Determination by LC-MS/MS of colistin A and B in plasma and ultrafiltrate from critically ill patients undergoing continuous venovenous haemodiafiltration

Background Colistin is a 50-years old antibiotic, whose use was ceased in the ’70s and recently resumed as a “salvage therapy” against multidrug-resistant Gram-negative bacteria, such as Pseudomonas aeruginosa and Acinobacter baumanii. The narrow therapeutic range of colistin makes the choice of its correct dosage crucial, and monitoring of blood concentration is occasionally necessary for critically ill patients, including intensive care patients subjected to continuous renal replacement therapy. Methods Two LC-MS/MS methods were developed and fully validated for the quantitative determination of colistin A and B in plasma and dialysis ultrafiltrate samples, ultimately arising from four patients undergoing continuous venovenous haemodiafiltration. Results The developed methods proved to be both specific and selective. They showed good fit and linearity over the entire range of interest. Trueness and accuracy proved satisfactory. Both methods have excellent intra-assay precision (CV% were lower than 10%) and limit of detection (LOD) values in the 20-100 ng/mL range, about 1-2 orders-of-magnitude below the concentrations commonly detected in real samples. The mean sieving coefficient (SC) values, measured after 10 min of continuous venovenous haemodiafiltration (CVVHDF), were 0.42 for colistin A and 0.48 for colistin B. SC values proved to be quite stable for 24 h, but then declined respectively to 0.24 for colistin A and 0.32 for colistin B, after 48 hours. At the median blood flow and effluent flow rate of 120 mL/min and 28 mL/min, clearance values for colistin B were higher than 15 mL/min. During the entire duration of CVVHDF sessions, the SC and clearance values for colistin A were significantly lower than those for colistin B. Conclusions Two simple methods for the simultaneous determination of colistin A and B have been developed and validated. Their application in the clinical setting demonstrates that CVVHDF treatment lasting 48 hours produce a relatively constant and efficient removal of the drug.