Colistin Alone Matches Antibiotic Combos for Deadly Infections in Children, Study Finds

When children in intensive care units develop infections caused by bacteria that shrug off nearly every modern antibiotic, doctors often reach for colistin, an old drug from the 1950s that most physicians had shelved decades ago. Because colistin alone has sometimes seemed insufficient against these multidrug-resistant Gram-negative organisms, many clinicians pair it with a second agent, most commonly the carbapenem meropenem, hoping that the combination will boost bacterial killing and improve survival. A new retrospective study from Chiang Mai University Hospital in Thailand now offers some of the most detailed evidence yet on whether that gamble pays off in children, and the answer is more nuanced than many clinicians might expect.

The research, published in New Microbes and New Infections, analyzed the medical records of 218 pediatric patients aged zero to eighteen years who received intravenous colistin for at least 72 hours for confirmed multidrug-resistant Gram-negative bacterial infections between 2010 and 2024. Seventy-four children received colistin as monotherapy, while 144 received colistin-based combination therapy with at least one additional antibiotic active against Gram-negative bacteria. The investigators, led by Ajaree Rayanakorn, Peninnah Oberdorfer, and Wasan Katip, set out to answer a question that has divided infectious disease specialists for years: does adding a second drug to colistin actually save lives in children, or does it merely add cost, complexity, and potential side effects?

The stakes are considerable. Multidrug-resistant Gram-negative bacteria, including Acinetobacter baumannii, Pseudomonas aeruginosa, Klebsiella pneumoniae, and carbapenem-resistant Enterobacteriaceae, are among the most feared pathogens in modern hospitals. Reported mortality rates from these infections range from 26 to 80 percent, and a meta-analysis cited by the authors found that patients with these infections face a 1.78-fold higher risk of death compared with those infected by other organisms. The burden falls especially heavily on the World Health Organization’s South-East Asia region. In Thailand alone, approximately 19,000 additional deaths each year are attributed to multidrug-resistant bacteria, and more than 80 percent of those deaths involve Gram-negative organisms. The World Health Organization has designated the development of new treatments for these pathogens as a global research priority, but new drugs remain scarce, leaving older agents like colistin as last-resort options.

Colistin, also known as polymyxin E, works through a mechanism that differs from most modern antibiotics. The molecule carries both hydrophilic and lipophilic properties, allowing it to attack the lipopolysaccharide component of Gram-negative bacterial membranes. By disrupting the membrane structure, colistin exerts a direct bactericidal effect, and pediatric studies have reported clinical cure rates of up to 89 percent in multidrug-resistant Gram-negative infections. However, its clinical efficacy as a single agent remains controversial, and in vitro experiments have repeatedly shown that colistin can act synergistically with carbapenems, cephalosporins, sulbactam-containing regimens, fluoroquinolones, and fosfomycin. Those laboratory findings drove widespread adoption of combination therapy in clinical practice, even though rigorous clinical evidence, particularly in children, has been largely absent.

To untangle the effects of treatment from the effects of illness severity, the Thai team employed a sophisticated statistical technique known as inverse probability of treatment weighting. This is critical because, in everyday practice, sicker children tend to receive combination therapy. In the unadjusted data, the combination group indeed appeared more severely ill: their median Pediatric Index of Mortality 3 score was higher, septic shock was more common, ventilator-associated pneumonia occurred more frequently, and roughly half were receiving vasoactive drugs compared with about a third of the monotherapy group. Propensity scores were estimated through multivariable logistic regression incorporating variables including sex, age, pediatric Glasgow Coma Scale, pediatric Sequential Organ Failure Assessment score, PIM3 score, septic shock, intensive care unit status, vasoactive drug use, baseline serum creatinine, ventilator-associated pneumonia, and bacteremia. After weighting, all standardized mean differences fell below the 0.1 threshold, indicating that the two groups had become statistically comparable across these measured prognostic factors.

The results were striking in what they did and did not show. After the weighting adjustment, colistin-based combination therapy was not associated with any significant difference in the primary outcome of 30-day mortality compared with monotherapy, with an adjusted odds ratio of 2.50 and a wide confidence interval spanning 0.69 to 9.15. Nor did the combination improve mortality at the end of treatment, clinical response, or the rate of nephrotoxicity, the most feared side effect of colistin, which was defined using consensus criteria based on serum creatinine increases or the RIFLE kidney injury classification. Across all of these clinical endpoints, the two strategies performed essentially the same.

