J Antimicrob Chemother 2020; 75: 2641–2649

doi:10.1093/jac/dkaa187 Advance Access publication 22 May 2020

Optimization of flucloxacillin dosing regimens in critically ill patients

using population pharmacokinetic modelling of total and unbound
concentrations

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

1,2
Nynke G. L. Jager *, Reinier M. van Hest3, Jiao Xie4, Gloria Wong1, Marta Ulldemolins 5
,
Roger J. M. Brüggemann2, Jeffrey Lipman1,6,7 and Jason A. Roberts 1,6–8

1
University of Queensland Centre for Clinical Research, The University of Queensland, Brisbane, Australia; 2Department of Pharmacy,
Radboud University Medical Center and Radboud Institute for Health Sciences, Nijmegen, The Netherlands; 3Department of Hospital
Pharmacy - Clinical Pharmacology, Amsterdam UMC, University of Amsterdam, Amsterdam, The Netherlands; 4Department of
Pharmacy, The Second Affiliated Hospital of Xi’an Jiaotong University, Shaanxi, China; 5Internal Medicine - Infectious Diseases
Departments, Hospital Universitari de Bellvitge, l’Hospitalet de Llobregat, Spain; 6Departments of Pharmacy (J.A.R.) and Intensive Care
(J.L.), Royal Brisbane and Women’s Hospital, Brisbane, Australia; 7Nı̂mes University Hospital, University of Montpellier, Nı̂mes, France;
8
Centre for Translational Anti-infective Pharmacodynamics, School of Pharmacy, The University of Queensland, Brisbane, Australia

*Corresponding author. E-mail: nynke.jager@radboudumc.nl

Received 17 February 2020; returned 2 April 2020; revised 15 April 2020; accepted 16 April 2020

Background: Initial appropriate anti-infective therapy is associated with improved outcomes in patients with
severe infections. In critically ill patients, altered pharmacokinetic (PK) behaviour is common and known to
influence the achievement of PK/pharmacodynamic targets.
Objectives: To describe population PK and optimized dosing regimens for flucloxacillin in critically ill patients.
Methods: First, we developed a population PK model, estimated between-patient variability (BPV) and identified
covariates that could explain BPV through non-linear mixed-effects analysis, using total and unbound concen-
trations obtained from 35 adult critically ill patients treated with intermittent flucloxacillin. Second, we validated
the model using external datasets from two different countries. Finally, frequently prescribed dosing regimens
were evaluated using Monte Carlo simulations.
Results: A two-compartment model with non-linear protein binding was developed and validated. BPV of the
maximum binding capacity decreased from 42.2% to 30.4% and BPV of unbound clearance decreased from
88.1% to 71.6% upon inclusion of serum albumin concentrations and estimated glomerular filtration rate (eGFR;
by CKD-EPI equation), respectively. PTA (target of 100%fT>MIC) was 91% for patients with eGFR of 33 mL/min and
1 g q6h, 87% for patients with eGFR of 96 mL/min and 2 g q4h and 71% for patients with eGFR of 153 mL/min
and 2 g q4h.
Conclusions: For patients with high creatinine clearance who are infected with moderately susceptible patho-
gens, therapeutic drug monitoring is advised since there is a risk of underexposure to flucloxacillin. Due to the
non-linear protein binding of flucloxacillin and the high prevalence of hypoalbuminaemia in critically ill patients,
dose adjustments should be based on unbound concentrations.

Introduction significantly improved clinical outcomes in patients with severe

infections.2–4 However, antibiotic selection and dosing is often
Severe infection is recognized as an important determinant of out- challenging in critically ill patients because of disease complexity,
come for patients in the ICU.1 Initial appropriate anti-infective reduced antibiotic susceptibility of causative pathogens and
therapy, implying both a timely commencement of treatment pathophysiological changes associated with critical illness. These
with a spectrum appropriate for the targeted pathogen and ad- pathophysiological changes, for example altered renal function or
equate exposure to this antimicrobial agent, is associated with hypoalbuminaemia, can influence antibiotic pharmacokinetics

C The Author(s) 2020. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy.
V
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.
org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly
cited. For commercial re-use, please contact journals.permissions@oup.com
2641
Jager et al.

