Empiric Colistimethate May Reduce ICU Survival in MDR Gram-Negative Bacteria: A Retrospective Study

Introduction: Multidrug-resistant (MDR) Gram-negative bacteria (GNB) in critically ill patients prolong ICU stay and increase morbidity and mortality. Polymyxins are commonly used empirically, but their effectiveness remains uncertain due to variability in resistance patterns, patient and hospital factors.


Methods: We conducted a retrospective cohort study in the ICU of Hospital Civil de Guadalajara "Dr. Juan I Menchaca" from 2022 to 2024. We included critically ill adults with ICU stays >48 hours and positive respiratory, urinary, or bloodstream cultures for MDR GNB identified using the VITEK®2 system. Data was retrieved from microbiology databases and hospital’s Electronic Health Records (EHR). EHR was used to distinguish active infections from colonization. The primary outcome was ICU mortality; secondary outcomes were 30-day mortality, ICU length of stay, clinical cure, and acute kidney injury (AKI) requiring renal replacement therapy (RRT). Other variables were demographic characteristics, comorbidities, infection site, mechanical ventilation, shock, APACHE II, SOFA, admission diagnosis, surgical status, and adequacy of empirical antibiotic therapy. We used logistic and Cox regression analyses to identify factors of ICU mortality. The study was approved by the local ethics committee and registered on OSF (osf.io/h9f4k). Writing was refined by using LLM tools.


Results: 224 patients were eligible (mean age 48.2 years; 55.8% male); hypertension (35.3%) and diabetes (27.2%) were common comorbidities. MDR Acinetobacter baumannii (56.7%) and MDR Klebsiella pneumoniae (27.2%) were the predominant pathogens. Tracheal secretions were the main culture site (66.1%), and 83.9% had single-site infections. ICU mortality was 53.6%, AKI requiring RRT occurred in 17.9% and carbapenem resistance (CR) in 78.5%. Colistimethate sodium (CMS) was used in 51.8% and was associated with higher ICU mortality (64.5% vs. 43.0%; p = 0.001; HR 1.52, 95% CI 1.04–2.22) and increased risk of AKI requiring RRT (OR 2.52, 95% CI 1.22–5.20). CR (OR 3.76, 95% CI 1.84–7.34) and AKI requiring RRT (OR 6.45, 95% CI 2.58–16.12) were also associated with ICU mortality.


Conclusions: Empirical CMS may reduce ICU survival, likely from severe adverse events and antimicrobial resistance patterns, highlighting the need for local risk–benefit assessment