Abstract Objectives To assess the pharmacokinetics of formed colistin in plasma and the safety of two different high doses of colistimethate sodium administered via nebulization in critically ill surgical patients with hospital-acquired pneumonia (HAP) or ventilator-associated pneumonia (VAP). Patients and methods Formed colistin plasma concentrations were measured in critically ill surgical patients with pneumonia treated with two different doses of nebulized colistimethate sodium (3 MIU/8 h versus 5 MIU/8 h). Adverse events possibly related to nebulized colistimethate sodium were recorded. Results Twenty-seven patients (15 in the 3 MIU/8 h group and 12 in the 5 MIU/8 h group) were included. Colistin plasma concentrations were unquantifiable (<0.1 mg/L) in eight (53.3%) patients in the 3 MIU/8 h group and in seven patients (58.3%) in the 5 MIU/8 h group. Median (IQR) quantifiable colistin plasma concentrations before nebulization and at 1, 4 and 8 h were 0.17 (0.12–0.33), 0.20 (0.11–0.24), 0.17 (0.12–0.23) and 0.17 (0.11–0.32) mg/L, respectively, in the 3 MIU/8 h group and 0.20 (0.11–0.35), 0.24 (0.12–0.44), 0.24 (0.10–0.49) and 0.23 (0.11–0.44) mg/L, respectively, in the 5 MIU/8 h group, with no differences between the two groups at any time. Renal impairment during nebulized treatment was observed in three patients in each group, but was unlikely to be related to colistimethate sodium treatment. Nebulized colistimethate sodium therapy was well tolerated and no bronchospasms or neurotoxicity events were observed. Conclusions In this limited observational case series of critically ill patients with HAP or VAP treated with high doses of nebulized colistimethate sodium, systemic exposure was minimal and the treatment was well tolerated.