Key Points
Question
What are global patterns and forecasts in childhood antimicrobial resistance (AMR) in priority pathogens by World Health Organization (WHO) Access, Watch, and Reserve (AWaRe) antibiotic categories?
Findings
In this cross-sectional study including 106 581 individuals, resistance to any Access-group antibiotic was higher than resistance to those in the Watch and Reserve groups; however, resistance to all 3 increased steeply in resource-limited settings, children aged 0 to 2 years, individuals with sepsis, and intensive care settings. By 2035, carbapenem resistance in children was projected to be significantly higher in Klebsiella species and in Acinetobacter baumannii.
Meaning
High and increasing AMR, especially to first-line antibiotics, threatens empiric treatment and may widen gaps between WHO guidance and clinical reality for children, particularly in resource-limited settings.
This cross-sectional study evaluates antimicrobial resistance to World Health Organization–designated priority pathogens among children.
Abstract
Importance
Antimicrobial resistance (AMR) threatens effective treatment of severe childhood infections, but multiregional data about pediatric AMR remain limited.
Objective
To assess geographic and temporal AMR trends among children for World Health Organization (WHO) priority pathogens using the WHO Access, Watch, and Reserve (AWaRe) antibiotic classification.
Design, Setting, and Participants
This multiregional surveillance study analyzed pediatric bacterial isolates from the Antimicrobial Testing Leadership and Surveillance (ATLAS) database between January 2004 and December 2022. The analysis included 106 581 isolates from 106 581 children aged 0 to 18 years in 82 countries. Data were analyzed from February 2024 to April 2026.
Exposures
Resistance to AWaRe-categorized antibiotics in WHO priority bacterial pathogens.
Main Outcomes and Measures
Temporal trends in resistance were evaluated using linear and nonlinear models, by geographic region, age, setting, and infection. Spatiotemporal generalized additive models were applied to estimate resistance trajectories and forecast to 2035.
Results
Of 106 581 children included in this study, 47% were aged 0 to 2 years, 35% were aged 3 to 12 years, and 18% were aged 13 to 18 years; 58 620 (55%) were male. From 2004 to 2022, pediatric AMR increased in all regions, with consistently higher resistance levels and faster growth in resource-limited settings. Resistance to any Access-group antibiotic was highest overall at a mean of 36% (range, 2%-66%), with resistance to Watch-group antibiotics at a mean of 22% (range, 1%-47%) and Reserve-group antibiotics at a mean of 13% (range, 0%-30%). Resistance to antibiotics in the Watch and Reserve groups increased most rapidly in intensive care units, wherein Watch-group antibiotic resistance increased from 15% (517/3564) to 33% (2910/8748) (P < .001), especially in those aged 0 to 2 years (12% [325/2649] to 32% [1257/3959]; P < .001) and those with sepsis (15% [298/2030] to 30% [1409/4705]; P < .001) and respiratory infections [12% [657/5324] to 29% [3311/11507]; P < .001). Acinetobacter baumannii had the highest overall resistance (more than 55% in every AWaRe category), while Klebsiella species increased fastest, particularly to third- or fourth-generation cephalosporins and carbapenems in Southeast Asia, Eastern Europe, and the Western Pacific. Projections suggested stabilization of resistance to Access antibiotics but continued increases in resistance to antibiotics in the Watch and Reserve groups, especially in Gram-negative pathogens. By 2035, carbapenem resistance was projected to be 35% (95% uncertainty interval [UI], 29%-40%) in Klebsiella species and 82% (95% UI, 77%-85%) in A baumannii.
Conclusions and Relevance
Pediatric AMR increased in all regions across the study period, driven primarily by escalating resistance among Gram-negative pathogens responsible for sepsis and pneumonia. These trends threaten the effectiveness of empiric therapy for severe childhood infections, particularly in settings with limited alternative treatment options.
Introduction
Antimicrobial resistance (AMR) is a global health crisis, recognized by the World Health Organization (WHO) as one of the greatest threats to human health. The emergence and spread of AMR jeopardizes the efficacy of antibiotics, complicating infectious diseases management and increasing morbidity, mortality, and health care costs. The WHO has prioritized AMR surveillance through 2 frameworks: the list of priority pathogens with the most concerning resistance, posing the most serious clinical challenges, and the Access, Watch, and Reserve (AWaRe) classification system, a new approach that stratifies antibiotics based on resistance potential and role in treatment. The AWaRe system provides a much-needed framework for antibiotic use, but relies on up-to-date surveillance data, which are frequently not available in children.
Children are especially vulnerable to infection due to immune immaturity and close contact with peers, facilitating transmission. They are also among the highest antibiotic recipients, with more than 90% receiving at least 1 by their first birthday in some settings. AMR contributes to severe outcomes in children, with an estimated 840 000 deaths in children younger than 5 years associated with AMR in 2021. In vulnerable groups, such as children with cancer and those in neonatal intensive care settings, AMR is associated with higher mortality and high social and economic burden. Evidence shows that AMR is increasing in children across all resource settings, with up to 10% of bloodstream infections attributed to multidrug-resistant organisms, with higher proportions reported in certain pathogen-setting combinations. Despite this, efforts to address AMR in children are hindered by limited surveillance and insufficient child-appropriate antibiotic formulations.
Research on AMR in children remains fragmented, with existing international data, especially from resource-limited settings, often difficult to compare or inaccessible to clinicians, researchers, and policymakers. This gap undermines effective clinical decision-making and development of evidence-based policies. Our study therefore leverages a large longitudinal AMR surveillance dataset and aims to analyze geographic and temporal trends in resistance to antibiotics by WHO AWaRe classification in children, map resistance in WHO-identified critical pathogens, and forecast changes over the next decade.
