Infections due to carbapenem-resistant Gram-negative bacteria (CR-GNB) are associated with considerable morbidity and mortality. Polymyxin B (PMB) is a first-line agent for CR-GNB-associated pneumonia, but limited data exist on the clinical use of inhaled (IH) PMB. A single-center, prospective randomized controlled trial was conducted in China to compare IH PMB alone with IH plus intravenous (IV) PMB between February 2022 and February 2024. The primary outcome was the clinical cure rate. Twenty-two evaluable patients were assigned to the IH group, and 56 patients were included in the IH+IV group. Baseline characteristics were comparable between the two groups. No significant differences were observed in clinical cure rates, favorable clinical outcomes, microbiological outcomes, all-cause mortality, or pneumonia-related mortality. However, IH PMB alone was associated with a lower incidence of nephrotoxicity (P = 0.030). IH PMB demonstrated significantly higher drug concentrations in the epithelial lining fluid (ELF) compared to systemic administration. Patients with immunosuppressive therapy (OR, 0.066
95% CI, 0.010-0.433
P = 0.005), malignancies (OR, 0.112
95% CI, 0.016-0.797
P = 0.029), and higher SOFA scores (OR, 0.693
95% CI, 0.518-0.929
P = 0.014) were less likely to achieve favorable clinical outcomes. Conversely, higher PMB ELF 1-hour concentrations (OR, 1.085
95% CI, 1.026-1.148
P = 0.004) were associated with more favorable clinical outcomes. The combination of these four indicators demonstrated excellent diagnostic performance (AUC = 0.882). Plasma 1-hour PMB concentrations showed acceptable predictive performance for nephrotoxicity (AUC = 0.766). The potential benefits of IH PMB outweigh the risks, making it an effective treatment for CR-GNB-associated pneumonia in combination with other empirical antimicrobial agents.