How WHO Graded Its Child Obesity Advice: Strong Rules, Thin Evidence
WHO's first child and adolescent obesity guidelines pair firm advice with thin proof. Here are the grades behind each rule, the conditions on teen drug treatment, and questions to bring to a pediatrician.
On Wednesday, Oct. 7, the World Health Organization released two guidelines on treating obesity in young people, and in both it calls its core advice "strong" while grading the evidence beneath it "very low." The guideline itself names the mismatch: "discord between the quality of evidence and the corresponding recommendation strength."
On Oct. 7, 2026, WHO recommended against weight-loss drugs, bariatric (weight-loss) surgery and weight-loss devices for children aged 0 to 9, and it graded that a strong recommendation. For ages 10 to 19, WHO says medicines may be considered only for a young person who has obesity and an obesity-related complication, after at least six months of supervised lifestyle care has failed.
WHO's release says 170 million people aged 5 to 19 were living with obesity in 2024, including 70 million aged 5 to 9 and 100 million aged 10 to 19. Prevalence in that age group has quadrupled since 1990, rising from 2% to 8%.
The guidelines come from two separate expert panels, one for children from 29 days to under 10 years and one for adolescents. Some outlets, including CBC and Reuters, noted thin evidence on the drugs, but none of the CBC, Reuters, BBC, Health Policy Watch, Medical Daily or Nairametrics reports set out the certainty grade behind each rule. Strength says how firmly a panel advises acting. Certainty says how far the numbers behind the advice can be trusted. WHO defines the lowest grade this way:
"Very low certainty: We have very little confidence in the effect estimate. The true effect is likely to be substantially different from the effect."
WHO, the children's guideline, Annex 2
A "strong" grade, in WHO's handbook, means confidence that the good effects outweigh the bad: "Most people would want the recommended intervention, and only a small proportion would not." A "conditional" grade means less certainty on that balance. The panels voted for strong recommendations on diet, activity and behaviour support because of "the practicality and universality of adopting interventions," not because trials were convincing. WHO's reasoning for the diet rule: "While benefits are small, so is the risk of adverse events."
Every grade, side by side:
| Treatment | Ages 0–9 | Ages 10–19 |
|---|---|---|
| Structured diet plan | Recommended (strong; very low certainty), over age 2 | Recommended (strong; very low) |
| Physical activity plan | Recommended (strong; very low) | Recommended (strong; very low) |
| Behaviour-changing support | Recommended (strong; very low) | Recommended (strong; very low) |
| Combined programme | Recommended (strong; very low) | Recommended (strong; very low) |
| Digital tools, parent-supervised | Suggested (conditional; very low), over age 2 | Suggested (conditional; very low) |
| Weight-loss drugs | Not recommended (strong; low) | May be considered (conditional; low), under set conditions |
| Bariatric surgery | Not recommended (strong; no evidence) | May be considered (conditional; low), severe obesity only |
| Weight-loss devices | Not recommended (strong; no evidence) | No recommendation (lack of evidence) |
Every strong lifestyle recommendation WHO makes is graded "very low." The firmest rules are not the best-proven ones.
A strong rule can rest on almost no evidence
The clearest cases are surgery and devices, where WHO graded the evidence "no evidence." The under-10 drug ban rests on low-certainty evidence, and almost none for the youngest. WHO commissioned a Cochrane review of drug trials and the guideline reports that it "did not identify any studies that included children under 6 years of age." Eight studies included children 6 or older; seven tested metformin and one tested liraglutide. WHO says the recommendation accounts for unknown effects on "normal growth, development, physiological parameters, and mental health."
For surgery and devices the panel said so plainly. It "agreed on a strong recommendation without relevant evidence," because surgery at such a young age "may result in significant harm."
So a strong "no" here is a precaution, not a finding that drugs hurt young children. Laurence Grummer-Strawn, a WHO nutrition official, said, as the BBC reported, where the worry lies:
"We are particularly concerned in this age group, if you are starting children very early on to using a GLP-1 or drugs like this, when does that stop?"
Laurence Grummer-Strawn, WHO nutrition official, as reported by the BBC
He added that "we don't really have evidence that you can use this as a temporary treatment and then turn to something dietary later on in life." GLP-1 drugs are weight-loss drugs such as the brands Wegovy and Saxenda.
Nor is it a wall against research. Dr. Luz María De Regil, who directs WHO's Department of Nutrition and Food Safety, said at a news conference that "it is clear that the research is going in this area," as CBC reported.
Lifestyle care works a little, and WHO says so
Lifestyle care is where both panels agreed, so it deserves an honest look at what it delivers. For ages 0 to 9, WHO's commissioned review found that healthcare-based programmes "may result in little to no difference in BMI z-scores at 12 months follow-up": -0.15 (95% range -0.23 to -0.06) across 16 studies, with very low certainty. A BMI z-score compares a child's BMI with other children of the same age and sex, so it cannot be set beside the drug numbers below.
Community programmes in schools, churches and community centres showed 0.02 (range -0.08 to 0.12) at 12 months, with moderate certainty. That range straddles zero. WHO adds that "information on adverse events was scant."
On physical activity the guideline says "no clear effect" was found on BMI, weight loss or body fat, yet "although small, the benefits of physical activity are clear for all ages and health statuses." Small benefit, small risk, open to nearly everyone: that is the whole case, and it is a fair one.
