Real-world effectiveness and treatment costs of ceftazidime-avibactam versus colistin-meropenem for ICU-acquired MDR gram-negative infections in liver transplant recipients.

BACKGROUND: Multidrug-resistant (MDR) Gram-negative infections remain a threat after liver transplantation, yet comparative treatment data in critically ill recipients are scarce. Therapeutic decisions are complicated by immunosuppression, resistance mechanisms, and antimicrobial toxicity. We evaluated ceftazidime-avibactam (CAZ/AVI) versus colistin-meropenem for ICU-acquired MDR Gram-negative infections. MATERIALS AND METHODS: This retrospective single-center cohort included 190 liver transplant recipients treated between October 2021 and April 2025; 100 received CAZ/AVI-based therapy and 90 colistin-meropenem. Detailed data were available for 169 patients. The primary outcome was clinical response. Secondary outcomes were microbiological eradication, sepsis-related mortality, 30-day all-cause in-hospital mortality, and direct antimicrobial costs. Multivariable logistic regression adjusted for severity-related, transplant-related, and microbiological covariates. RESULTS: In the 169-patient cohort, clinical response was higher with CAZ/AVI than colistin-meropenem (86.9% vs. 56.5%; p < 0.001), as was microbiological eradication (94.0% vs. 72.9%; p < 0.001). CAZ/AVI remained associated with clinical response (aOR, 4.83; 95% CI, 2.10-11.11) and lower sepsis-related mortality (11.9% vs. 43.5%; aOR, 0.19; 95% CI, 0.08-0.45), whereas 30-day all-cause mortality did not differ (41.7% vs. 47.1%; aOR, 1.02; 95% CI, 0.52-2.00). Mean treatment cost was higher with CAZ/AVI, yielding 48,571.76 TL per additional clinical response; this estimate was descriptive and did not represent formal cost-effectiveness. CONCLUSIONS: When microbiologically appropriate, CAZ/AVI was associated with better clinical and microbiological outcomes and lower sepsis-related mortality than colistin-meropenem, without lower 30-day all-cause mortality. Given the observational design, incomplete treatment-appropriateness data, and limited economic scope, these associations require prospective validation before causal conclusions are drawn.