To evaluate the efficacy and safety of colistimethate sodium (CMS) for the treatment of critical patients infected by pan-drug resistant Acinetobacter baumannii (PDR-AB) or pan-drug resistant Pseudomonas aeruginosa (PDR-PA).

321 isolates of PDR-AB and 204 isolates of PDR-PA from critical patients admitted to 35 intensive care units (ICUs) of grade two or above were collected from the Anhui Antimicrobial Resistance Investigation Net (AHARIN) program from September 2012 to September 2015, while the minimal inhibitory concentrations (MIC) of colistin were determined by the E-test. A series of Monte Carlo simulations was performed for CMS regimens (1 MU q8h, 2 MU q8h, and 3 MU q8h, and MU meant a million of unit), and the probability of achieving a 24-hour area under the drug concentration time curve (AUC₂₄)/MIC ratio > 60 and risk of nephrotoxicity for each dosing regimen was calculated. Each simulation was run over three CL_Cr ranges: < 60, ≥ 60-90, ≥ 90-120 mL/min. The probability of target attainment (PTA) for the AUC₂₄/MIC ratio was calculated using the partial MIC value, while the cumulative fraction of response (CFR) was determined by integrating each PTA with the MIC distributions, the value greater than or equal to 90% or more than 80% was set as the optimal dosing regimen or suboptimal dosing regimen respectively. The probability of average 24-hour serum concentrations up to 4 mg/L for three dosage regimens was used to predict the risks of nephrotoxicity.

All 321 isolates of PDR-AB and 204 isolates of PDR-PA were susceptible to colistin, the MIC_50/90 against PDR-AB were 0.5 mg/L and 1.0 mg/L, and those against PDR-PA were 0.5 mg/L and 1.5 mg/L, respectively. When recommended dose (1 MU q8h) was used for patients with CL_Cr of < 60 mL/min, high CFR value (89.78% for PDR-AB, 81.06% for PDR-PA) were obtained, but with a high risks of nephrotoxicity (> 32.51%). Moreover, low value of PTA (< 66.56%) was yielded for isolates with MIC of ≥ 1 mg/L. Recommended dose also yielded a low CFR value (56.97%-69.31% for PDR-AB, 44.76%-56.94% for PDR-PA) in patients with CL_Cr of ≥ 60-120 mL/min. When dose was increased to 2 MU q8h, CFR (77.45%-92.87%) and the risks of nephrotoxicity (< 0.15%) was optimal for patients with CL_Cr ≥ 60-120 mL/min, but low value of PTA (< 75.36%) was also yielded for isolates with MIC of ≥ 1 mg/L. The most aggressive dose of 3 MU q8h provided high CFR (> 89.24%) even in patients with CL_Cr ≥ 90-120 mL/min, and PTA was < 76.20% only for isolates with MIC of ≥ 1.5 mg/L, but this dosing scheme was associated with unacceptable risks of nephrotoxicity (> 33.68%).

Measurement of MIC, individualized CMS therapy and therapeutic drug-level monitoring should be considered to achieve the optimal drug exposure and ensure the safety of CMS.

Colistin; Monte Carlo simulation; Pharmacokinetic; Pharmacodynamic; Toxico-dynamics