WHO draws a hard line on GLP-1 drugs for children with obesity
The World Health Organization has set a notably conservative standard for treating obesity in children and adolescents, putting diet, physical activity, and caregiver-led behavioral programs ahead of medication. For children younger than 10, the guidance advises against obesity drugs, surgery, and devices because of lingering uncertainty about safety and long-term effectiveness. In adolescents 10 to 19, GLP-1 drugs may be considered only after supervised interventions fail.
The practical logic is straightforward. Childhood obesity develops in the context of family eating patterns, school environments, sleep disruption, sedentary behavior, and emotional stress, so a drug alone rarely addresses the biology and the behavior at the same time. GLP-1 therapy can reduce appetite and improve satiety signaling, but the global evidence base in young patients is still thin enough that WHO is not treating these agents as first-line care. Luz María De Regil, who leads WHO’s Department of Nutrition and Food Safety, said the current evidence on GLP-1 drugs is insufficient and could be revisited as trial data emerge.
The age cutoff matters clinically because growth, pubertal development, and long-term adherence all change the risk-benefit equation in pediatrics. A medication that may be acceptable in an adult with established cardiometabolic disease has to clear a much higher bar in a 10-year-old who may need treatment decisions measured in years, not months. In the U.S., Wegovy and Saxenda are already approved for certain patients 12 and older, so the WHO position does not reverse national approvals; it simply signals that global pediatric use should remain cautious until stronger long-term data accumulate.
UNICEF data cited in the article frame obesity as the most common form of malnutrition among school-age children worldwide, which is an uncomfortable but accurate clinical reality: excess calories and undernutrition can coexist, especially when diets are heavy in ultra-processed foods and light in protein, fiber, and micronutrients. For clinicians, the new guidance reinforces a stepped-care model in which structured lifestyle treatment is not a placeholder but the main therapy. Medication may still have a role, but only after intensive behavioral support has been tried in a supervised way.
Chronic back pain persisted despite years of daily opioid exposure
A Northwestern University study published Oct. 1 in Brain adds another hard clinical data point to the long-running opioid debate: patients with chronic back pain who used prescription opioids daily for an average of 7.8 years reported no better pain control than closely matched patients who did not use opioids. Both groups of 70 rated pain at about 5.4 on a 10-point scale, while the opioid group scored slightly worse on physical function and mental health.
The finding is stark because it cuts against the expectation that long-term opioid therapy should deliver durable analgesia in chronic noncancer pain. Instead, the biology of persistent opioid exposure can work against the desired effect. Tolerance blunts analgesic response, reward circuitry adapts, and pain processing can become more entrenched over time, leaving patients exposed to the adverse effects of opioids without the relief that justified ongoing treatment in the first place.
Brain scans in the study showed little anatomical change but clear differences in activity in regions tied to reward and decision-making. That pattern fits with a disorder that is not only nociceptive but also neurobehavioral, in which chronic pain and opioid exposure may both reshape how the brain evaluates discomfort, motivation, and perceived control. Senior author A. Vania Apkarian said the team saw no sign that pain improved after so many years on the drug, and the participants were taking low or moderate doses under medical supervision.
The clinical implication is direct. Long-term opioid continuation should not be assumed to equal pain improvement, especially in chronic back pain where function often matters more than a single pain score. If the drug is not improving mobility, mood, or day-to-day coping, continued prescribing may be preserving dependence rather than restoring health. The study was observational, so it cannot prove causation, but it does strengthen the case for reassessing chronic opioid regimens with the same rigor used to start them.
Biohub’s $1.8 billion push aims to turn cell behavior into trainable AI data
A separate piece of the day’s medical news points in a very different direction: the push to build large-scale biological datasets for artificial intelligence. Meta, Google DeepMind, and the drug discovery company Isomorphic Labs are jointly investing $300 million in the Virtual Biology Initiative run by Biohub, the nonprofit supported by Meta CEO Mark Zuckerberg and Priscilla Chan, M.D. The Department of Energy is contributing more than $500 million over five years, and the National Institutes of Health will coordinate datasets built with more than $500 million in earlier federal funding.
The scientific premise is compelling. Modern biology still relies on narrow slices of experimental conditions, while real cells respond to a vast range of signals, environments, and perturbations. By measuring how cells react across far more conditions than scientists have previously studied, the project aims to generate data rich enough to train AI models that may shorten drug development and help identify therapeutic pathways more efficiently.
Data access will not be immediate. Commercial funders receive an embargo period before the information becomes public, and Biohub’s head of science, Alex Rives, expects the first dataset in about a year. That delay is not a footnote; it reflects the tension between open science, public investment, and commercial incentives in modern biomedical research. The central question is whether the resulting datasets will be broad and rigorous enough to move beyond elegant computation and into actual therapeutic discovery.
Taken together, the three stories sketch a medicine field moving in opposite directions at once. Pediatric obesity care is becoming more conservative about pharmacology, chronic pain data are challenging assumptions about the value of prolonged opioid use, and federal as well as private capital is pouring into AI-enabled biology. Each reflects the same underlying demand: treatment claims now need stronger evidence, longer follow-up, and clearer biological justification before they are treated as settled clinical practice. The unanswered question is whether the next wave of data will make those thresholds easier to meet, or simply reveal how many current habits were built on incomplete evidence.