There was, however, one clear exception. Children receiving combination therapy were significantly more likely to achieve microbiological response, meaning that follow-up cultures from the originally infected site no longer grew the multidrug-resistant organism. After adjustment, the odds of microbiological clearance were 3.62 times higher in the combination group, a result that reached statistical significance with a p-value of 0.037. The authors suggest this enhanced bacterial eradication may reflect genuine synergistic activity between colistin and carbapenems, particularly against Acinetobacter baumannii, which dominated the cohort, accounting for more than 82 percent of isolates in both treatment groups. The most frequently used combination regimen was colistin with meropenem, given to nearly 63 percent of the combination group.

The disconnect between better bacterial clearance and no survival benefit is a recurring puzzle in antimicrobial research, and the authors offer several plausible explanations. Outcomes in critically ill children are shaped by host immune response, organ dysfunction, baseline disease severity, and comorbidities, any of which may blunt the clinical impact of eradicating bacteria. Colistin also penetrates lung tissue poorly, which matters in a cohort where roughly 70 percent of infections were pulmonary. Microbiological clearance may simply arrive too late in the course of critical illness to alter its trajectory. The findings align with a systematic review and meta-analysis by Wang and colleagues, which likewise found no mortality or clinical response differences between colistin monotherapy and combination therapy but demonstrated improved microbiological eradication with combination regimens. On the safety front, the absence of excess nephrotoxicity in the combination group is reassuring and supports the view that colistin-associated kidney injury is primarily dose-dependent and driven by patient-specific risk factors rather than synergistic toxicity between colistin and its partners.

The study’s limitations deserve careful attention. As a retrospective, single-center analysis at a 1,400-bed tertiary teaching hospital, its findings may not generalize to institutions with different resistance patterns and prescribing habits. The cohort, while among the largest pediatric studies in this field, remained modest in size, and the wide confidence intervals around key estimates, particularly the 30-day mortality odds ratio, reflect that imprecision. The inclusion of all multidrug-resistant Gram-negative infections, rather than only carbapenem-resistant organisms, complicates extrapolation to the subgroup where combination therapy is most often recommended. Residual confounding from unmeasured variables, including the timing of antibiotic initiation and dosing appropriateness, cannot be excluded, and the predominance of Acinetobacter baumannii may limit applicability to other pathogens. The authors themselves characterize the results as hypothesis-generating rather than definitive.

Even with those caveats, the study carries practical weight for clinicians confronting these infections in children. The message is not that combination therapy is useless, but that its benefits may be narrower than assumed: it may help clear bacteria without adding kidney risk, yet it does not appear to reduce mortality. The authors conclude that the decision to use combination therapy should be individualized, potentially reserved for situations where rapid bacterial clearance is most critical, such as bloodstream infections or ventilator-associated pneumonia caused by Acinetobacter baumannii. For many other children with multidrug-resistant Gram-negative infections, colistin alone may deliver comparable clinical outcomes. Confirming that conclusion will require the kind of study that has long been missing in this vulnerable population: large, multicenter, prospective trials, ideally randomized, designed specifically to define optimal treatment strategies for pediatric patients facing some of the most dangerous bacteria in modern medicine.

Subject of Research: Colistin monotherapy versus combination therapy for pediatric multidrug-resistant Gram-negative infections

Article Title: Efficacy and safety of colistin monotherapy versus colistin-based combination therapy in pediatric patients with multidrug-resistant gram-negative infections: A propensity score analysis

Article References: Rayanakorn, A., Oberdorfer, P., & Katip, W. (2026). Efficacy and safety of colistin monotherapy versus colistin-based combination therapy in pediatric patients with multidrug-resistant gram-negative infections: A propensity score analysis. New Microbes and New Infections, 74, Article 101857. https://doi.org/10.1016/j.nmni.2026.101857

Image Credits: AI Generated

DOI: 10.1016/j.nmni.2026.101857

Keywords: colistin, multidrug-resistant Gram-negative bacteria, pediatric infections, combination therapy, Acinetobacter baumannii, meropenem, nephrotoxicity, propensity score analysis, antimicrobial resistance, Thailand, hospital-acquired pneumonia, microbiological response