Methods
Study Design
This surveillance study used pediatric data from the Antimicrobial Testing Leadership and Surveillance (ATLAS) dataset. Covering 82 countries from January 2004 to December 2022, ATLAS includes isolates from blood, cerebrospinal fluid, urine, sterile site fluids, sputum, skin and wounds, and stool (eTable 1 in Supplement 1) and also tests susceptibility. Data are publicly accessible on the Vivli web platform (access ID: 00010421 [previously 00009059]). The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline was followed. Pfizer provided a regulatory waiver of ethical review because of the use of deidentified data.
Data Variables
The ATLAS dataset groups children by ages 0 to 2 years, 3 to 12 years, and 13 to 18 years. Analysis focused on WHO 2024 critical, high, and medium priority pathogens. Exclusions were: Salmonella spp, Mycobacterium tuberculosis, and Shigella spp (low isolate numbers) and Neisseria gonorrhoeae (limited pediatric relevance) (eTable 2 in Supplement 1). Antibiotics were grouped as per WHO AWaRe classification (eTable 3 in Supplement 1): Access, Watch, and Reserve. Resistance by AWaRe category was defined pragmatically as resistance to at least 1 antibiotic in that category, to standardize category-level comparisons. A sensitivity analysis using the median number of resistant antibiotics showed 95% similarity (eTable 9 in Supplement 1). ATLAS provides minimum inhibitory concentrations and susceptibility results based on Clinical and Laboratory Standards Institute (CLSI) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints. Resistance included intermediate and resistant categories. A 2-stage spatiotemporal model estimated resistance by pathogen-antibiotic pair, location, and year (eFigure 1 in Supplement 1). Isolates were assigned to 10 infectious syndromes using International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) codes and published AMR burden methods (eTable 4 in Supplement 1). Regions were defined geographically per the WHO’s division: the Americas (subdivided into North, Central, and South to avoid masking heterogeneity in pediatric AMR), Europe (subdivided into Eastern and Western), Eastern Mediterranean (which includes North African countries), Africa, Southeast Asia, and the Western Pacific (eTable 5 in Supplement 1). For analysis by resource setting, World Bank income groups were used, combining high with upper-middle (high) and low with lower-middle (low) income countries (eTable 5 in Supplement 1).
Risk of Bias
Because ATLAS is maintained by Pfizer, isolates must be representative and not biased toward testing new antibiotics. For regulatory purposes, a specified number of isolates is requested from each site regardless of susceptibility, with selection at site discretion. Urinary isolates are capped at 20% to ensure diversity, and testing is centralized in an independent laboratory to standardize methods. Independent reviews (Wellcome Trust, Open Data Institute, UK Health Security Agency) have validated data quality; methods are described elsewhere. We assessed pathogen distribution (including residual estimates) for anomalies and removed outliers (<0.05% of data).
Statistical Analysis
To examine resistance patterns within the AWaRe framework, temporal trends were assessed through both linear and nonlinear approaches, with the model selected according to best fit. Cochran-Armitage tests were used for linear trends, and quadratic binomial regression where nonlinearity was evident, with P values <.05 considered significant, and a minimum effect size threshold (β > 0.001) applied to avoid overinterpretation of nonmeaningful changes. Subgroup analyses excluded groups with fewer than 10 isolates. To forecast changes in AMR to the year 2035, generalized additive models with binomial distribution adjusting for age group and region were used (sex was excluded because preliminary analyses showed no meaningful difference in resistance by sex), with 95% uncertainty intervals (UIs) computed from 500 simulation draws. To assess temporal trends in AMR to cephalosporins and carbapenems while reducing year-to-year volatility, aggregated resistance proportions within 3-year intervals were calculated and weighted linear regression models fitted, with time as a continuous ordered variable. Annualized rates of change (AROCs) and 95% CIs were estimated using nonparametric bootstrapping (1000 replicates) to generate bias-corrected 95% CIs, providing robust quantification of uncertainty, significant when CIs excluded 0. Analyses used R version 4.5.1 (R Foundation) packages, including ggplot2, sf, rnaturalearth, and rnaturalearthdata for plotting maps and mgcv for forecasting. Data were analyzed from February 2024 to April 2026.
Results
Sample Characteristics
There were 106 581 isolates from 106 581 children in 82 countries over 19 years (Figure 1). Of these, 47% were from children aged 0 to 2 years, 35% 3 to 12 years, and 18% 13 to 18 years, with 58 620 male and 47 961 (45%) female. Most isolates were from medical wards (47%) and intensive care units (ICUs) (27%), and from sputum (30%), skin/wounds (21%), and blood (16%). The most frequent pathogen was Staphylococcus aureus (19%), followed by Klebsiella spp (11%) and Escherichia coli (10%).
Figure 1. Choropleth Map, Stacked Bar Charts, and Dot Plot Showing Characteristics of Sample Isolates.
Each dot represents 1% of the region’s total isolates.
