Colistin Combined with Ampicillin–Sulbactam versus Colistin Monotherapy for Carbapenem-Resistant Acinetobacter baumannii Ventilator-Associated Pneumonia: A Retrospective Comparative Non-Randomized Cohort Analysis of Microbiological Eradication and 30-Day

Abstract

Background: Carbapenem-resistant Acinetobacter baumannii (CRAB) ventilator-associated pneumonia (VAP) is associated with high mortality and limited therapeutic options. In settings where sulbactam–durlobactam is unavailable, colistin-based regimens and ampicillin–sulbactam are frequently used. We performed a limited retrospective comparative cohort analysis of critically ill patients with CRAB-VAP treated with colistin plus ampicillin–sulbactam or colistin monotherapy. The aim of the study is to compare microbiological eradication and 30-day all-cause mortality between colistin combined with ampicillin–sulbactam (combination) and colistin monotherapy in critically ill patients with CRAB-VAP.
Methods: This single-center retrospective, non-randomized comparative analysis included adults with microbiologically confirmed CRAB-VAP treated in a tertiary mixed ICU in Plovdiv, Bulgaria, from January 2025 through March 2026. In the combination group, 15/16 patients received intravenous colistin plus ampicillin–sulbactam 3 g every 8 h (sulbactam 3 g/day) by extended infusion over 3–4 h, while only 1 patient received 4 g of ampicillin–sulbactam every 8 h by extended infusion plus colistin. All patients received a colistimethate sodium (CMS) loading dose of 9 MIU followed by 3 MIU every 8 h in patients with normal renal function, and maintenance adjustment was done according to renal function in patients with renal injury; the monotherapy group received intravenous colistin alone. The primary outcomes were microbiological eradication on follow-up tracheal aspirate culture and 30-day all-cause mortality. Fisher’s exact test and Mann–Whitney U test were used for unadjusted comparisons. Given the limited sample size and baseline imbalances, results are interpreted as associations rather than treatment effects.
Results: Eighteen patients were screened for combination treatment; two were excluded: one had bloodstream infection as the primary infectious focus, and one died before antimicrobial treatment initiation, leaving 16 patients in the combination group. Sixteen patients receiving colistin monotherapy in the close time frame were analyzed, for a final cohort of 32 patients. Microbiological eradication occurred in 8/16 (50.0%) patients receiving combination therapy and 4/16 (25.0%) receiving monotherapy (unadjusted odds ratio [OR], 3.00; 95% confidence interval [CI], 0.68–13.3; Fisher’s exact p=0.273). Thirty-day all-cause mortality was 5/16 (31.25%) in the combination group and 12/16 (75.0%) in the monotherapy group (unadjusted OR for death, 0.15; 95% CI, 0.03–0.76; Fisher’s exact p=0.032). Colistin treatment duration was shorter in the combination group than in the monotherapy group: median 9.0 days (IQR, 7.75–10.0) versus 13.5 days (IQR, 9.75–14.0), respectively (Mann–Whitney U=47.0; exact p=0.0017). Patients receiving monotherapy were older and had substantially higher baseline procalcitonin concentrations. These imbalances may have contributed to the observed outcome differences.
Conclusion: In this small, retrospective, non-randomized cohort, colistin plus ampicillin-sulbactam was associated with lower observed 30-day mortality than colistin monotherapy. Microbiological eradication was numerically higher but was not statistically significantly different between groups. Causal inference is not justified because of the small sample, non-randomized treatment allocation, and imbalance in baseline age and inflammatory severity. Prospective multicenter studies with standardized dosing and adequate confounder control are required.