Davide Bindellini (1, 2), David Busse (1, 2), Philipp Simon (3), Robin Michelet (1), David Petroff (4), Linda B.S. Aulin (1), Christoph Dorn (5), Markus Zeitlinger (6), Hermann Wrigge (7), Wilhelm Huisinga (8), Charlotte Kloft (1)
(1) Dept. of Clinical Pharmacy and Biochemistry, Institute of Pharmacy, Freie Universitaet Berlin, Germany, (2) and Graduate Research Training program PharMetrX, Germany, (3) Dept. of Anesthesiology and Intensive Care, Faculty of Medicine, University of Augsburg, Germany and Integrated Research and Treatment Center (IFB), Adiposity Diseases, University of Leipzig, Germany, (4) Clinical Trial Centre Leipzig, University of Leipzig, Germany, (5) Institute of Pharmacy, University of Regensburg, Germany, (6) Department of Clinical Pharmacology, Medical University of Vienna, Vienna, Austria (7) Bergmannstrost Hospital Halle, Department of Anaesthesiology, Intensive Care and Emergency Medicine, Pain Therapy, Halle, Germany (8) Institute of Mathematics, University of Potsdam, Germany
Objectives: Cefazolin (CEZ) is frequently used for the treatment of skin and soft tissue infections (SSTI), e.g. after surgery. Obesity has been identified as a risk factor for surgical site infections, but data for a quantitative evaluation of CEZ pharmacokinetics (PK) in the interstitial space fluid (ISF) of subcutaneous (s.c.) adipose tissue (i.e. the target site for the treatment of SSTI) in obese patients are scarce. Additionally, CEZ dosing recommendations and guidelines are currently inconsistent for both obese and nonobese patients. The aims of this mechanistic model-based analysis were (i) to identify covariates explaining and providing mechanistic interpretation for PK differences between obese and nonobese patients, and to quantify (ii) CEZ protein binding kinetics and (iii) target-site penetration in obese versus nonobese surgical patients.
Methods: Previously published data [1, 2] from 15 obese (BMImedian=52.6 kg/m2) and 15 nonobese patients (BMImedian=26.0 kg/m2), receiving a single dose of 2000 mg CEZ (30-min intravenously) for infection prophylaxis before abdominal surgery [1, 3], were included in the analysis. PK data from dense sampling were available over 8 h in plasma (ntotal=240) and via microdialysis in the ISF of s.c. adipose tissue (unbound target-site concentrations, ntotal=591). Additionally, ultrafiltration was performed on part of the plasma samples to measure unbound CEZ plasma concentration (ntotal=120). Nonlinear mixed-effects (NLME) PK modelling was applied on all data simultaneously. Covariate selection and inclusion was based on mechanistic principles, such as allometric scaling. Several body size descriptors (BSDs) were tested for “theory-based” allometric scaling of volumes (V) and flows (fixed exponents 1 and 0.75, respectively), as well as semi-mechanistic approach based on lean body weight (LBW), fat mass (FM) and a drug- and PK-related scaling parameter R (LBW/FM approach) [4]. In the LBW/FM approach, central V is scaled with LBW and peripheral V is scaled with LBW and FM: The drug-related parameter R represents the fraction of the peripheral V scaled with LBW, whereas the remaining fraction (1-R) is scaled with FM. The comparison between models including different BSDs or LBW/FM approach was performed using Akaike information criterion. Concerning clearance, beside several BSDs, creatinine-based estimates of glomerular filtration rate and age were tested. The application of NLME PK modelling allowed to characterise CEZ protein binding kinetics: Both linear and saturable binding models were tested. Then, the developed model was leveraged to evaluate differences in target-site penetration index (PI), as the ratio of unbound CEZ AUC0-t (t=4 or 8 hours) in target-site:plasma (PI=fAUCtarget, 0-t/fAUCpla, 0-t), between patient groups.
Results: A two-compartment NLME PK model (with target-site concentrations attributed to the peripheral compartment) best described CEZ concentrations over time. Interindividual variability was implemented on all parameters assuming lognormal distribution, except for the drug-related parameters CEZ dissociation constant (Kd) and R. Allometric scaling of distribution parameters (central and peripheral V and intercompartmental flow) was performed using the LBW/FM approach [4], whereas no covariates were found to have a significant influence on CEZ clearance. The LBW/FM approach showed that LBW has a larger impact on CEZ distribution than FM (R=0.764, 95%CI=0.668-0.860), reflecting CEZ hydrophilicity. CEZ plasma protein binding was best characterised via a nonlinear saturable binding equation, with an estimated maximum binding capacity (Bmax) of 247 mg/L (95%CI=207-286 mg/L) and Kd of 65.3 mg/L (95%CI=49.9-80.7 mg/L). PI was evaluated at 4 and 8 hours: Both at 4 hours (p=0.683) and at 8 hours (p=0.345), no statistical significant differences were observed between obese and nonobese patients.
Conclusions: The implementation of the LBW/FM approach described CEZ distribution differences between obese and nonobese patients well. Despite covariates on CEZ clearance being expected and mechanistically motivated, their inclusion led to unstable models or unidentifiable parameters. CEZ plasma protein binding results were in line with previous studies that suggested saturable binding [1, 5]. Lastly, no impact of obesity was found on CEZ PI both at 4 and 8 hours.
References:
[1] Dorn C et al. JAC (2021) 76, 8, 2144-2120
[2] Dorn C et al. J Chromatogr B (2019) 1118-1119, 51-54
[3] Simon P et al. Contemp Clin Trials Commun (2019) 15, 100375
[4] Huisinga W et al. ASCPT (2012) 1, 9, 1-10
[5] Brill M et al. JAC (2014) 69, 3, 715-723
Reference: PAGE 30 (2022) Abstr 10177 [www.page-meeting.org/?abstract=10177]
Poster: Drug/Disease Modelling - Infection