AMR Using AWaRe Antibiotic Classifications
WHO Priority Pathogens and WHO Regions
Using AWaRe as a framework, AMR among WHO priority pathogens varied substantially by region and time. Resistance to Access-group antibiotics was higher for most pathogens during the study years (mean [range], Access: 36% [2%-66%] vs Watch: 22% [1%-47%] and Reserve: 12% [0%-30%]). The exceptions were Enterococcus spp, where the highest resistance was to Watch (48%) and Pseudomonas aeruginosa to Reserve antibiotics (59%) (Figure 2). Enterobacterales had persistently high resistance to Access antibiotics, with large recent increases in Watch and Reserve resistance in Klebsiella spp and E cloacae especially in Africa and Southeast Asia (Figure 2). Up to 2012, Watch resistance in Klebsiella spp in Africa was 29% (185/645), increasing in the last 5 years from 2018 to 2022 to 49% (1777/3654), with 0 Reserve resistance before 2012 increasing to 33% (845/2577). Likewise, there was no Reserve resistance in Klebsiella spp in Southeast Asia before 2012, and over the last 5 years, it increased to 43% (376/865). A baumannii had high resistance across the surveillance dataset, exceeding 55% in every AWaRe category in 2022. Although Reserve resistance in A baumannii in the last 5 years was higher in lower-resource regions (range, 65%-78%) than in higher-resource regions (range, 51%-71%), the steepest increases were in higher-resource regions (eFigure 7 in Supplement 1). In North America, for example, Reserve resistance in A baumannii increased from 12% (86/747) prior to 2012 to 51% (370/730) in the last 5 years (P < .001). A similar pattern occurred in Europe with Eastern Europe having higher Reserve resistance in earlier years, with steeper increases in Western Europe over the whole study period (eFigure 7 in Supplement 1). In contrast, Access resistance in S aureus declined overall, except in Southeast Asia. For example, in South America, resistance decreased from 57% (594/1042) prior to 2012 to 28% (998/3546) in 2018 to 2022 (P < .001). Regional and pathogen-specific summaries are in eFigure 3 in Supplement 1.
Figure 2. Line Plots Showing Resistance in World Health Organization Priority Pathogens to Access, Watch, and Reserve Antibiotic Categories Over Time.
aP < .05.
Clinical Setting
Resistance patterns differed by age and setting. In 2022, children aged 0 to 2 years had the highest Access resistance (33% [2008/6101]), while adolescents aged 13 to 18 years had the highest Watch and Reserve resistance (both 28%: [2062/7443] and [1336/4824], respectively). Between 2004 and 2022, in all inpatient settings (medical, surgical, ICUs), resistance to Access antibiotics declined, while to Watch and Reserve it increased (eFigure 4 in Supplement 1). In ICUs this was significant: Watch resistance increased from 15% (517/3564) to 33% (2910/8748) (P < .001) and Reserve resistance increased from 9% (69/765) to 32% (1765/5520) (P < .001). This was particularly evident in those aged 0 to 2 years, where Watch resistance increased from 12% (325/2649) to 32% (1257/3959) (P < .001), compared to children aged 3 years and older: 21% (192/915) to 35% (1653/4789, P = .20). While resistance in all ages and all AWaRe categories was higher in ICUs in low-income countries, this was exemplified in children aged 0 to 2 years: Watch resistance in ICUs in low-income countries in 2022 was 56% (364/699) compared to 26% (863/3260) in high-income countries (P < .001). In emergency departments, Watch resistance increased from 8% (88/1043) to 16% (549/3529, P < .001): adolescents aged 13 to 18 years had higher resistance at 21% (209/1009) than children 12 years and younger at 13% (340/2570) (P < .001). Access resistance changed little across all outpatient settings (eFigure 4 in Supplement 1).
Infection Syndrome
Watch and Reserve resistance increased from 2004 to 2022 in almost all infection types, most markedly in sepsis and respiratory infections (eFigure 4 in Supplement 1). For sepsis, Watch resistance increased from 15% (298/2030) to 30% (1409/4705) (P < .001) and Reserve from 3% (16/474) to 26% (746/2824) (P < .001). In respiratory infections, Watch resistance increased from 12% (657/5324) to 29% (3311/11507) (P < .001) and Reserve from 9% (96/1113) to 30% (2055/6780) (P < .001). There was no difference in rate of increase between age groups, but by 2022 for sepsis, children aged 0 to 2 years had the highest Watch resistance at 34% (696/2070) compared to children aged 3 years and older at 27% (713/2635) (P < .001). For respiratory infections, adolescents aged 13 to 18 years had the highest Watch resistance in 2022 at 36% (804/2212) compared to those aged 12 years and younger at 27% (2507/9295) (P < .001). In contrast, Access resistance in sepsis and respiratory infections remained stable at approximately 30%, while resistance in skin infections, due predominantly to S aureus, decreased by 71% to 24% (1197/5008) (eFigure 4 in Supplement 1).
Critical and High-Priority Pathogens and Resource Setting
For the 8 critical and high-priority pathogens, country-specific income categories were used to compare the impact of resource setting. In 2004, resistance across pathogens was similar in each category comparing high- and low-income countries: Access, 48% and 44%; Watch, 14% and 16%; Reserve, 9% and 4%, respectively. By 2022, the differences were stark. In high-income countries, Access resistance had decreased to 29% (3999/13 687) compared to that in low-income countries, which remained at 47% (1273/2702) (P < .001). For Watch, in high-income countries resistance increased to 24% (5503/22 648), but was double that in low-income countries at 48% (2170/4548) (P < .001). Reserve resistance was also higher in low-income countries at 37% (1107/3014) vs 25% (3830/15 303) (P < .001). Differences existed across all ages, but the greatest differences between resource settings were in children aged 0 to 2 years (eTable 8 in Supplement 1).
