Semaglutide in Children Under 12: Reading STEP Young's 40.4%

On September 7, Novo Nordisk announced the first results of STEP Young, a Phase III trial of once weekly semaglutide in children aged 6 to under 12 with obesity. The trial randomized 165 children, and the company says it met its primary endpoint: a larger reduction in body mass index (BMI) at week 68 than placebo. The number that traveled was 40.4%. That share of the semaglutide group was no longer classified as having obesity at week 68, compared with none of the children on placebo. The results are topline only and have not been peer reviewed.

What the trial tested

STEP Young is a randomized, double blind, placebo controlled, multinational trial. Both groups received lifestyle modification throughout, including a reduced calorie diet and more physical activity. Children received a maximum dose of either 1.7 mg or 2.4 mg of semaglutide, chosen by their body weight at the start. More than 85% of the children had class II or class III severe obesity at baseline. In children, obesity means a BMI at or above the 95th percentile for age and sex on CDC growth charts, class II means 120% or more of that threshold and class III means 140% or more. Novo says the trial is being run to meet post marketing requirements from regulators, which explains why it follows the adult and adolescent programs.

The primary endpoint was the percent change in BMI from baseline to week 68. A confirmatory secondary endpoint was improvement in BMI classification at week 68, and that is where the 40.4% comes from. Supportive secondary endpoints include changes in cardiovascular risk factors, glucose metabolism and body composition in a DXA subgroup, measured at week 68 and week 104.

How to read 40.4% against zero

Two details change the meaning. First, a footnote in the release says the figure is based on the trial product estimand, meaning the treatment effect if all children adhered to treatment. That answers a different question from the one parents ask, which is what happens to children who start the drug and sometimes miss doses. The release does not give the result under the stricter approach that counts everyone as randomized, and that figure can be lower.

Second, "no longer classified as having obesity" is a threshold crossing, not an average. A child who falls from the 96th BMI percentile to the 94th crosses it, and a child who drops sharply but stays above the line does not. Novo says the 40.4% means two in five children moved to a normal weight or overweight category. Given that most children started with severe obesity, that is a substantial shift. But it tells us nothing about the average change in BMI, which is the primary endpoint, and the release does not report that number.

The placebo comparison deserves a comment too. Zero of the children in the placebo group crossed the line, and the company says both groups got lifestyle support. Yoni Freedhoff of the University of Ottawa, who wrote a commentary on the results, called the placebo arm a reflection of decades of failure for lifestyle treatment alone in this population. It is a fair reading, but it is also the view of a commentator who discloses a clinical grant from Novo Nordisk to his company, and the underlying numbers are not yet public.

The argument that is already forming

The company frames the need in large numbers: it cites an estimated 177 million children aged 5 to 19 living with obesity in 2025, rising to 228 million by 2040. Whatever the exact figures, the clinical point is that the options for young children with severe obesity are limited. Freedhoff's column, published September 23, argued that effective treatment would improve quality of life and cut medical risk at once, and he asked whether it will trigger a new moral panic. He quotes a critic who wrote that prescribing weight loss for six year olds "is not, and never has been, about health," even if the drug works.

That objection is not about this trial's data, and it deserves a straight answer rather than dismissal. The relevant questions are empirical. Do the medicines affect growth and puberty over years? Do they change a child's relationship with food? What happens when treatment stops? STEP Young cannot answer the second or third, and the topline answers the first only in a sentence. Freedhoff himself says careful study of growth, puberty and psychological effects is needed. In adults, the benefit that mattered most surfaced in a trial that followed people for a mean of 39.8 months, as we explain in our piece on semaglutide's heart benefit.

The hard part of this story is that both camps will find support in it. The result is large enough that it would be negligent to ignore, and incomplete enough that it would be careless to act on it as settled. I would wait for the November presentation and look for four things: the average BMI change, the analysis counting every randomized child, the discontinuation and adverse event rates by age, and any measured change in height velocity. If those hold up, the debate will shift from whether the drug works in young children to who should get it, for how long, and what happens after. For the checks that apply to any trial like this, see our five minute guide to reading a study.