In adolescent obesity, medication and lifestyle treatment are not alternatives

Design

Systematic review and random-effects network meta-analysis of randomised trials, search to June 2025

Population

3835 adolescents aged 10 to 19 with obesity across 42 trials, median age 14.5 years

Primary outcome

BMI and BMI z score, with waist circumference, fat mass and lean mass secondary

Effect

Semaglutide plus counselling BMI -8.31 (95 per cent CI -12.33 to -4.28); lifestyle treatment alone BMI -3.85 (-4.91 to -2.80); all drugs better combined than alone

Forty-two randomised trials and 3835 adolescents with obesity, median age 14.5 years and 59 per cent female, were combined in a network meta-analysis comparing structured lifestyle treatment, glucagon-like peptide-1 receptor agonists, metformin, orlistat, phentermine-topiramate, and combinations of drug and lifestyle treatment.

Every drug performed better paired with lifestyle treatment than alone, and combination treatments led on every adiposity outcome. Semaglutide with counselling produced the largest reduction in BMI, a mean difference of -8.31 (95 per cent CI -12.33 to -4.28) and in BMI z score of -1.80 (-2.39 to -1.21), though the authors flag this rests on few studies. Lifestyle treatment on its own was far from negligible: BMI -3.85 (-4.91 to -2.80) and BMI z score -0.89 (-1.17 to -0.61) against control.

The framing the authors reach is the useful one. Pharmacotherapy is not an adjunct bolted onto a failing lifestyle programme, and lifestyle treatment is not a preliminary hurdle to clear before drugs are permitted; the two are components of the same treatment and perform best together. The horizon, though, is short - typically 6 to 12 months - and adolescence is not a 12-month problem. Nothing here speaks to what happens when the drug stops, which is the question a parent will ask first.

• Offer structured lifestyle treatment alongside any medication, not as a precondition for it.
• Quote the BMI z score change rather than raw BMI when talking to families about a growing adolescent.
• Set expectations against a 6-to-12-month evidence horizon and say that longer-term data are absent.
• Discuss what happens on stopping before starting; the trials do not answer it.
• Note that the largest effect estimate rests on a small number of studies with a wide interval.

The statistics, in plain English

A BMI reduction of 8.31 units is very large for an adolescent, and an interval running from -12.33 to -4.28 tells you the estimate is imprecise as well as impressive; it comes from few trials, so treat it as promising rather than established. Network meta-analysis compares treatments that were often never tested against each other, which adds assumptions on top of the usual pooling. The follow-up is short across the whole network, so none of these estimates describes maintained weight loss.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.