Forecasts in Resistance to AWaRe Antibiotic Classification by 2035
WHO Priority Pathogens and WHO Regions
Forecasting to 2035 based on current trajectories projected A baumannii to have the highest overall resistance by 2035, with Access resistance at 70% (95% UI, 68%-71%), Watch at 70% (95% UI, 69%-72%), and Reserve at 45% (95% UI, 43%-47%) (eTable 7 in Supplement 1). P aeruginosa ranked next, with corresponding estimates of 56% (95% UI, 54%-57%), 35% (95% UI, 33%-36%), and 30% (95% UI, 29%-32%). E cloacae and Klebsiella spp were both forecasted to have Access resistance of approximately 50%, with Watch and Reserve resistance ranging from 25% to 30% (eFigure 6 in Supplement 1). Across most regions, Access resistance was forecasted to stabilize or decline with Watch and Reserve resistance forecasted to continue to increase, particularly in lower-income regions (eFigure 6 in Supplement 1). Forecasted antibiotic resistance in Southeast Asia by 2035 was 47% (95% UI, 44%-50%) for Access, 84% (95% UI, 83%-85%) for Watch, and 43% (95% UI, 40%-47%) for Reserve. Corresponding estimates for Africa were 43% (95% UI, 40%-46%), 77% (95% UI, 75%-79%), and 37% (95% UI, 33%-40%) and for Central America, 41% (95% UI, 38%-43%), 78% (95% UI, 77%-80%), and 37% (95% UI, 33%-40%). North America showed the same pattern, with lower absolute AMR levels (eFigure 6 and eTable 7 in Supplement 1).
Clinical Setting and Infection Syndrome
By clinical setting, ICUs were projected to have the highest resistance rates in 2035, to Access antibiotics at 44% (95% UI, 42%-47%), Watch 80% (95% UI, 79%-82%), and Reserve 42% (95% UI, 38%-45%). Gastrointestinal infections were forecasted to have the highest resistance to both Access (48%; 95% UI, 44%-52%) and Watch antibiotics (81%; 95% UI, 79%-83%), with intra-abdominal infections predicted to have the highest Reserve resistance (37%; 95% UI, 33%-40%). Sepsis was also projected to have high resistance to Watch (79%; 95% UI, 77%-80%) and Reserve antibiotics (35%; 95% UI, 32%-38%). Forecasts did not differ by age (eFigure 6 in Supplement 1).
Cephalosporin and Carbapenem AMR in WHO Critical Priority Pathogens
Increases in resistance to either third- or fourth-generation cephalosporins, and to carbapenems among critical pathogens, Enterobacterales and A baumannii were shown through AROCs (Figure 3) and region-specific comparisons over time (Figure 4; eTable 7 in Supplement 1).
Figure 3. Forest Plot and Heat Map Showing Annualized Rate of Change (AROC) in Resistance to Third- or Fourth-Generation Cephalosporins and Carbapenems in World Health Organization (WHO) Critical Pathogens.
a95% CI (1000 bootstrap replicates) excludes 0.
Figure 4. Choropleth Maps Showing Geographic Distribution of Resistance in World Health Organization Critical Pathogens to Carbapenems.
Raw data were used. Shown are actual distributions in 2021 and forecasted distribution in 2035.
Third- or Fourth-Generation Cephalosporin Resistance
From 2004 to 2022, third- and fourth-generation cephalosporin resistance among critical priority pathogens in children increased from 16% to 31%, and by 2035 was projected to be 36% (95% UI, 26%-39%) (Table; eTable 7 and eFigure 8 in Supplement 1). E coli cephalosporin resistance was 33% by 2022, with a projected increase to 56% by 2035 (95% UI, 51%-61%). There was a significant E coli AROC increase in Africa (2.8; 95% CI, 2.3-3.3). Klebsiella spp had 45% resistance in 2022, forecast to be 70% by 2035 (95% UI, 65%-74%), also with significant AROC increases in Africa (2.0; 95% CI, 1.4-2.5) and Eastern Europe (1.0; 95% CI, 0.7-1.3) (Figure 3). E cloacae is similarly forecast to have 71% (95% UI, 66%-75%) resistance across all regions by 2035. S marcescens had the lowest current resistance (7%) among critical pathogens and the lowest projections (47%; 95% UI, 41%-52%). In contrast, A baumannii had the highest resistance across all regions at 64%, forecasted to increase to 83% (95% UI, 80%-86%). Regions with the highest projected third- or fourth-generation cephalosporin resistance among all critical pathogens by 2035 were Eastern Europe, Southeast Asia and Western Pacific (Table; Figure 4; eTable 7 in Supplement 1).
Table. Third- and Fourth-Generation Cephalosporin and Carbapenem Resistance in World Health Organization (WHO) Critical Pathogens and Regions.
