Drug-Resistant Superbugs Found in Most Ethiopian Children’s ICU Infections

In a pediatric intensive care unit in central Ethiopia, more than two out of every three clinically significant bacterial infections are now caused by organisms that doctors would classify as high-risk for antimicrobial resistance. That is the stark conclusion of a five-year retrospective study published in BMC Pediatrics, which combed through the medical and microbiological records of children admitted to the pediatric intensive care unit, or PICU, at Asella Teaching and Referral Hospital between July 2020 and June 2025. The findings, drawn from one of the few datasets of its kind for non-neonatal pediatric intensive care in the country, paint a sobering picture of how thoroughly resistant bacteria have penetrated the wards of a resource-limited referral hospital, and they offer a rare quantitative glimpse into the epidemiology of infections that too often go unmeasured.

The research team, led by Abdi Hayato and Tesfa Gebremeskel of the Department of Pediatrics and Child Health at Arsi University College of Health Sciences, together with Solomon Tejineh of the Department of Public Health, reviewed the records of 217 pediatric patients aged 29 days to 14 years who had culture results available during the study window. Because the study was retrospective, the investigators worked entirely from existing hospital documentation rather than prospectively collecting samples, an approach that trades some methodological control for access to five full years of real-world clinical practice. Ethical clearance came from the Arsi University College of Health Sciences Ethical Review Committee, and individual informed consent was waived because the analysis used anonymized retrospective data, in line with the Declaration of Helsinki.

The raw numbers tell a story of both diagnostic challenge and microbial burden. Of the 217 patient records reviewed, just over half, 52.5 percent, produced any positive culture at all, a figure the researchers describe as the crude culture yield. But positive cultures are not always meaningful: some organisms, particularly coagulase-negative staphylococci, are common skin dwellers that frequently contaminate samples without causing disease. When the team excluded these likely contaminants, the proportion of patients with bacteriologically significant cultures fell to 31.3 percent, or 68 of the 217 children. That distinction matters enormously for clinical interpretation, because treating a contaminant as an infection drives unnecessary antibiotic use, which in turn fuels the very resistance problem the study documents.

From those significant cultures, the researchers identified 112 bacterial isolates across all specimen types. Gram-negative organisms, bacteria whose cell walls are structured in a way that makes them inherently harder to treat and more prone to acquiring resistance genes, accounted for 54.5 percent of the total. Among the Gram-negatives, Klebsiella species emerged as the most frequent pathogen, a genus that has become notorious worldwide for its capacity to harbor powerful resistance enzymes on plasmids, the mobile genetic elements that bacteria swap like trading cards. The predominance of Klebsiella in a pediatric intensive care setting is particularly concerning because the organism thrives in hospital environments, colonizes ventilated patients, and readily spreads through contaminated hands and equipment.

To quantify the severity of resistance, the team applied the World Health Organization’s classification scheme for high-risk phenotypes. These categories include multidrug-resistant organisms, which shrug off multiple antibiotic classes; extensively drug-resistant organisms, which remain susceptible to only a handful of last-line drugs; pan-drug-resistant organisms, against which no available agent works; bacteria producing extended-spectrum beta-lactamases, enzymes that dismantle many penicillins and cephalosporins; and carbapenem-resistant Enterobacterales, which defeat even the carbapenems, the antibiotics typically reserved as a final line of defense. Importantly, these categories are not mutually exclusive, so a single isolate can satisfy several definitions at once. Antimicrobial susceptibility testing was performed on 66 clinically significant isolates, one per patient, providing the analytical backbone of the study.

The susceptibility results were alarming. Among the 66 tested isolates, 45, or 68.2 percent, carried at least one high-risk resistance phenotype. Extended-spectrum beta-lactamase production was the single most common phenotype at 19.7 percent, followed by extensive drug resistance at 15.2 percent, multidrug resistance and pan-drug resistance at 13.6 percent each, and carbapenem resistance at 6.1 percent. Resistance to older beta-lactam antibiotics was close to total: 91.3 percent of isolates resisted ampicillin and 92.3 percent resisted ceftazidime, a third-generation cephalosporin that many treatment guidelines still position as a workhorse drug. In practical terms, for a clinician at this hospital, empiric therapy with those agents would fail roughly nine times out of ten.