(PK) and consequently the achievement of PK/pharmacodynamic PK analysis

(PD) targets.5,6 We developed an integrated PK model for total and unbound flucloxacillin
Flucloxacillin is a b-lactam antibiotic frequently used in the PK using the non-linear mixed-effects modelling package NONMEM.
treatment of different infections caused by Gram-positive bacteria, Detailed information on methodological model building and validation is
such as penicillinase-producing staphylococci, including MSSA. It is available in the Supplementary data.
metabolized to a limited extent and the unchanged drug and
metabolites are excreted in the urine by glomerular filtration and External validation
tubular secretion. Flucloxacillin is approximately 95% bound to

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

Two external datasets were used for the external validation of the predict-
serum proteins and has a half-life of about 1 h.7 The clinical out- ive performance of the PK model. The first external dataset, the Brisbane
come of b-lactam antibiotics is related to the time the unbound (or dataset, consisted of 28 unbound flucloxacillin concentrations from 20 crit-
free) drug concentration remains above the MIC of the targeted ically ill patients admitted to the ICU of the Royal Brisbane and Women’s
pathogen, fT>MIC.8 In order to devise rational and individualized Hospital. The second external dataset, the Nijmegen dataset, consisted of
34 total and unbound flucloxacillin concentrations from 14 critically ill
dosing regimens of flucloxacillin that ensure sufficient fT>MIC, it is
patients admitted to the ICU of the Radboud University Medical Center in
of importance to understand its PK behaviour and to identify rele- Nijmegen, The Netherlands. Detailed information on the external patient
vant covariates influencing this PK. Because of its high protein bind- cohorts is available in the Supplementary data.
ing and the relationship between unbound drug and outcome, the Total and unbound flucloxacillin plasma concentrations were predicted
unbound PK of flucloxacillin are of special interest. There is, how- by fixing the population PK parameters to the final estimates of the previ-
ever, limited knowledge on the PK of flucloxacillin, especially in crit- ously developed model and setting maximum evaluations (MAXEVAL) to 0.
ically ill patients. The impact of the identified covariates was validated by means of differen-
In view of the above, we performed a population PK study of ces in predictive performance of the population PK model without covari-
ates (the structural model) and with covariates (the final model). To this
total and unbound flucloxacillin in an adult critically ill patient
end, the differences in bias (percentage error) and precision (absolute per-
population. Our objectives were to describe population PK and opti- centage error), calculated using Equations 1 and 2, between the structural
mized dosing regimens for flucloxacillin in critically ill patients. and final model were assessed. A P value of 0.05 was used as a cut-off value
Furthermore, we sought to externally validate the model as well for statistical significance. Since the data were not normally distributed, as
as quantify the value of various patient-specific covariates explain- tested with the Shapiro–Wilk test, the median errors were compared.11
ing altered flucloxacillin PK behaviour.
Cpredi  Cobsi
Percentage errori ¼  100% (1)
Cobsi
Patients and methods
Cpredi  Cobsi
Patients and samples Absolute percentage errori ¼  100% (2)
Cobsi
PK data were obtained from 35 patients treated with intermittent flucloxa-
cillin for a (suspected) infection in the 30 bed tertiary referral ICU at the where Cpredi and Cobsi represent the ith predicted (PRED, population predic-
Royal Brisbane and Women’s Hospital in Brisbane, Australia. This study tion) and observed concentration, respectively.
comprised data from two different sources. The first dataset was from a Also, visual predictive checks (VPCs) were compared visually and it was
previously published prospective clinical study, performed between May evaluated whether there was bias present in the model itself. Models where
and December 2009.5 This dataset consisted of 10 patients with hypoalbu- 0 was included in the IQR of the median percentage error were considered
minaemia (serum albumin concentration 32 g/L). After at least 24 h of unbiased.
treatment, timed samples were collected at 30, 45, 60, 90, 120, 150 and
180 min (q4h dosing regimens) or 300 min (q6h dosing regimens) after the
Monte Carlo dosing simulations
30 min infusion. In all samples, both total and unbound flucloxacillin con-
centrations were measured. Data for the other 25 patients originated from Using the final population PK model, total and unbound flucloxacillin con-
a prospective observational study that was conducted as part of the b-lac- centration–time profiles were predicted based on Monte Carlo simulations
following two frequently prescribed dosing regimens: 1 g q6h and 2 g q4h.
tam therapeutic drug monitoring (TDM) programme, between May 2012
These dosing regimens were simulated for the first 24 h of treatment of a
and July 2014.9 For 19 of these patients, only unbound flucloxacillin con-
typical patient with all median characteristics of the population but with an
centrations were available; for the other 6 patients, both total and unbound
estimated glomerular filtration rate (eGFR) of the 10th (33 mL/min) and
concentrations were available. For the majority of these patients, both mid
90th percentile (153 mL/min) and with a serum albumin concentration of
and trough levels were obtained in one dosing interval. Samples were
the 10th (15 g/L) and 90th percentile (30 g/L).
obtained after at least four doses, between Days 1 and 5.
Second, the PTA, being the percentage of patients with an unbound flu-
cloxacillin concentration remaining above a specified MIC during the whole
Bioanalytical method dosing interval (100%fT>MIC),8 was calculated for different dosing regimens.
Simulations of four frequently prescribed dosing regimens (1 g q6h, 1 g q4h,
The samples obtained for the prospective study were frozen at #80 C and 2 g q6h and 2 g q4h) for patients with all median characteristics of the
analysed within 8 months after the sample was drawn, in accordance with population but with three different eGFR values (33, 96 and 153 mL/min)
the results of the long-term stability investigation.10 Samples collected were performed. Each dosing regimen was simulated 1000 times per eGFR
during the TDM programme were centrifuged within 1 h of collection and value. Since the MIC distribution of flucloxacillin for Staphylococcus aureus
analysed directly. All samples were analysed using a validated HPLC-UV is lacking but suggested to be similar to that of cloxacillin,12 the target MICs
method. were based on the MIC distribution of cloxacillin for S. aureus according to
Detailed information on the bioanalytical methods is available as EUCAST. The epidemiological cut-off value (ECOFF) of this pathogen for
Supplementary data (Appendices 1–3) at JAC Online. cloxacillin is 0.5 mg/L, although the WT distribution of this pathogen shows

