The Ozempic-for-kids problem nobody is asking out loud.

What actually changed in 2022 and 2023

In December 2022, the FDA expanded semaglutide 2.4 mg (Wegovy) labeling to include adolescents aged 12 years and older with obesity, on the strength of the STEP TEEN trial. In January 2023, the American Academy of Pediatrics released its first comprehensive Clinical Practice Guideline on the evaluation and treatment of children and adolescents with obesity. That guideline explicitly recommended offering pharmacotherapy as an adjunct to intensive health-behavior and lifestyle treatment for adolescents 12 and older with obesity, and recommended evaluating bariatric surgery for adolescents 13 and older with severe obesity.

Two things were true at the same time. Pediatric obesity is a real disease with documented long-term cardiometabolic, orthopedic, and psychosocial consequences, and the previous standard of care — lifestyle counseling alone — was failing the kids who needed help most. The 2023 guideline was, in part, an honest admission that "watch and wait" had not worked. The new tools work better than the old tools at moving weight. That is the part the guideline got right.

The part it skipped — and the part the public conversation has mostly skipped since — is whether moving weight is the right outcome measure in a growing adolescent at all, and what the body-composition cost of that weight loss is going to look like over decades the trials have not yet covered.

The STEP TEEN trial: what it showed and what it did not

STEP TEEN randomized 201 adolescents aged 12 to under 18 with obesity to weekly subcutaneous semaglutide 2.4 mg plus lifestyle intervention or placebo plus lifestyle intervention for 68 weeks. The headline result was a roughly 16% reduction in body mass index in the semaglutide arm versus an essentially flat trajectory in placebo — a between-group difference of about 16 percentage points of BMI change. About three-quarters of the semaglutide group achieved at least 5% body weight reduction.

That is a clinically meaningful, internally valid result. Cardiometabolic markers — waist circumference, lipids, glycemic measures, ALT — moved in favorable directions. Safety in the 68-week window was consistent with the adult Wegovy profile: gastrointestinal events dominated, serious adverse events were not increased meaningfully over placebo.

What STEP TEEN did not measure with the precision the question deserves is body composition. The trial reported BMI, body weight, and waist circumference. It did not include DEXA-grade fat-versus-lean separation as a primary or major secondary endpoint for the full cohort. The closest published sub-analysis suggested favorable changes in body composition broadly consistent with adult patterns, but the trial was not powered or designed to answer the question that matters most in a growing body: of the weight that left, how much of it was supposed to be there?

The AAP 2023 guideline and what it endorsed

The AAP's 2023 Clinical Practice Guideline is a 73-page document with 13 key action statements. The headline summary that made it into news coverage — "AAP endorses Ozempic and bariatric surgery for kids" — undersold the nuance. The guideline is clear that intensive health-behavior and lifestyle treatment (IHBLT) remains the foundation, that pharmacotherapy is an adjunct to IHBLT not a replacement, and that the decision belongs to the family and clinician together with informed-consent emphasis on the lifelong nature of obesity as a chronic disease.

It is also clear, in the body of the document, that the long-term safety database in adolescents is thin. The committee explicitly notes the absence of multi-year body-composition and bone-density data and recommends ongoing surveillance. That language did not survive the news cycle. The takeaway in the public conversation became "the AAP says this is fine." The takeaway in the guideline itself is closer to "the alternative is worse, the short-term data is acceptable, the long-term data does not yet exist, here is a framework for proceeding anyway."

That is a defensible clinical position. It is not a settled scientific one.

The adult body composition data is the warning the pediatric conversation is choosing not to look at directly. Twenty to forty percent of lost weight in adult GLP-1 trials is lean tissue. The growing body does not have spare lean tissue to give.

The lean-mass question nobody is answering

In the adult body-composition sub-analyses of GLP-1 and dual-agonist trials — semaglutide STEP, tirzepatide SURMOUNT — DEXA data consistently shows that roughly 20% to 40% of total weight lost is fat-free mass rather than fat mass. The exact proportion varies by trial, by analytic method, by baseline body composition, and by the rate of weight loss. The direction does not vary. Aggressive GLP-1-driven weight loss without a deliberate resistance training and protein protocol leaves a body with proportionally less lean tissue than it started with.

In a middle-aged adult with a settled adult body, that proportion is concerning but manageable. The lean-mass loss is mostly skeletal muscle, some bone mineral content, and connective tissue. In an adult with appropriate protein intake, resistance training, and a finite duration of treatment, much of it is recoverable when intake normalizes.

In a 13-year-old, the question is different. That adolescent is not redistributing a settled lean-mass compartment. They are building it. The window from roughly age 11 to 21 is when humans deposit most of their adult skeletal muscle, the architectural muscle attachments, and — critically — about 90% of their adult bone mass. That deposition runs on appetite and intake — the exact two levers a GLP-1 drug exists to pull down. A pharmacologic agent that suppresses appetite, suppresses gastric motility, and forces a sustained caloric deficit is operating directly on the inputs to that deposition program.

There is no adolescent DEXA trial that has run long enough at semaglutide therapeutic doses to tell us what that does to peak lean mass acquisition. There is no adolescent bone-density trial that has run long enough to tell us what it does to peak bone mass. The conversation in the popular media — and in some of the pediatric endocrinology coverage — has treated the absence of red flags in 68 weeks of STEP TEEN as evidence of safety. The absence of red flags in 68 weeks is not the same as evidence of long-term safety in a tissue compartment that takes a decade to accrue.

