Arkadiusz Adamiszak1,2, Agnieszka Bienert1
1Department of Pharmacology, Poznan University of Medical Sciences, 2Doctoral School, Poznan University of Medical Sciences
Introduction/Objectives: Levofloxacin is a third-generation fluoroquinolone with a broad spectrum, including gram-positive and -negative bacteria [1]. Currently, no dosing standards have been recommended by The European Committee on Antimicrobial Susceptibility Testing (EUCAST), and the doses proposed in Stanford Health Care Antimicrobial Dosing Reference Guide seem to be insufficient for critically ill patients undergoing continuous renal replacement therapy (CRRT) [2,3]. This study aimed to develop a preliminary levofloxacin population pharmacokinetic (PopPK) model and investigate dosages for CRRT and non-CRRT intensive care unit (ICU) patients. Methods: Pharmacokinetic and clinical data was obtained retrospectively from 3 articles describing critically ill adult patients treated with 125 – 500 mg of levofloxacin per 24 – 48h in 0.5 – 1h infusion [4–6]. The PK data was extracted from the graphs using WebPlotDigitizer [7] or directly from the article results. Levofloxacin concentration-time data was analyzed using PopPK modeling according to the SAEM algorithm for non-linear mixed effects models implemented in the nlmixr2 R package [8]. Model selection was based on goodness-of-fit criteria such as Bayesian Information Criteria (BIC), the objective function value (OFV) defined as -2×Log Likelihood, the precision of the parameter estimates, and diagnostic plots. One- and two-compartment models with linear elimination were tested during the model development procedure. Any available clinical variables were examined as potential covariates using a stepwise covariate model procedure (SCM). Simulations of dosing regimens were performed based on the final model using the rxode2 [9] and posologyr [10] R packages. We tested 250/500/750/1000 mg in a 1h infusion every 24 and 12 hours as standard and high dosages, respectively [2]. For each combination, we simulated 2500 patients as 50 patients in a group replicated 50 times. For the probability of target attainment (PTA) analysis, we assumed the ratio of the area under the unbound concentration-time curve in the first 24h to the minimum inhibitory concentration (fAUC0 – 24/MIC) = 80 as a PK/PD targets [11,12]. The PTA = 90% for MIC values between 0.0625 – 8 mg/L was considered an acceptable probability of success [11]. The free fraction of levofloxacin was calculated based on a percentage (~35%) of protein binding mentioned in SmPC [13]. Results: The study included 25 adult ICU patients and 77 plasma samples, of which 32 were from 16 CRRT patients. The median (range) of the continuous covariates tested for the entire population was 57.0 years (23.0 – 76.0) for age and 95.6 kg (60.0 – 189.0) for weight. Additionally, we tested the influence of creatinine clearance (ClCr) estimated by the Cockcroft-Gault formula (median = 77.7 mL/min and range = 55.0 – 235.0 mL/min) and creatinine serum (median = 1.0 mg/dL and range = 0.7 – 1.6 mg/dL) on non-CRRT patents clearance (CLnon-CRRT). In turn, for patients undergoing CRRT, we tested the influence of blood flow rate (median = 9 L/h and range 5.4 – 12.0 L/h) and effluent flow rate (median = 1.5 and range 0.84 – 2.38 L/h) on CRRT patients clearance (CLCRRT). The two-compartment model with separate clearances for CRRT and non-CRRT patients described the PK data well. The mean population parameter estimates (%RSE) for CLnon-CRRT was 4.51 L/h (6.63%), CLCRRT was 2.23 L/h (7.77%), central compartment volume (Vc) was 35.01 L (24.70%), inter-compartmental clearance (Q) was 42.10 L/h (22.49%) and peripheral compartment volume (Vp) was 46.46 L (16.82%). Inclusion of the impact of creatine clearance and CRRT (as a categorical variable) on CLnon-CRRT and Q, respectively, significantly improved model fitting and decreased the unexplained inter-individual variability (expressed as % of the coefficient of variation, %CV) of CLnon-CRRT from 37% to 18% and Q from 54% to 39%. The remaining %CVs were 30% for CLCRRT, 80% for V2, and 51% for V2. The residual error equal to 7.71% (%CV) was estimated using a proportional error. According to simulations for fAUC0 – 24/MIC = 80, ICU patients undergoing CRRT are able to achieve assumed PTA = 90% using at least one of the tested dosages only for MIC = 1 mg/L. The same observation is for non-CRRT patients with ClCr between 50 – 150 mL/min. In turn, tested levofloxacin doses are sufficient for non-CRRT patients with ClCr between 180 – 240 mg/L for MIC up to 0.5 mg/L. Conclusions: The results of our study indicate a need for the establishment of levofloxacin doses for CRRT and increased ClCr patients. Assuming the mentioned PK/PD and PTA targets, the MIC = 2 mg/L should be considered resistance. When investigating higher levofloxacin doses than currently registered, the increased risks of side and toxic effects characteristic of fluoroquinolones should be taken into account. Our results should be used with caution, given the scarce number of patients included in the model.
