Sami Ullah (1)*, Michael Zoller (2)*, Usman Arshad (1), Mikayil Huseyn-Zada (2), Uwe Fuhr (1), Johannes Zander (3)#, Max Taubert (1)#
(1) University of Cologne, Faculty of Medicine and University Hospital Cologne, Center for Pharmacology, Department I of Pharmacology, Cologne, Germany (2) Department of Anaesthesiology, Hospital of the Ludwig Maximilians University of Munich, Munich, Germany (3) Institute of Laboratory Medicine, Hospital of the Ludwig Maximilians University of Munich, Munich, Germany *, #: equal contribution
Introduction
Pathophysiological changes in ICU patients with severe infections lead to a high pharmacokinetic variability (1) and risk of improper exposure to antibiotics. While inter-individual pharmacokinetic differences have been explored for cefepime (2), less information is available on intra-individual variability during the treatment. Especially in critically ill subjects, such variability might be of relevance since a stable exposure to antibiotics is endangered due to unstable renal function, concomitant medication, fluid resuscitation and further disease-/treatment-related factors. Consequently, ambiguities on the choice of optimum dosing regimens for drugs with a rather narrow therapeutic window, such as cefepime (2, 3), result. To individualize cefepime dosing during the treatment, predictors of pharmacokinetic parameters may be useful. Recently, Jonckheere et al. (3) concluded that Cockcroft-Gault creatinine clearance (eCRCL) might be a better predictor of cefepime clearance than measured creatinine clearance (mCRCL). This finding is challenged by the invalidity of eCRCL in presence of unstable creatinine kinetics (4) and potential confounding by components of the equation (body weight, sex and serum creatinine concentrations). Therefore, it is unclear whether eCRCL should be employed to predict time-dependent changes of cefepime clearance and whether intra-individual variability poses a clinically relevant risk to treatment efficacy.
Objectives:
- To evaluate inter- and intra-individual variability of cefepime pharmacokinetics in critically ill subjects.
- To evaluate eCRCL as a predictor of intra-individual changes in cefepime clearance during treatment.
Methods:
A total of 15 critically ill patients (median age and body weight of 62 [range 22-94] years and 71 [45-120] kg) treated in a medical-surgical ICU were enrolled after approval by the local ethics committee and informed consent from patients or legal representatives. Five patients received intravenous hemodialysis throughout the four study days. Cefepime was administered as 30-min intravenous infusions (median dose 2g) twice daily as per local guidelines. Post-dose plasma samples (median 12, 6, 7, 6 per subject on days 1 to 4) were obtained on four consecutive days and cefepime concentrations were quantified via validated HPLC (5). A compartmental model was developed using NONMEM 7.4.3 (6), whereas different models were compared in terms of changes in objective function value (dOFV), unexplained variability and goodness of fit. Changes in clearance (CL) and central volume of distribution (V1) throughout the treatment course were explored via random (intra-individual variability) and fixed effects (systematic changes in parameters). For patients on dialysis, CL and V1 were estimated separately. Based on the base model, the influence of covariates (age, weight, sex, height, BMI, urea, serum creatinine, 24 hours urine creatinine and the corresponding mCRCL observed once daily) on cefepime clearance was evaluated. Assuming Cockcroft-Gault creatinine clearance might represent relationships to body weight, sex and baseline kidney function, a covariate model with eCRCL calculated on each treatment day was compared to a model comprising solely eCRCL obtained at study start.
Results:
387 plasma concentrations (median 61 mg/L [range 0.6-848 mg/L]) were available for analysis. A two-compartment model with linear elimination best described cefepime concentrations in dialysis and non-dialysis patients. While inclusion of eCRCL estimated for each individual day relevantly improved the model (dOFV -9, IIV reduced by 36%, IOV increased by 1.4%), no model improvement was observed for mCRCL (dOFV -0.1, IIV reduced by 2%, IOV reduced by 0.1%). The best model, however, was the one based on eCRCL estimates obtained on the first day (dOFV of -13, IIV reduced by 45%, IOV increased by 1.7%) (comparison to base model). In the final model, population CL (median of 3.67 L/h) and V1 (median of 3.89 L) showed a slightly more pronounced inter- than intra-individual variability (38% versus 36% and 68% versus 64%) and a clear reduction (median CL of 2.35 L/h, median V1 of 1.86L) in patients on dialysis.
Conclusions:
When accounting for Cockcroft-Gault creatinine clearance, residual inter- and intra-individual variability were of similar magnitude. Adapting the therapy to changes in serum creatinine concentrations using Cockcroft-Gault or measured creatinine clearance during therapy seems inappropriate.
References:
[1] Blot SI, Pea F, Lipman J. 2014. The effect of pathophysiology on pharmacokinetics in the critically ill patient — Concepts appraised by the example of antimicrobial agents. Adv Drug Deliv Rev 77:3–11.
[2] Chapuis TM, Giannoni E, Majcherczyk PA, Chioléro R, Schaller M-D, Berger MM, Bolay S, Décosterd LA, Bugnon D, Moreillon P. 2010. Prospective monitoring of cefepime in intensive care unit adult patients. Crit Care 14:R51.
[3] Jonckheere S, De Neve N, De Beenhouwer H, Berth M, Vermeulen A, Van Bocxlaer J, Colin P. 2016. A model-based analysis of the predictive performance of different renal function markers for cefepime clearance in the ICU. J Antimicrob Chemother 71:2538–2546.
[4] Chen S, Greene T, Kusek J, Beck G. 2013. Retooling the creatinine clearance equation to estimate kinetic GFR when the plasma creatinine is changing acutely. J Am Soc Nephrol 24:877–88.
[5] Zander J, Maier B, Suhr A, Zoller M, Frey L, Teupser D, Vogeser M. 2015. Quantification of piperacillin, tazobactam, cefepime, meropenem, ciprofloxacin and linezolid in serum using an isotope dilution UHPLC-MS/MS method with semi-automated sample preparation. Clin Chem Lab Med 53:781–91.
[6] Beal S, Sheiner L, Boeckmann A, Bauer R. 2018. NONMEM users guides. ICON Dev Solut.
Reference: PAGE 28 (2019) Abstr 9109 [www.page-meeting.org/?abstract=9109]
Poster: Drug/Disease Modelling - Infection