2642
Flucloxacillin dose optimization JAC
that the majority (55%) of the isolated S. aureus have an MIC of 0.25 mg/ 0.5–168 h) and a median of 2 h (range: 0.5–6 h) after the most
L.13 Additionally, the PTA for the ECOFF of oxacillin as a surrogate for fluclox- recent administration. Measured total concentrations ranged
acillin12 for MSSA, 2.0 mg/L,13 was also evaluated. from 1.0 to 202 mg/L and measured unbound concentrations
ranged from 0.1 to 30 mg/L.
Statistical analysis
Spearman correlation tests were used to test the correlation between the
protein binding (%) of flucloxacillin, calculated as (1 # unbound
fraction) % 100, and (i) the total flucloxacillin concentration and (ii) the
Protein binding

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

serum albumin concentration. The Shapiro–Wilk test was used to assess The observed protein binding of flucloxacillin (%), calculated per
whether the data were normally distributed. Wilcoxon matched-pairs tests analysed patient sample as (1 # unbound fraction) % 100, ranged
were used to assess the differences between the predictive performance of from 63.4% to 97.2%, with a median of 89.4%. The relationship be-
the structural model and the predictive performance of the final model in tween unbound and total flucloxacillin concentrations is depicted
the external datasets. All abovementioned tests were performed using
in Figure 1; it shows non-linear, saturable protein binding that is
GraphPad Prism version 8.0.2 for Windows, GraphPad Software, San Diego,
CA, USA, www.graphpad.com.
concentration dependent. This was confirmed by the association
between the protein binding of flucloxacillin and total flucloxacillin
concentrations, which was not constant (Spearman correlation
Ethics r = #0.63, P = <0.0001). In line with this finding, the protein binding
Approval for the prospective study, from which 10 patients for the method of flucloxacillin was positively associated with serum albumin con-
development dataset originated, was obtained by the ethics committee of
centrations (Figure 2, Spearman correlation r = 0.52, P = <0.0001).
the Royal Brisbane and Women’s hospital (protocol HREC/09/QRBW/85)
and written informed consent was obtained for each patient prior to
entering the study. For the other 25 patients included in the model- Unbound flucloxacillin concentration (mg/L)
40
building dataset and the 20 patients in the Brisbane external dataset, a
waiver for informed consent was granted by the ethics committee of
the Royal Brisbane and Women’s Hospital, since blood sampling
was performed as part of the local TDM programme. For the Nijmegen 30
external dataset, the study protocol was evaluated by the local ethics
committee and the need for written informed consent was waived due
to its observational nature. 20

Results
10
Patients and samples
Patient characteristics are shown in Table 1. The majority of the
patients were treated with flucloxacillin for bloodstream or re- 0
spiratory infections, at doses ranging from 1 g q6h to 2 g q2h. 0 50 100 150 200 250
For all patients, flucloxacillin dose was at the discretion of the
Total flucloxacillin concentration (mg/L)
treating intensivist. In total, 79 total and 104 unbound flucloxa-
cillin plasma concentrations were collected, obtained at a me- Figure 1. Unbound versus total flucloxacillin concentrations, measured
dian of 30 h after the start of flucloxacillin treatment (range: in 79 samples from 16 patients.

