Population pharmacokinetics of fluconazole in critically ill patients: an individual patient data meta-analysis.
My Luong Vuong (1,2)*, Omar Elkayal (1)*, Jan-Willem C. Alffenaar (3,4,5), Yves Debaveye (6,7), Joost Wauters (8,9), Beatrijs Mertens (1,10), Jasper M. Boonstra (11), Jason A. Roberts (12,13,14), Raoul Bergner (15), Steven Buijk (16), Indy Sandaradura (17,18,19), Roger J. Brüggemann (20,21), Jeroen A. Schouten (22), Isabel Spriet (1,10)+, Erwin Dreesen (1)+ *Shared co-first authorship, +Shared co-senior authorship
(1) Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium, (2) Department of Family Medicine and Population Health, University of Antwerp, Antwerp, Belgium, (3) The University of Sydney, Faculty of Medicine and Health, School of Pharmacy, Sydney, Australia, (4) Westmead Hospital, Sydney, Australia, (5) Marie Bashir Institute of Infectious Diseases and Biosecurity, University of Sydney, Sydney, Australia, (6) Department of Cellular and Molecular Medicine, KU Leuven, Leuven, Belgium, (7) Intensive Care Unit, University Hospitals Leuven, Leuven, Belgium, (8) Department of Microbiology, Immunology and Transplantation, KU Leuven, Leuven, Belgium, (9) Medical Intensive Care Unit, University Hospitals Leuven, Leuven, Belgium, (10) Pharmacy Department, University Hospitals Leuven, Leuven, Belgium, (11) University of Groningen, University Medical Center Groningen, Department of Clinical Pharmacy and Pharmacology, Groningen, the Netherlands, (12) University of Queensland Centre for Clinical Research (UQCCR), The University of Queensland, Brisbane, Queensland, Australia, (13) Departments of Pharmacy and Intensive Care Medicine, Royal Brisbane and Women’s Hospital, Brisbane 4029, Australia, (14) Division of Anaesthesiology Critical Care Emergency and Pain Medicine, Nîmes University Hospital, University of Montpellier, 30029 Nîmes, France, (15) Division of Rheumatology, Department of Medicine A, Ludwigshafen Medical Center, Ludwigshafen, Germany, (16) Department of Surgery, IJsseland Hospital, Capelle aan den Ijssel, the Netherlands, (17) Centre for Infectious Diseases and Microbiology, Westmead Hospital, (18) Faculty of Medicine and Health, Westmead Clinical School, The University of Sydney, (19) Institute of Clinical Pathology and Medical Research, New South Wales Health Pathology, Westmead Hospital, Sydney, Australia, (20) Radboud University Medical Center, Radboud Institute for Health Sciences, Department of Pharmacy, Nijmegen, The Netherlands, (21) Center of Expertise in Mycology Radboudumc/CWZ, Nijmegen, The Netherlands, (22) Department of Intensive Care, Canisius Wilhelmina Hospital, Nijmegen, The Netherlands
Introduction: Fluconazole is a triazole antifungal commonly used to prevent and treat invasive candidiasis [1]. It is largely eliminated through renal excretion and undergoes tubular reabsorption [2,3]. Its plasma protein binding is about 11 to 12% [3]. The pharmacokinetic/pharmacodynamic (PKPD) target of unbound fluconazole is the ratio between the area under the concentration-time curve to the minimum inhibitory concentration (MIC) of 100, which corresponds to trough concentrations (Cmin) of 7.5-15 mg/L for MIC values of 2-4 mg/L [4-6]. Recent studies have shown large variability in exposure and significant underexposure to fluconazole in critically ill patients, including those with augmented renal clearance, renal replacement therapy and obesity, suggesting the need for dose optimization [4].
Objectives: The objectives of our work were
- to develop a population PK (popPK) model for intravenous (IV) fluconazole in critically ill patients, and
- to identify covariates with a relevant impact on fluconazole PK.
Methods: An individual patient data meta-analysis (IPDMA) was performed including data from eight prospective observational PK studies of IV fluconazole in critically ill patients admitted to the intensive care unit (ICU) [4,7-13]. A popPK analysis was performed (NONMEM 7.5) to describe the PK of IV fluconazole. Different model structures were explored. Interindividual variability (IIV), and residual unexplained variability (RUV) were quantified. The base model was selected based on objective function value comparisons (difference ≥3.84 points; P ≤0.050), plausibility and precision of parameter estimates, and goodness-of-fit plots.