[1] Podder, V.; Patel, P.; Sadiq, N.M. Levofloxacin. In StatPearls; StatPearls Publishing: Treasure Island (FL), 2025. [2] EUCAST Clinical breakpoints and dosages 2024 (v 14.0 (https://www.eucast.org/clinical_breakpoints) [3] Stanford Antimicrobial Safety & Sustainability Program (https://med.stanford.edu/bugsanddrugs/guidebook.html) [4] Sánchez Navarro, A.; Gandarillas, C.-I.C.; Lerma, F.A.; Menacho, Y.A.; Domínguez-Gil, A. Pharmacokinetics and Pharmacodynamics of Levofloxacin in Intensive Care Patients. Clin Pharmacokinet 2005, 44, 627–635, doi:10.2165/00003088-200544060-00004. [5] Guenter, S.G.; Iven, H.; Boos, C.; Bruch, H.; Muhl, E. Pharmacokinetics of Levofloxacin During Continuous Venovenous Hemodiafiltration and Continuous Venovenous Hemofiltration in Critically Ill Patients. Pharmacotherapy 2002, 22, 175–183, doi:10.1592/phco.22.3.175.33546. [6] Malone, R.S.; Fish, D.N.; Abraham, E.; Teitelbaum, I. Pharmacokinetics of Levofloxacin and Ciprofloxacin during Continuous Renal Replacement Therapy in Critically Ill Patients. Antimicrob Agents Chemother 2001, 45, 2949–2954, doi:10.1128/AAC.45.10.2949-2954.2001. [7] Rohatgi, A. WebPlotDigitizer User Manual Version 4.6. 2022. [8] Fidler, M.; Xiong, Y.; Schoemaker, R.; Wilkins, J.; Wang, W.; Trame, M.; Xu, H.; Harrold, J.; Denney, B.; Papathanasiou, T.; et al. Nlmixr2: Nonlinear Mixed Effects Models in Population PK/PD 2024. [9] Fidler, M.L.; Wang, W.; Hindmarsh, A.; Srinivasan, A.; Al-Mohy, A.H.; Denney, B.; Moler, C.; Cooley, D.; Schmidt, D.; Hairer, E.; et al. Rxode2: Facilities for Simulating from ODE-Based Models 2025. [10] Leven, C.; Fidler, M.; Comets, E.; Lavenu, A.; Lavielle, M. Posologyr: Individual Dose Optimization Using Population Pharmacokinetics 2025. [11] Setiawan, E.; Abdul-Aziz, M.-H.; Cotta, M.O.; Susaniwati, S.; Cahjono, H.; Sari, I.Y.; Wibowo, T.; Marpaung, F.R.; Roberts, J.A. Population Pharmacokinetics and Dose Optimization of Intravenous Levofloxacin in Hospitalized Adult Patients. Sci Rep 2022, 12, 8930, doi:10.1038/s41598-022-12627-1. [12] Roberts, J.A.; Cotta, M.O.; Cojutti, P.; Lugano, M.; Rocca, G.D.; Pea, F. Does Critical Illness Change Levofloxacin Pharmacokinetics? Antimicrobial Agents and Chemotherapy 2016, 60, 1459–1463, doi:10.1128/aac.02610-15. [13] Tavanic – Referral | European Medicines Agency (EMA) Available online: https://www.ema.europa.eu/en/medicines/human/referrals/tavanic (accessed on 10 March 2025).
Reference: PAGE 33 (2025) Abstr 11315 [www.page-meeting.org/?abstract=11315]
Poster: Drug/Disease Modelling - Infection