Table 1. Patient characteristics

External patient cohorts

Characteristic Model-building patient cohort (n = 35) Brisbane (n = 20) Nijmegen (n = 14)

Female, n (%) 12 (34) 7 (35) 3 (21)

Age (years) 52 (43–67) 55 (41–62) 61 (51–71)
Total body weight (kg) 95 (73–120) 80 (62–115) 83 (74–96)
BMI (kg/m2) 31 (25–35) 26 (20–34) 27 (24–29)
SOFA score 8 (5–13) 6 (3–10) 9 (5–11)
eGFR (mL/min) 96 (26–166) 52 (11–159) 51 (22–177)
RRT, n (%) 4 (11) 3 (15) 3 (21)
Albumin (g/L) 21 (15–34) 21 (15–33) 15 (10–26)

Values are expressed as median (IQR), unless stated otherwise. RRT, renal replacement therapy.

2643
Jager et al.

Analysed serum albumin concentrations ranged from <15 mg/L to Ctotal ¼ Cunbound þ ðBmax  Cunbound Þ=ðKd þ Cunbound Þ (3)
34 g/L, with a median of 21 g/L.
In Equation 3, Ctotal is the total flucloxacillin concentration, Cunbound
is the unbound flucloxacillin concentration and (Bmax % Cunbound)/
(Kd ! Cunbound) represents the bound flucloxacillin concentration,
PK analysis where Bmax is the maximum binding capacity and Kd is the equilib-
A two-compartmental model with first-order elimination provided rium dissociation constant.
the best fit for logarithmically transformed data. In line with the The multivariate covariate analysis revealed that there was a

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

observations in Figure 1, non-linear protein binding resulted in a statistically significant association between serum albumin con-
better fit than linear protein binding and was described by the fol- centration and Bmax and between eGFR and CL, as described by
lowing equation:14 Equations 4 and 5.

Bmax ðmmol=LÞ ¼ 0:469  ðAlbumin=20Þ1:51 (4)

100
CL ðL=hÞ ¼ 55:4  ðeGFR=90Þ0:809 (5)

90 The associations between the covariates and the PK parameters

Protein binding (%)

are depicted in Figure 3. With these associations in the model,

the between-patient variability (BPV) of Bmax decreased from
80 42.2% to 30.4% and the BPV of CL decreased from 88.1% to
71.6%.
Parameter estimates of the different model-building steps are
70
shown in Table 2. Detailed information on the results of methodo-
logical model building and validation, including the VPC (Figure S1),
60 is available in the Supplementary data (Appendix 2).

15 20 25 30 35
External validation
Table 1 shows the characteristics of the patients in the different
Albumin concentration (g/L)
datasets. Table 3 shows that, for both external datasets, the pre-
Figure 2. Protein binding of flucloxacillin (%), calculated as dictive performance of the final model was statistically significant-
(1 # unbound fraction) % 100, versus serum albumin concentrations, ly better than the predictive performance of the structural model,
based on 79 patient samples. with lower median percentage errors and lower median absolute

(a) (b)
250
1.2

1.0 200
Flucloxacillin clearance (L/h)
Bmax (mmol/L)

0.8
150

0.6
100
0.4

50
0.2

y = 0.469* (x/20)1.51 y = 55.4* (x/90)0.809

0 0
0 10 20 30 0 50 100 150 200

Albumin (g/L) eGFR (mL/min)

Figure 3. Covariate relationship between (a) serum albumin concentrations and Bmax and (b) eGFR and CL of unbound flucloxacillin for the final
model, for all patients for whom both total and unbound concentrations were measured (n = 16). The dots represent the individual estimates of (a)
Bmax and (b) unbound flucloxacillin CL. The line represents the model-predicted association between the parameter estimate and the covariate of
interest.

2644
Flucloxacillin dose optimization JAC
Table 2. Parameter estimates of the different model-building steps

Structural model Final model Bootstrap (n = 1000) of model with covariates

Parameters estimate RSE (%) estimate RSE (%) estimate 95% CI

Bmax (mmol/L) 0.46 14.5 0.469 14.1 0.478 0.316–0.622

Kd (mmol/L) 0.0397 15.3 0.0441 16.6 0.0450 0.0260–0.0621
CL (L/h) 54.6 13.6 55.4 11.4 55.2 42.8–68.1

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

V1 (L) 51.5 11.5 52.7 12.0 53.3 36.9–68.6
V2 (L) 55.9 11.9 56.8 11.8 57.3 41.6–72.0
Q (L/h) 66.4 25.8 67.2 26.0 65.6 32.6–102
BPV
Bmax (%CV) 42.2 23.3 30.4 19.2 28.5 14.4–41.1
CL (%CV) 88.1 11.7 71.6 15.8 68.9 43.7–96.1
Residual variability
proportional error, unbound flucloxacillin 0.222 11.6 0.222 11.2 0.212 0.169–0.275
proportional error, total flucloxacillin 0.161 11.5 0.160 11.6 0.150 0.112–0.203
Covariates
albumin — — 1.51 28.6 1.52 0.521–2.50
eGFR — — 0.809 24.2 0.809 0.365–1.02

CV, coefficient of variation; RSE, relative standard error.

