- Design
- Multiregional surveillance study with forecasting models
- Population
- 106,581 bacterial isolates from children aged 0-18 in 82 countries, 2004-2022
- Primary outcome
- Trends in resistance by WHO AWaRe antibiotic group
- Effect
- ICU Watch-group resistance 15% to 33%; projected 2035 carbapenem resistance 35% in Klebsiella
This surveillance study analysed 106,581 bacterial isolates from children aged 0 to 18 in 82 countries between 2004 and 2022, grouping resistance by the WHO Access, Watch and Reserve (AWaRe) classes and projecting trends to 2035. Published in JAMA Pediatrics on 1 October 2026.
Resistance rose in every region and grew fastest in resource-limited settings. In intensive care, resistance to Watch-group antibiotics more than doubled, from 15% to 33%, and in children under two from 12% to 32%. Acinetobacter baumannii was resistant in more than half of isolates across every class, and Klebsiella resistance to third-generation cephalosporins and carbapenems rose fastest in South-East Asia and the Western Pacific. By 2035, carbapenem resistance is projected at 35% in Klebsiella and 82% in Acinetobacter.
For Indian paediatricians, this is the setting the data describe. Empiric regimens for neonatal and paediatric sepsis that worked a decade ago may now fail more often.
The ATLAS database draws on hospital isolates sent for testing, which may over-represent severe and resistant infections.
- Send cultures before starting antibiotics in every child with suspected sepsis
- Review empiric sepsis regimens against your own unit's antibiogram at least yearly
- Narrow or stop antibiotics at 48 hours when cultures are negative and the child is well
- Prefer Access-group antibiotics in the community for uncomplicated infections
- Treat infection control in NICU and PICU as part of antibiotic stewardship
Why it matters
Empiric therapy for the sickest children is losing ground fastest in the regions where alternatives are scarcest.
Don't overread it
The 2035 figures are model projections, not observed resistance.
The statistics, in plain English
The rises within ICUs come from large numbers of isolates and are very unlikely to be chance. The 95% uncertainty intervals for 2035 describe the model's range, not certainty about the future.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for paediatrics, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free