Yet the antibiogram was not uniformly bleak. Two drugs retained meaningful activity: amikacin, an aminoglycoside, remained effective against 83.3 percent of tested isolates, and meropenem, a carbapenem, against 85.7 percent. These susceptibility profiles are the study’s most actionable contribution, because an antibiogram, the aggregated map of which drugs still work against which local organisms, is the foundation on which hospital antibiotic stewardship programs are built. The authors suggest that these data can inform the development of such a program at Asella Teaching and Referral Hospital, guiding which agents should be used for empiric therapy and which should be conserved. In settings where diagnostic microbiology is limited, a reliable local antibiogram is often the only rational basis for treatment decisions.

The study also included an exploratory statistical analysis of factors associated with high-risk resistance, using binary logistic regression with a significance threshold of 0.05. One association stood out: each additional day of PICU stay was linked to a dramatically increased odds of high-risk antimicrobial resistance, with an adjusted odds ratio of 7.59 and a 95 percent confidence interval stretching from 1.10 to 52.60. The authors are careful, and rightly so, to flag that this estimate is statistically unstable. The enormous width of the confidence interval reflects the small sample size and the retrospective design, and they explicitly caution that the figure should not be interpreted as a valid per-day effect. Still, the direction of the association is biologically plausible: longer intensive care stays mean more invasive procedures, more ventilator time, more central lines, and more cumulative antibiotic exposure, all of which select for resistant organisms.

The broader context makes these local findings globally significant. Antimicrobial resistance is recognized by the World Health Organization as one of the top public health threats facing humanity, and low- and middle-income countries bear a disproportionate share of the burden, partly because of limited diagnostic capacity, constrained access to newer antibiotics, weaker infection prevention and control infrastructure, and over-the-counter antibiotic availability that encourages misuse. Pediatric intensive care units concentrate all of these risk factors in the most vulnerable patients: children whose immune systems are immature, whose bodies are small, and whose infections can progress from fever to septic shock within hours. Data from such settings are chronically scarce, which is why the authors emphasize that prospective, multicenter surveillance with standardized microbiological protocols is urgently needed across Ethiopia and comparable settings.

What happens next will determine whether this study becomes a turning point or another warning that goes unheeded. The authors call for institutional stewardship development anchored in the antibiogram they have generated, alongside strengthened infection prevention and control measures to interrupt transmission of resistant organisms within the unit. The study’s limitations, its single-center design, retrospective data collection, modest isolate numbers, and the instability of the regression estimate, are real, but they do not diminish the central signal: in this Ethiopian PICU, the era in which standard beta-lactam antibiotics could be relied upon to save critically ill children is effectively over, and the remaining effective drugs, amikacin and meropenem, must be protected before they too are lost. As resistant bacteria continue their quiet global expansion, studies like this one serve as both a measurement and a warning, documenting precisely how much ground has already been surrendered in the fight against one of medicine’s most implacable adversaries.

Subject of Research: Antimicrobial resistance patterns among bacterial infections in a pediatric intensive care unit in Ethiopia

Article Title: Bacterial profile and antimicrobial resistance patterns among pediatric intensive care unit patients in Ethiopia: a retrospective cohort study, 2020–2025

Article References: Hayato, A., Gebremeskel, T., & Tejineh, S. (2026). Bacterial profile and antimicrobial resistance patterns among pediatric intensive care unit patients in Ethiopia: a retrospective cohort study, 2020–2025. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07748-9

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07748-9

Keywords: antimicrobial resistance, pediatric intensive care, Klebsiella, ESBL, carbapenem-resistant Enterobacterales, Ethiopia, antibiogram, Gram-negative bacteria, antibiotic stewardship, retrospective cohort, hospital-acquired infection, BMC Pediatrics