that motivate obesity treatment — cardiovascular events, diabetes incidence, and function — rather than only in the number on the scale. An accompanying commentary published alongside the trials framed the results as another step towards addressing obesity-mediated disease [s3].

This article describes results from two randomised trials and is informational only. It is not medical advice and does not recommend any drug or dose. Semaglutide dosing is a matter for a prescribing clinician.

Sources

• [s1] Wharton S, Freitas P, Hjelmesæth J, Kabisch M, Kandler K, Lingvay I, Quiroga M, Rosenstock J, Garvey WT, STEP UP trial group, Once-weekly semaglutide 7·2 mg in adults with obesity (STEP UP): a randomised, controlled, phase 3b trial, The Lancet Diabetes & Endocrinology, 2025;13:949-963, published online 2025-09-14.
• [s2] Lingvay I, Bergenheim SJ, Buse JB, Freitas P, Garvey WT, Harder-Lauridsen NM, Rosenstock J, Sahu K, Wharton S, STEP UP T2D trial group, Once-weekly semaglutide 7·2 mg in adults with obesity and type 2 diabetes (STEP UP T2D): a randomised, controlled, phase 3b trial, The Lancet Diabetes & Endocrinology, 2025;13:935-948, published online 2025-09-14.
• [s3] Sattar N, Lean MEJ, Semaglutide 7·2 mg: another step towards tackling diseases mediated by obesity, The Lancet Diabetes & Endocrinology, 2025;13:901-903, published online 2025-09-14.