| Resistance to third- and fourth-generation cephalosporins | Resistance to carbapenems | |||
|---|---|---|---|---|
| 2021/2022 Rate, % | Forecasted rate, % (UI) | 2021/2022 Rate, % | Forecasted rate, % (UI) | |
| Critical pathogens | ||||
| Escherichia coli | 33 | 56 (51-61) | 2.7 | 7 (5-9) |
| Klebsiella spp | 45 | 70 (65-74) | 15 | 35 (29-40) |
| Enterobacter cloacae | 24 | 71 (66-75) | 5 | 23 (19-29) |
| Serratia marcescens | 7 | 47 (41-52) | 4 | 25 (20-31) |
| Acinetobacter baumannii | 56 | 83 (80-86) | 51 | 82 (77-85) |
| WHO regions | ||||
| Higher income | ||||
| North America | 33 | 61 (57-66) | 10 | 50 (44-56) |
| Western Europe | 42 | 65 (61-70) | 10 | 50 (44-56) |
| Eastern Europe | 61 | 77 (74-80) | 25 | 65 (59-70) |
| South America | 50 | 79 (75-82) | 22 | 64 (58-69) |
| Western Pacific | 45 | 79 (76-82) | 11 | 61 (55-67) |
| Lower income | ||||
| Central America | 48 | 83 (80-85) | 16 | 67 (62-72) |
| Eastern Mediterranean | 65 | 82 (78-84) | 22 | 60 (54-66) |
| Southeast Asia | 69 | 90 (87-91) | 43 | 79 (75-83) |
| Africa | 69 | 83 (80-86) | 20 | 61 (55-67) |
Carbapenem Resistance
From 2004 to 2022, carbapenem resistance among critical priority pathogens in children increased from 9% to 17%, and by 2035 was projected to be 21% (95% UI, 19%-23%). E coli carbapenem resistance was the lowest among critical pathogens at 3% in 2022, with a projected rise to 7.5% (95% UI, 6%-10%) by 2035. Klebsiella spp resistance was the highest among Enterobacterales at 15%, forecasted to reach 35% (95% UI, 29%-40%), with significant overall AROC (0.72; 95% CI, 0.7-0.8), reflected regionally in Southeast Asia (3.2; 2.6-4.0), Eastern Europe (1.4; 95% CI, 1.2-1.6), and Eastern Mediterranean (1.5; 95% CI, 1.1-1.9) (Figure 3). Carbapenem resistance in both E cloacae and S marcescens was 4%, projected to increase to 22% and 25%, respecitively. A baumannii had the highest resistance at 51% with a forecasted rate by 2035 of 82% (95% UI, 77%-85%). Regions with the highest projected carbapenem resistance across critical pathogens by 2035 were again Southeast Asia and Eastern Europe (Table; Figure 4).
Discussion
In this cross-sectional analysis of more than 100 000 pediatric bacteria isolates, there was a substantial increase in AMR in children over the past 2 decades. Key findings were that resistance to Watch and Reserve antibiotics, kept for severe and multidrug-resistant infections, was especially high in ICUs, children aged 0 to 2 years, those with sepsis, and those with respiratory infections. Forecasts based on current trends projected steep increases in Watch resistance and slower increases in Reserve resistance over the next decade, especially in A baumannii, P aeruginosa, and Klebsiella spp. These findings were driven by higher AMR in resource-limited settings across all AWaRe categories. Conversely, resistance to Access antibiotics was projected to decline.
The higher resistance to Access antibiotics reflects their widespread historical use as first-line agents for childhood infections. However, the declining trend in Access resistance, particularly in S aureus, likely reflects focus on methicillin-resistant S aureus (MRSA) control: a multicenter study in the US—where MRSA infections have been an infection control target—reported a more than 50% decline in MRSA infections between 2016 and 2021. This is encouraging and aligns with antimicrobial stewardship (AMS) prioritization of first-line drugs.
However, high Gram-negative resistance to Access antibiotics suggests that these are becoming ineffective for first-line use in many pediatric settings. The finding of increasing resistance in emergency department isolates mirrors the results of other studies and may signal increasing community rates in children, where surveillance is limited. The problem extends beyond first-line agents: high and increasing Watch and Reserve resistance in ICUs and high-burden infections mark key system stress points. A study of European ICUs reported high third-generation cephalosporin resistance in A baumannii (64%-91%) and P aeruginosa (2%-16%), pathogens with substantial AMR-attributed mortality. Resistance to Watch antibiotics, now frequently first-line in pediatric and neonatal ICUs, and to Reserve antibiotics with limited pediatric pharmacokinetic data or clinical trials, is deeply concerning for infection management in children.
The highest current and projected resistance rates were concentrated in resource-limited regions, including Southeast Asia, Africa and Central America, with previous work determining this to result from a combination of high infection burden, unregulated antibiotic access, and limitations in sanitation and infection control, exacerbated by conflict and natural disasters. Although only phenotypic data were available in this study, in E coli and Klebsiella spp, resistance to third- and fourth-generation cephalosporins and carbapenems was likely driven by horizontally transmitted enzymes (eg, extended-spectrum β-lactamases and carbapenemases on mobile plasmids), selected by widespread broad-spectrum use and spread in health care settings. Targeted interventions include strengthening antimicrobial stewardship, reducing cephalosporin use, and improving infection control. In E cloacae and S marcescens, resistance may arise from inducible AmpC β-lactamase expression, often emerging during treatment despite in vitro susceptibility. The key driver was likely prolonged individual exposure to third-generation cephalosporins, particularly in hospitalized children and intensive care. Interventions include shorter treatment, improving awareness, and access to alternative agents in resource-limited settings. In A baumannii, resistance may reflect intrinsic (eg, oxacillinase-type β-lactamases) and acquired mechanisms, compounded by heavy carbapenem use and environmental persistence, particularly in intensive care. Interventions include reducing carbapenem use, strengthening environmental cleaning, and limiting invasive devices. Overall, while mechanisms differ, resistance was consistently amplified by antibiotic pressure and health care transmission, highlighting the need for coordinated stewardship and infection control adapted to local contexts.
These findings are broadly comparable with the WHO’s Global Antimicrobial Resistance and Use Surveillance System (GLASS) 2022 country ranges (cephalosporin resistance, 33%-100%; carbapenem, 23%-100%), with lower pediatric point estimates expected given case-mix and setting. Without urgent intervention, by 2035 third- and fourth-generation cephalosporin resistance in Klebsiella spp could exceed 70% and carbapenem-resistance in A baumannii could reach 82%, compromising the efficacy of last-resort antibiotics. These projections align with those of WHO models predicting disproportionate increases in multidrug-resistant infections in resource-limited settings, where children face the dual problems of high AMR and limited access to effective antibiotics. Region-specific data can guide pediatric priorities.
