STEP 3: semaglutide with intensive behavioural therapy

STEP 3 asks whether semaglutide still adds something when the lifestyle programme is already intensive. The answer in the paper is yes, and the size of the placebo response is the most instructive part.

What the trial asked

Whether semaglutide 2.4 mg, as an adjunct to intensive behavioural therapy with an initial low-calorie diet, produces more weight loss than placebo with the same programme.

Design

Randomised, double-blind, parallel-group, phase 3a, 68 weeks, at 41 US sites between August 2018 and April 2020. Randomised 2:1 to semaglutide 2.4 mg (n=407) or placebo (n=204). Both arms had a low-calorie diet for the first 8 weeks and intensive behavioural therapy, defined as 30 counselling visits, across the 68 weeks.

Who was enrolled

611 adults without diabetes, with either overweight (BMI 27 or higher) plus at least one comorbidity, or obesity (BMI 30 or higher). 495 (81.0 percent) were women. Mean age 46 years (SD 13), mean weight 105.8 kg (SD 22.9), mean BMI 38.0 (SD 6.7). 567 (92.8 percent) completed the trial and 505 (82.7 percent) were still on treatment at the end.

Primary endpoint and result

Coprimary endpoints at week 68: percent change in body weight, and loss of 5 percent or more of baseline weight.

Difference -10.3 percentage points (95% CI -12.0 to -8.6; P less than 0.001). Loss of 5 percent or more: 86.6 versus 47.6 percent (P less than 0.001).

Key secondary results

Confirmatory secondary endpoints were losses of at least 10 and 15 percent: 75.3 versus 27.0 percent, and 55.8 versus 13.2 percent (P less than 0.001 for both).

Adverse events and discontinuation

Gastrointestinal adverse events: 82.8 percent with semaglutide versus 63.2 percent with placebo. Treatment was discontinued because of these events in 3.4 percent of semaglutide participants and 0 percent of placebo participants.

Funding and registration

The abstract does not name the funder. ClinicalTrials.gov NCT03611582.

What this trial does not show

• The placebo arm lost 5.7 percent, more than double the placebo loss in STEP 1. That is the behavioural programme at work. The semaglutide arm's -16.0 percent is not much larger than STEP 1's -14.9 percent, so the placebo-subtracted difference is smaller here (-10.3 versus -12.4 points). The trial does not tell you whether intensive therapy adds to the drug for an individual; it tells you both arms received it.
• 30 counselling visits is more than most clinical programmes provide. The result may not transfer to lighter-touch care.
• A single US-only population, 81 percent women.
• The authors state that further research is needed on durability; the trial stops at week 68.

Where the numbers come from

PubMed abstract of the primary paper (PubMed 33625476), fetched 4 September 2026. The tirzepatide trial built around a lifestyle programme, with the programme before randomisation rather than alongside it, is SURMOUNT-3.