A final model including covariate effects was built through two-way stepwise covariate modelling (αforward=0.010, αbackward=0.001). The tested time-varying covariates were body weight, body mass index, fat-free mass (FFM; approximated by lean body weight calculated using Boer’s formula [14]), and the estimated glomerular filtration rate (calculated using the Chronic Kidney Disease Epidemiology Collaboration equation; eGFRCKD-EPI). Median imputation was used for handling covariate missingness at the patient level. A prediction-corrected visual predictive check (pcVPC; 1,000 replicates) was used to evaluate the final model.
Results: Eight clinical centres shared PK data from a total of 177 critically ill patients admitted to the ICU, contributing 1616 plasma samples of which 392 were taken at trough (24.3%). Cmin target attainment of 7.5 vs. 15 mg/L were 64% vs. 18% on day 1 and 81% vs. 53% during overall treatment period.
The popPK of IV fluconazole was best described by a two-compartment model with linear elimination. The estimated clearance when on continuous renal replacement therapy (CLCRRT) was 1.71 L/h [8.1%] (typical value [relative standard error]), and 0.66 L/h [4.9%] when not on CRRT (CLno CRRT). Intercompartmental clearance (Q) was 9.40 L/h [31.4%], the volume of distribution in the central compartment (Vc) was 42.5 L [7.5%], and the volume of distribution in the peripheral compartment (Vp) was 7.4 L [38.1%]. The IIV in CLCRRT, CLno CRRT and Vc were estimated at 45.6% (coefficient of variation; CV), 49.1%, and 55.1%CV, resp. A combined additive (±0.796 mg/L) and proportional (13.9%CV) error model described the RUV.
Equations 1 and 2 show how fluconazole Vc and CLno CRRT of patient i at time point j increase with FFM and eGFRCKD-EPI, resp.
Vc(i,j) = 42.5 L x (FFM(i,j)/58.05 kg)1.26 x ehi
with hi ~ N(0, 0.265) (Eq.1)
CLno CRRT(i,j) = 0.66 L/h x (eGFRCKD-EPI(i,j)/91.55 ml/min/1.73 m2)0.696 x ehi
with hi ~ N(0, 0.216) (Eq.2)
With FFM increasing from 31.97 kg to 87.44 kg, fluconazole Vc increased from 20.0 L to 71.2 L, resulting in an increase in elimination half-life (t1/2,e) from 28.8 h to 82.1 h. With eGFRCKD-EPI increasing from 6.38 ml/min/1.73 m2 to 257.30 ml/min/1.73 m2, fluconazole CLno CCRT increased from 0.10 L/h to 1.35 L/h, with a decrease in t1/2,e from 346.0 h to 25.7 h.
Conclusion: An IPDMA of IV fluconazole popPK model was performed based on the largest pooled patient population to date. Our analysis identified eGFRCKD-EPI and FFM as statistically significant and clinically relevant predictors of fluconazole PK. The developed model will be the basis for a simulations study to identify a fluconazole dosing strategy for critically ill patients.
References
[1] Tissot et al. Haematologica. 2017; 102(3):433-444.
[2] Bellmann et al. Infection. 2017; 45(6):737-779.
[3] Pfizer. Summary of Product Characteristics (SmPC) Diflucan. 2012.
[4] Boonstra et al. Antimicrob Agents Chemother. 2021; 65(3).
[5] ECoAST. Rationale for EUCAST clinical breakpoints Fluconazole. v3. 2020.
[6] Rodríguez-Tudela et al. Antimicrob Agents Chemother. 2007; 51(10):3599-3604.
[7] Alobaid et al. Antimicrob Agents Chemother. 2016; 60(11):6550-6557.
[8] Bergner et al. Nephrol Dial Transplant. 2006; 21(4):1019-1023.
[9] Buijk et al. Intensive Care Med. 2001; 27(1):115-121.
[10] Muilwijk et al. Antimicrob Agents Chemother. 2020; 64(10).
[11] Sandaradura et al. Eur J Clin Microbiol Infect Dis. 2021; 40(7):1521-1528.
[12] Sinnollareddy et al. Critical Care. 2015; 19.
[13] Van Daele et al. Microorganisms. 2021; 9(10).
[14] Boer. Am J Physiol. 1984; 247(4 Pt 2):F632-636.