These results have clear policy and practice implications. Children remain underrepresented in global AMR strategies despite distinct resistance patterns. Our findings highlight the need for tailored AMS, targeting sepsis and respiratory infections, especially in settings with potentially weaker oversight (eg, emergency departments).
To enhance data accessibility, we developed the AMR in Kids interactive platform, which aggregates pediatric resistance data to support equitable surveillance, data sharing, and context-specific analysis. Aligned with WHO GLASS, this resource aims to give visibility and hands-on access to granular country-level analyses and stratification, particularly in high-burden, low-resource settings. The next step would be to link these surveillance data to AMS strategies, as outlined in the WHO Global Action Plan on AMR, to target and evaluate interventions. National examples from the Netherlands and Denmark show how AMR action plans and integrated surveillance systems can support stewardship, monitor antimicrobial use, and track resistance trends over time. Future research should characterize molecular drivers of resistance in high-burden infections, to better target interventions.
Limitations
This study has several limitations. First, ATLAS is not population-based, and site participation and isolate submission practices may influence pathogen mix and resistance patterns across regions and time. Although testing was performed in a central US laboratory to minimize variability, selection bias may still exist. Findings therefore reflect a large multinational surveillance network rather than fully population-representative global estimates; however, ATLAS remains a valuable resource given the lack of comparable datasets. Second, heterogeneity within AWaRe antibiotic categories complicates interpretation of aggregate resistance rates, as category-level trends may obscure differences between individual antibiotics. Third, molecular mechanisms underpinning phenotypic resistance were not analyzed, limiting interpretation of underlying drivers. Fourth, forecasts assume continuation of current trends and cannot account for future changes in policy, treatment, diagnostics, vaccination, stewardship, conflict, or health system disruption.
Conclusions
The findings in this study indicate that AMR in children increased globally across the study period, driven by increasing resistance among critical Gram-negative pathogens. Integrated, age-stratified AMR surveillance that aligns AWaRe antibiotic recommendations with microbiology and outcomes is urgently needed to guide stewardship and protect the efficacy of essential antibiotics for future generations.
eMethods.
eTable 1. Study datasets contributed to ATLAS antibiotic dataset
eTable 2. WHO 2024 Priority Pathogen list and their resistant antibiotics
eTable 3. WHO Access (Green), Watch (Yellow) and Reserve (Red) antibiotic list
eTable 4. Infectious syndrome contributing to antimicrobial resistance assigned by sample source, treatment site and age group
eTable 5. WHO regions and countries
eTable 6. Pathogen and their intrinsic resistant antibiotics
eTable 7. Prediction of WHO critical pathogens to carbapenem resistance proportion in 2035 by WHO regions and critical pathogens
eTable 8. AWaRe Resistance Among Clinical Setting, Infection Syndrome and WHO Critical and High-Priority Pathogens by Age Group
eTable 9. Sensitivity analysis using median estimates compared with primary analysis
eFigure 1. Number of patients among WHO regions by year
eFigure 2. WHO priority pathogens distribution by year
eFigure 3. Trends of resistance to WHO access, watch, reserve antibiotics among 12 priority pathogens (a) by WHO region (b) over 2004-2022
eFigure 4. Resistant trends to AWaRe antibiotics a) sex, b) age group, c) outpatient setting d) inpatient setting, e) infection syndromes, and f) pathogens
eFigure 5. ATLAS in and out-of-sample residual estimates
eFigure 6. Prediction of WHO roups and among WHO priority pathogens
eFigure 7. Resistance trends in WHO priority pathogens to AWaRe antibiotic categories over time (2004-202) by WHO regions: a) age b) higher resource region and c) lower resource region
eFigure 8. Geographic distribution of 3rd/4th generation cephalosporin resistance in critical priority pathogens by WHO region in 2021 and forecast in 2035
Data sharing statement
References
- 1.World Health Organization . WHO bacterial priority pathogens list, 2024. World Health Organization. Accessed 09 September 2024. https://www.who.int/publications/i/item/9789240093461
- 2.Hsia Y, Lee BR, Versporten A, et al. ; GARPEC and Global-PPS networks . Use of the WHO Access, Watch, and Reserve classification to define patterns of hospital antibiotic use (AWaRe): an analysis of paediatric survey data from 56 countries. Lancet Glob Health. 2019;7(7):e861-e871. doi: 10.1016/S2214-109X(19)30071-3 [DOI] [PubMed] [Google Scholar]
- 3.Rogawski ET, Platts-Mills JA, Seidman JC, et al. Use of antibiotics in children younger than two years in eight countries: a prospective cohort study. Bull World Health Organ. 2017;95(1):49-61. doi: 10.2471/BLT.16.176123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.GBD 2021 Antimicrobial Resistance Collaborators . Global burden of bacterial antimicrobial resistance 1990-2021: a systematic analysis with forecasts to 2050. Lancet. 2024;404(10459):1199-1226. doi: 10.1016/S0140-6736(24)01867-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Williams A, Coombs GW, Bell J, et al. Australian Group on Antimicrobial Resistance surveillance outcome programs—bloodstream infections and antimicrobial resistance patterns from patients less than 18 years of age, January 2020-December 2021. Commun Dis Intell. 2024;48. doi: 10.33321/cdi.202/4.48.32 [DOI] [PubMed] [Google Scholar]