In practice, the guidelines are more concrete than "eat well and move." A structured diet plan, for children older than 2, should be supervised by "health workers with knowledge on nutrition," involve parents or caregivers, and include a check for "risk factors for eating disorders." Staff should be "non-stigmatizing." For ages 5 and over, WHO restates an average of at least 60 minutes a day of moderate to vigorous activity and a two-hour daily maximum for recreational screen time.
Cost is the catch: WHO acknowledges that "financial constraints may hinder parents from affording healthy food or meal plans." The guidelines are advice for health systems, not law.
What a teenager has to clear before a drug is on the table
For ages 10 to 19, the adolescent guideline suggests drug treatment only for young people who have obesity and obesity-related complications, "only after" a structured, supervised lifestyle programme "has failed." The footnotes carry the real conditions. Any prescription should come in a specialized multidisciplinary setting, after close evaluation of physical development and psychological stability, with a structured long-term follow-up.
The grade is conditional, with low certainty. And the recommendation names no drug. Health Policy Watch reported that the guidelines "don't focus on or recommend a specific weight loss drug," reasoning it attributed to Grummer-Strawn. WHO's adult guideline of Dec. 1, 2025, about 10 months earlier, named three agents: liraglutide, semaglutide and tirzepatide. Headlines about "GLP-1s for teens" are an interpretation; the text speaks of pharmacological interventions as a class.
The evidence behind it is a review of 37 randomized trials with 4,218 participants. Seventeen ran in high-income countries. Follow-up lasted six to 31 months, "with a median of 11 months," and every drug was given on top of lifestyle care.
Pooled against placebo, drug users ended 5.47 kilograms lighter on average, about 12 pounds (range 7.7 to 16.4 pounds, low certainty). Any adverse event struck 846 per 1,000 young people on placebo and 872 per 1,000 on a drug: 26 more, with a range from none to 60 more. Stopping treatment because of side effects is in the chart.
A relative risk of stopping of 1.70 sounds large. In absolute terms it is 17 more young people per 1,000, about 1 in 59, and the range (0.90 to 3.22) includes no difference at all.
Treat all of these as averages of different medicines. Metformin is the largest group in the pool, and WHO's panel said benefits "varied according to medication type." The pool cannot tell a family how any single drug performs. In one 2024 trial of liraglutide in children 6 to under 12, adverse events occurred in 50 of 56 children (89%) on the drug and 23 of 26 (89%) on placebo.
The panel also flagged practicalities: injections "are viewed less favourably than oral medications," and access "may be hampered due to cost, lack of coverage" by insurance, or logistics such as refrigeration. Surgery is a narrower door: WHO says it may be considered for severe obesity, meaning BMI-for-age more than four standard deviations above the median, or more than three plus a complication. Evidence is low certainty, and WHO's children's guideline notes that "data are lacking on long-term effectiveness."
Three guidelines, three thresholds
WHO is not alone, and the others disagree with it. The Associated Press reported on Jan. 9, 2023 that under American Academy of Pediatrics guidance, doctors should offer adolescents 12 and older who have obesity access to appropriate drugs. Co-author Dr. Ihuoma Eneli of Nationwide Children's Hospital, as the Associated Press reported:
"Waiting doesn't work."
Dr. Ihuoma Eneli, co-author of the American Academy of Pediatrics guidance
Canada's pediatric guidance, published in CMAJ, recommends GLP-1 use only after age 12, in CBC's words, and rejects the sequence WHO sets:
"There is no evidence to support a stepwise approach whereby pharmacologic and surgical interventions should be offered only if behavioural and psychological interventions prove ineffective."
Canadian pediatric obesity guidance published in CMAJ, as quoted by CBC
Set side by side, the same evidence yields three thresholds. WHO asks for at least six months of supervised lifestyle care plus a complication before drugs at 10 to 19. The AAP offers drugs from 12. Canada also starts at 12, with no required lifestyle-first step. That comparison is an editorial reading, not WHO's, and the gap is one of judgement under thin data. For a related sequencing question on the coverage side, see this explainer on fail-first rules.
Use is already rising. A study in Pediatrics from NYU Langone, published Sept. 4, 2026, found 20,282 children aged 8 to 11 prescribed GLP-1 drugs since 2019, which is 0.6% of more than 3.5 million children with obesity in the data. It shows who got the drugs, not whether they were safe.
For a family heading into an appointment, the guidelines' own conditions suggest questions. These questions are a synthesis of the guidelines' conditions, not WHO text, and decisions belong with the child's clinician:
- Where is my child on the BMI-for-age chart, and does any complication (blood pressure, blood sugar, cholesterol or mood) show up?
- Is there a structured, supervised programme covering food, activity and behaviour, involving the family and led by someone with nutrition training?
- Will the plan screen for eating-disorder risk and stay free of blame?
- If medicine comes up for a 10-to-19-year-old, how long has the lifestyle programme run (the guideline says at least six months), which complication is it aimed at, and who will monitor puberty, growth and mood and follow up long term?
- If a drug is offered to a child under 10, is it off-label, and what evidence is that based on?
The strongest rule in the package sits exactly where the data run out, at the age-10 line. WHO calls the children's guideline a living one, promising that "future iterations will incorporate new evidence," while trials in six-year-olds are running. It gives no date for an update.
