Clinical Infectious Diseases 870 • CID 2017:65 (1 September) • CORRESPONDENCE Clinical Infectious Diseases ® 2017;65(5):870 Dosing Colistin Properly: Let’s Save “Our Last Resort Old Drug!” Dear Editor, Nation and coworkers recently pub- lished algorithms for individualized colistin dosing to achieve optimal plasma drug concentration in critically ill patients with a wide range of renal function (RF) and those on renal replacement therapy (RRT) [1]. e updated dosing sugges- tions are greatly improved compared with previously published algorithms from the same authors based on interim analysis [2]. Indeed, the updated version is based on a large cohort of patient population and includes specific information on how to estimate the maintenance colis- tin dose also in patients receiving or not RRT or undergoing sustained low-effi- ciency dialysis. Authors attempted to bal- ance potential antibacterial activity and nephrotoxicity targeting probability of attainment rates of >80% for colistin con- centration ≥2 and <30% for colistin con- centration ≥4 mg/L. When the developed algorithms were applied back to individ- ual patients stratified to different degrees of RF, it was shown that more than 80% of patients with creatinine clearance (Cr- Cl) <80 mL/min achieved colistin con- centration ≥2 mg/L, whereas less than 30% achieved more than 4 mg/L. Authors concluded that the study generated cli- nician-friendly algorithms to be used in critically ill patients over a wide range of RF. Some methodological drawbacks may partially challenge these conclusions. It shouldn’t be correct to validate the good- ness of the proposed equations through a “back application” in the same popula- tion used to develop the models but in an independent cohort of patients. Actually, the risk might be to develop algorithms working very well in the development populations but largely failing in real life. As shown in Table 1, it seems that hyperfiltrating patients (Cr-Cl ≥300 mL/ min) were included in the study, but no specific colistin dosing suggestions have been provided for this population. is may eventually explain why the pro- posed equations largely failed in patients with Cr-Cl >80 mL/min, with less than 40% of subjects achieving colistin con- centration ≥2 mg/L. Indeed, it is well known that augmented renal clearance is frequent in critically ill patients and can result in elevated renal elimination and subtherapeutic plasma antibiotic concentrations [3]. Similarly, looking at Table 1, patients enrolled in the study weighted up to 130 kg, rendering virtu- ally impossible to verify if the equations can fit for obese patients too. Finally, the authors do not consider as an additional covariate the levels of serum albumin. Indeed, colistin has a protein binding of 40–75% and colistin A has a concentra- tion dependent binding [4]. erefore, it cannot be excluded that the application of the proposed equations for the prediction of colistin loading/maintenance dosing to hypoalbuminemic patients may result in higher than expected colistin clearance, suboptimal drug exposure and develop- ment of colistin resistance [5]. Note Potential conflicts of interest. Both authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. Alberto Corona, 1 and Dario Cattaneo 2 1 Intensive Care Unit and 2 Unit of Clinical Pharmacology, Luigi Sacco Hospital, ASST Fatebenefratelli Sacco, University of Milano, Italy References 1. Nation RL, Samira MG, Thamlikitkul V, et al. Dosing guidance for intravenous colistin in criti- cally ill patients. Clin Inf Dis 2017; 64:565–71. 2. Garonzik SM, Li J, Thamlikitkul V, et al. Population pharmacokinetics of colistin methanesulfonate and formed colistin in critically ill patients from a multicenter study provide dosing suggestions for various categories of patients. Antimicrob Agents Chemother 2017; 55:3284–94. 3. Udy AA, Roberts JA, Boots RJ, Paterson DL, Lipman J. Augmented renal clearance: implications for antibacterial dosing in the critically ill. Clin Pharmacokinet 2010; 49:1–16. 4. Mohamed AF, Karaiskos I, Plachouras D, et al. Application of a loading dose of colistin meth- anesulfonate in critically ill patients: population pharmacokinetics, protein binding, and prediction of bacterial kill. Antimicrob Agents Chemother 2012; 56:4241–9. 5. Roberts JA, Pea F, Lipman J. The clinical rele- vance of plasma protein binding changes. Clin Pharmacokinet 2013; 52:1–8. CORRESPONDENCE © The Author 2017. Published by Oxford University Press for the Infectious Diseases Society of America. All rights reserved. For permissions, e-mail: journals.permissions@oup.com. DOI: 10.1093/cid/cix388 Correspondence: A. Corona, Intensive Care Medicine and Infectious Diseases Specialist, Biostatistician and Epidemiology Methodologist, Intensive Care Unit, Luigi Sacco Hospital, ASST Fatebenefratelli Sacco, University of Milano, via GB Grassi 74, 20154 Milano, Italy (corona.alberto@libero.it). Reply to Corona and Cattaneo To the Editor—We thank Corona and Cattaneo [1] for their comments on our recent article in which we described dos- ing guidance for intravenous colistin in critically ill patients [2]. Because of the very substantial inter- patient variability in pharmacokinetics of formed colistin, even at a given creatinine clearance, it was most appropriate to use the data from the entire cohort to derive the dosing suggestions, instead of divid- ing our available cases into a “learning” and a “validation” data set. e patients included in the study were being cared for in 4 centers across 3 continents, and displayed characteristics typical of crit- ically ill patients requiring intravenous colistin [2]. Our report included the largest number of patients in a single analysis. In relation to the influence of renal function, in another study only 4 of 12 patients with creatinine clear- ance >80 mL/minute and receiving Downloaded from https://academic.oup.com/cid/article-abstract/65/5/870/3962720 by guest on 01 June 2020