- 6.Kaur J, Singh H, Sethi T. Emerging trends in antimicrobial resistance in bloodstream infections: multicentric longitudinal study in India (2017-2022). Lancet Reg Health Southeast Asia. 2024;26:100412. doi: 10.1016/j.lansea.2024.100412 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dickson BFR, Harrison M, Villanueva-Uy MET, et al. Pathogen distribution and antimicrobial resistance among neonatal bloodstream infections in Southeast Asia: results from NeoSEAP, a multicentre retrospective study. The Lancet Reg Health West Pac. 2025;62:101617. doi: 10.1016/j.lanwpc.2025.101617 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Russell NJ, Stöhr W, Plakkal N, et al. Patterns of antibiotic use, pathogens, and prediction of mortality in hospitalized neonates and young infants with sepsis: a global neonatal sepsis observational cohort study (NeoOBS). PLoS Med. 2023;20(6):e1004179. doi: 10.1371/journal.pmed.1004179 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sivasankar S, Goldman JL, Hoffman MA. Variation in antibiotic resistance patterns for children and adults treated at 166 non-affiliated US facilities using EHR data. JAC Antimicrob Resist. 2023;5(1):dlac128. doi: 10.1093/jacamr/dlac128 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.de Kraker ME, Stewardson AJ, Harbarth S. Will 10 million people die a year due to antimicrobial resistance by 2050? PLoS Med. 2016;13(11):e1002184. doi: 10.1371/journal.pmed.1002184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.World Health Organization . Paediatric drug optimization for antibiotics: meeting report, 30 November, 5-7 December 2022. 2023. Accessed 10 November 2025. https://www.who.int/publications/i/item/9789240068933
- 12.Catalán P, Wood E, Blair JMA, Gudelj I, Iredell JR, Beardmore RE. Seeking patterns of antibiotic resistance in ATLAS, an open, raw MIC database with patient metadata. Nat Commun. 2022;13(1):2917. doi: 10.1038/s41467-022-30635-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Li R, D’Arcy C, Hermsen E, Serio A, Srinivasan A, Bradford P. Creative usage of Vivli’s antimicrobial resistance surveillance open data through the re-use data challenge. Wellcome Open Res. 2024;9(235). doi: 10.12688/wellcomeopenres.21094.1 [DOI] [Google Scholar]
- 14.Jones M, Jernigan JA, Evans ME, Roselle GA, Hatfield KM, Samore MH. Vital signs: trends in Staphylococcus aureus Infections in Veterans Affairs medical centers—United States, 2005-2017. MMWR Morb Mortal Wkly Rep. 2019;68:220-224. doi: 10.15585/mmwr.mm6809e2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kourtis AP, Hatfield K, Baggs J, et al. Vital signs: epidemiology and recent trends in methicillin-resistant and in methicillin-susceptible Staphylococcus aureus bloodstream infections—United States. MMWR Morb Mortal Wkly Rep. 2019;68(9):214-219. doi: 10.15585/mmwr.mm6809e2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.World Health Organization . The WHO AWaRe (Access, Watch, Reserve) antibiotic book. Updated December 9, 2022. Accessed June 2023. https://www.who.int/publications/i/item/9789240062382
- 17.Romandini A, Pani A, Schenardi PA, Pattarino GAC, De Giacomo C, Scaglione F. Antibiotic resistance in pediatric infections: global emerging threats, predicting the near future. Antibiotics (Basel). 2021;10(4):393. doi: 10.3390/antibiotics10040393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Costelloe C, Metcalfe C, Lovering A, Mant D, Hay AD. Effect of antibiotic prescribing in primary care on antimicrobial resistance in individual patients: systematic review and meta-analysis. BMJ. 2010;340:c2096. doi: 10.1136/bmj.c2096 [DOI] [PubMed] [Google Scholar]
- 19.Cerini P, Meduri FR, Tomassetti F, et al. Trends in antibiotic resistance of nosocomial and community-acquired infections in Italy. Antibiotics (Basel). 2023;12(4):651. doi: 10.3390/antibiotics12040651 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Messina NL, Williamson DA, Robins-Browne R, Bryant PA, Curtis N. Risk factors for carriage of antibiotic-resistant bacteria in healthy children in the community: a systematic review. Pediatr Infect Dis J. 2020;39(5):397-405. doi: 10.1097/INF.0000000000002532 [DOI] [PubMed] [Google Scholar]
- 21.Ioannou P, Maraki S, Koumaki D, et al. A six-year retrospective study of microbiological characteristics and antimicrobial resistance in specimens from a tertiary hospital’s surgical ward. Antibiotics (Basel). 2023;12(3):490. doi: 10.3390/antibiotics12030490 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Cristina ML, Sartini M, Spagnolo AM. Serratia marcescens infections in neonatal intensive care units (NICUs). Int J Environ Res Public Health. 2019;16(4):610. doi: 10.3390/ijerph16040610 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Van Goethem S, Xavier BB, Glupczynski Y, et al. Genomic epidemiological analysis of a single-centre polyclonal outbreak of Serratia marcescens, Belgium, 2022 to 2023. Euro Surveill. 2024;29(48):2400144. doi: 10.2807/1560-7917.ES.2024.29.48.2400144 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Thompson G, Barker CI, Folgori L, et al. Global shortage of neonatal and paediatric antibiotic trials: rapid review. BMJ Open. 2017;7(10):e016293. doi: 10.1136/bmjopen-2017-016293 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Gandra S, Alvarez-Uria G, Turner P, Joshi J, Limmathurotsakul D, van Doorn HR. Antimicrobial resistance surveillance in low- and middle-income countries: progress and challenges in eight South Asian and Southeast Asian countries. Clin Microbiol Rev. 2020;33(3):e00048-19. doi: 10.1128/CMR.00048-19 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Michael CA, Dominey-Howes D, Labbate M. The antimicrobial resistance crisis: causes, consequences, and management. Front Public Health. 