Peak bone mass and a one-shot biology window

The most consequential one-shot biological process in the second decade of life is bone mass acquisition. Roughly 26% of adult total bone mineral content is acquired during the two years of peak skeletal growth (around age 12 to 14 in girls, age 14 to 16 in boys), and approximately 90% of peak adult bone mass is in place by the late teens. The remaining gain happens slowly through the early twenties. After the late twenties, bone mass plateaus and then starts the long lifetime decline.

Peak bone mass is the single strongest predictor of osteoporosis risk in the seventh and eighth decades of life. A 10% reduction in peak adult bone mass is estimated to increase lifetime osteoporotic fracture risk by 50% or more. The window for accruing that mass does not reopen. You either build the bone in your second decade or you do not, and the choice is essentially final by age 25.

Bone deposition during that window is mechanically coupled (loading from physical activity), nutritionally coupled (protein, calcium, vitamin D, adequate caloric intake), and hormonally coupled (growth hormone pulsatility, IGF-1, sex steroids during puberty). A drug that reduces appetite by 35% or more, reduces caloric intake correspondingly, and shifts body weight downward — all during the active deposition phase — is intervening on multiple inputs to the deposition program simultaneously.

Could that be neutral or favorable for bone? Possibly. Less mechanical loading from less body weight might be offset by better metabolic health and less inflammatory drag on bone formation. We do not know, because nobody has run the multi-year DEXA trial that would tell us. The honest answer is "we are intervening on a window of biology that does not reopen, with a drug whose body-composition effects in adults are known to be lean-mass-costly, and we are calling 68 weeks of normal-looking labs a green light."

The duration-of-treatment question

Then there is duration. Adult GLP-1 trials have made it clear that semaglutide is, functionally, a chronic medication. Weight rebounds when it is stopped. The STEP 1 extension data showed that participants regained roughly two-thirds of lost weight in the year after discontinuation. The implicit framing in adult prescribing has shifted toward indefinite use — "obesity is a chronic disease and chronic diseases need chronic treatment."

Applied to a 13-year-old, that framing becomes "this patient will likely be on a GLP-1 agonist for 60 to 70 years." We have no human data — none — on the cardiovascular, oncologic, pancreatic, thyroid, gallbladder, or psychiatric profile of GLP-1 receptor agonism over a seven-decade horizon. The longest controlled adult exposure data is about five years. The longest adolescent exposure data is the open-label extension of STEP TEEN.

That is not, on its own, a reason not to prescribe. There are plenty of medications whose long-horizon data does not exist when they enter clinical use — statins were prescribed for decades before we had clean multi-decade follow-up. But it is a reason to be explicit, with families, about what the unknowns actually are, and a reason for the prescribing decision to weigh those unknowns against the alternative of severe pediatric obesity with its own well-documented multi-decade consequences. The framing that makes the prescribing decision look obvious in either direction is the dishonest framing.

An honest position

The honest read on the pediatric GLP-1 question, as of 2026, is this:

Severe pediatric obesity is a serious disease. Doing nothing has real, documented costs. The STEP TEEN data is internally valid and the weight outcomes are clinically meaningful. The AAP guideline framework — pharmacotherapy as an adjunct to intensive lifestyle treatment, with informed consent and shared decision-making — is a reasonable clinical position. None of that is in dispute here.

What is in dispute, and what the conversation should be including but mostly is not, is the body-composition question. A drug that costs adults 20-40% of weight loss as lean mass is being prescribed to bodies that are mid-deposition of their entire adult skeletal and muscular architecture. The protein, resistance training, and DEXA monitoring protocols that adult body-composition specialists routinely build around GLP-1 prescribing have not been standardized into pediatric care. They should be.

If a 13-year-old goes on semaglutide for severe obesity, the minimum responsible package should include: a structured resistance training program from the first week, a protein intake target appropriate to growth (commonly cited as ~1.2-1.6 g/kg body weight in adolescents on caloric deficit), serial DEXA monitoring of bone and lean mass at baseline and at 6-12 month intervals, vitamin D and calcium status checks, and an explicit conversation with the family about peak bone mass acquisition and the multi-year nature of the commitment. None of that is in the FDA label. Almost none of it is in routine pediatric prescribing. That gap is the part the conversation should be having.

What this article is not saying

This is not an argument against GLP-1 prescribing in adolescents. The drugs work, the trials are real, the alternative of untreated severe pediatric obesity has its own multi-decade costs that are well documented. Anyone reading this as "Wellness Radar says don't give kids Ozempic" is reading it wrong.

This is also not an argument from precautionary anti-pharmaceutical reflex. Adolescent biology is not a museum piece that needs to be preserved from medication. It is a developmental window that responds to inputs, and intervening on the inputs requires understanding what is being deposited during that window and what the cost of the intervention is in that compartment. The argument is for that understanding to be part of the prescribing conversation, not absent from it.

Finally, this is not a critique of the families and physicians making these decisions today. The data they have to work with is what exists. The critique is of a public conversation that has skipped the body-composition question entirely — that has treated FDA approval and AAP endorsement as the end of the discussion when, on the parts of the biology that matter most for these specific patients over the next sixty years, the discussion has barely started.