Study Question: Can we identify patient factors that influence pharmacokinetics of IV polymyxin B in the critically ill and use them to develop dosing regimens?
Study Description: This was an observational pharmacokinetic study in which physicians individually determined the need and dosage for polymyxin B on each patient enrolled. Blood and urine samples were collected before after the completing the infusion of polymyxin B. Individual patient characteristics of total body weight (TBW), lean body weight (LBW), sex, age, creatinine clearance (CrCl), albumin, and APACHE II score were used in pharmacokinetic modeling. Monte Carlo simulations were then also performed for 6 dosing regimens with 5,000 virtual subjects.
Results: Twenty-four patients were evaluated with polymyxin B given at doses of 0.45 – 3.38 mg/kg/day, with every 12 hour dosing in 23 patients and every 24 hour dosing in one patient. The median CrCl was 33 mL/min with 2 patients requiring continuous renal replacement therapy (CRRT). The median TBW was 62.5 kg (with one patient morbidly obese at 250 kg), and the median AUC over 24 hours was 66.9 ± 21.6 mg/hr/L (range, 16.4 – 117 mg/hr/L).
Between-subject variability for both polymyxin B clearance and volume of distribution were reduced in allometric scaling by TBW and both linear and allometric scaling of LBW. Neither the unscaled nor scaled clearance estimate of polymyxin B showed any relationship to CrCl, APACHE II score, sex, age or serum albumin concentration. Six different dosing strategies are outlined and analyzed by Monte Carlo simulation with respect to achievable Cmin, Cmax, and 24-hour AUC. Simulation analysis showed that dosing regimens without a loading dose had lower drug concentrations on day 1 than at steady state.
Conclusion(s): Total body weight, not creatinine clearance, influences pharmacokinetics of polymyxin B in the critically ill patient.
Perspective: Polymyxin B is an important agent in the antimicrobial armamentarium to bear in mind with increasing prevalence of multidrug resistant pathogens. The authors used a Monte Carlo simulation modeled after pharmacokinetic data from 24 patients to simulate dosing in the critically ill population. The authors suggest using a loading dose to achieve optimal polymyxin B plasma concentrations quickly in this vulnerable patient population. Additionally, the authors conclude that a dose of 1.5 mg/kg q12h would reach a AUC/MIC of 20 in about half the patients when the MIC of the causative pathogen is 2 mg/L. According to this model, polymixin B does not require renal dose adjustments in patients with or without CRRT, unlike colistin. However, nephrotoxicity associated with polymyxin B is a significant adverse effect that clinicians must consider when using this medication, and although CrCl does not determine dosing strategies, poor renal function increases tubular cell exposure to polymycin B and can increase the incidence of nephrotoxicity. Colistin is currently the polymyxin of choice at many institutions due to the overall lack of data describing the use of polymyxin B. This article is an important step in describing a dosing strategy for polymyxin B and increasing clinician knowledge about it as an alternative agent.