2014;2:145. doi: 10.3389/fpubh.2014.00145 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Duguid RC, Ashley EA, Turner P, et al. Antimicrobial resistance among children in Southeast Asia: a systematic review. BMJ Public Health. 2025;3(1):e001312. doi: 10.1136/bmjph-2024-001312 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Tornberg-Belanger SN, Rwigi D, Mugo M, et al. Antimicrobial resistance including extended spectrum beta lactamases (ESBL) among E. coli isolated from Kenyan children at hospital discharge. PLoS Negl Trop Dis. 2022;16(3):e0010283. doi: 10.1371/journal.pntd.0010283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Flokas ME, Karanika S, Alevizakos M, Mylonakis E. Prevalence of ESBL-producing Enterobacteriaceae in pediatric bloodstream infections: a systematic review and meta-analysis. PLoS One. 2017;12(1):e0171216. doi: 10.1371/journal.pone.0171216 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Endale H, Mathewos M, Abdeta D. Potential causes of spread of antimicrobial resistance and preventive measures in one health perspective-a review. Infect Drug Resist. 2023;16:7515-7545. doi: 10.2147/IDR.S428837 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lim C, Hantrakun V, Klaytong P, et al. Frequency and mortality rate following antimicrobial-resistant bloodstream infections in tertiary-care hospitals compared with secondary-care hospitals. PLoS One. 2024;19(5):e0303132. doi: 10.1371/journal.pone.0303132 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lee SO, Lee ES, Park SY, Kim SY, Seo YH, Cho YK. Reduced use of third-generation cephalosporins decreases the acquisition of extended-spectrum beta-lactamase-producing Klebsiella pneumoniae. Infect Control Hosp Epidemiol. 2004;25(10):832-837. doi: 10.1086/502304 [DOI] [PubMed] [Google Scholar]
- 33.Mendelson M, Røttingen JA, Gopinathan U, et al. Maximising access to achieve appropriate human antimicrobial use in low-income and middle-income countries. Lancet. 2016;387(10014):188-198. doi: 10.1016/S0140-6736(15)00547-4 [DOI] [PubMed] [Google Scholar]
- 34.Logan LK, Gandra S, Trett A, Weinstein RA, Laxminarayan R. Acinetobacter baumannii resistance trends in children in the United States, 1999-2012. J Pediatric Infect Dis Soc. 2019;8(2):136-142. doi: 10.1093/jpids/piy018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.World Health Organization . Global antimicrobial resistance and use surveillance system (GLASS) report: 2022. Accessed 20 July 2023. https://www.who.int/publications/i/item/9789240062702
- 36.Otaigbe II. Mitigating inequitable access to appropriate antibiotics in low- and middle-income countries. JAC Antimicrob Resist. 2025;7(2):dlaf061. doi: 10.1093/jacamr/dlaf061 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Fay MJ, Bryant PA. Antimicrobial stewardship in children: where to from here? J Paediatr Child Health. 2020;56(10):1504-1507. doi: 10.1111/jpc.15209 [DOI] [PubMed] [Google Scholar]
- 38.AMR in kids. AMR in kids home page. Accessed June 18, 2026. www.amrinkids.com
- 39.Ministry of Health, Welfare and Sport . Netherlands: Dutch national action plan on AMR. Accessed 24 April 2025. https://www.who.int/publications/m/item/netherlands-dutch-national-action-plan-on-amr
- 40.Duarte ASR, Pessoa J, Attauabi M, Lindegaard M, Sönksen UW, eds. DANMAP 2023: use of antimicrobial agents and occurrence of antimicrobial resistance in bacteria from food animals, food and humans in Denmark. Statens Serum Institut and National Food Institute at the Technical University of Denmark; 2024:21-26. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods.
eTable 1. Study datasets contributed to ATLAS antibiotic dataset
eTable 2. WHO 2024 Priority Pathogen list and their resistant antibiotics
eTable 3. WHO Access (Green), Watch (Yellow) and Reserve (Red) antibiotic list
eTable 4. Infectious syndrome contributing to antimicrobial resistance assigned by sample source, treatment site and age group
eTable 5. WHO regions and countries
eTable 6. Pathogen and their intrinsic resistant antibiotics
eTable 7. Prediction of WHO critical pathogens to carbapenem resistance proportion in 2035 by WHO regions and critical pathogens
eTable 8. AWaRe Resistance Among Clinical Setting, Infection Syndrome and WHO Critical and High-Priority Pathogens by Age Group
eTable 9. Sensitivity analysis using median estimates compared with primary analysis
eFigure 1. Number of patients among WHO regions by year
eFigure 2. WHO priority pathogens distribution by year
eFigure 3. Trends of resistance to WHO access, watch, reserve antibiotics among 12 priority pathogens (a) by WHO region (b) over 2004-2022
eFigure 4. Resistant trends to AWaRe antibiotics a) sex, b) age group, c) outpatient setting d) inpatient setting, e) infection syndromes, and f) pathogens
eFigure 5. ATLAS in and out-of-sample residual estimates
eFigure 6. Prediction of WHO roups and among WHO priority pathogens
eFigure 7. Resistance trends in WHO priority pathogens to AWaRe antibiotic categories over time (2004-202) by WHO regions: a) age b) higher resource region and c) lower resource region
eFigure 8. Geographic distribution of 3rd/4th generation cephalosporin resistance in critical priority pathogens by WHO region in 2021 and forecast in 2035
Data sharing statement