Table 3. Predictive performance of the structural and final model in two external datasets

Brisbane external dataset (n = 20) Nijmegen external dataset (n = 14)

Characteristics structural model final model P structural model final model P

Total flucloxacillin
error (%) NA NA 18.1 (#54.8 to 66.4) #11.0 (#57.1 to 28.3) 0.0005
absolute error (%) NA NA 55.3 (28.5–77.6) 39.6 (19.2–59.6) 0.004
Unbound flucloxacillin
error (%) #59.4 (#83.4 to 14.3) #5.80 (#36.9 to 29.7) 0.01 #49.1 (#86.1 to 9.80) #27.8 (#64.5 to 21.2) 0.04
absolute error (%) 70.3 (40.4–92.0) 33.4 (10.9–58.4) 0.0005 59.2 (32.3–88.9) 51.7 (25.2–74.9) 0.01

Values are expressed as median (IQR). A one-sided Wilcoxon matched-pairs test was used to test differences between the performance of the struc-
tural model and the final model. NA, not available.

percentage errors. For both external datasets, the structural as 0.5 mg/L, a dosing regimen of 1 g q6h resulted in a PTA of 91% for
well as the final model proved to be unbiased for predicting the un- patients with an eGFR of 33 mL/min. A dosing regimen of 2 g q4h
bound and total flucloxacillin concentrations. The VPCs (Figures S2 resulted in a PTA of 87% for patients with an eGFR of 96 mL/min
and S3) show that the final model, with covariates, better pre- and a PTA of 71% for patients with an eGFR of 153 mL/min.
dicted the observed concentration–time data of both external The PTA for an MIC of 0.25 mg/L, the MIC of the majority of
datasets than the model without covariates. targeted pathogens,13 was >90% for patients with an eGFR of
96 mL/min treated with a dosing regimen of 2 g q4h. For patients
with a higher eGFR, a PTA of >90% was only reached for pathogens
Monte Carlo dosing simulations
with an MIC of 0.125 mg/L and a dosing regimen of 2 g q4h. The
Patients with higher eGFRs and lower serum albumin concentra- PTA for an MIC of 2.0 mg/L, the ECOFF of oxacillin for MSSA, and a
tions had lower total flucloxacillin concentrations (Figure 4a and c). dosing regimen of 2 g q4h was 95% for patients with an eGFR of
For unbound flucloxacillin, simulations indicated that higher eGFRs 33 mL/min, 57% for patients with an eGFR of 96 mL/min and 36%
resulted in lower unbound concentrations. Serum albumin concen- for patients with an eGFR of 153 mL/min.
trations did not affect unbound concentrations (Figure 4b and d).
Figure 5 shows the difference in PTA for the frequently
prescribed dosing regimens 1 g q6h, 1 g q4h, 2 g q6h and 2 g q4h,
Discussion
for different simulated eGFRs and different MICs of the targeted This is, to the best of our knowledge, the first study where both
pathogen. When targeting the ECOFF of cloxacillin for S. aureus, the PK and the covariates affecting the PK of total and unbound

2645
Jager et al.

(a) eGFR 33, Alb 15 (b)

250 eGFR 33, Alb 15
eGFR 33, Alb 30 20
eGFR 33, Alb 30
concentration (mg/L)

Unbound flucloxacillin
200 eGFR 153, Alb 15

concentration (mg/L)
Total flucloxacillin

eGFR 153, Alb 15

eGFR 153, Alb 30 15
eGFR 153, Alb 30
150
10
100

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

50 5

0 0
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Time after first dose of flucloxacillin (h) Time after first dose of flucloxacillin (h)

(c) eGFR 33, Alb 15 (d)

400 eGFR 33, Alb 15
eGFR 33, Alb 30 40
eGFR 33, Alb 30
eGFR 153, Alb 15
concentration (mg/L)

Unbound flucloxacillin
concentration (mg/L)
300 eGFR 153, Alb 15
Total flucloxacillin

eGFR 153, Alb 30 30

eGFR 153, Alb 30
200
20

100 10

0 0
0 4 8 12 16 20 24 0 4 8 12 16 20 24
Time after first dose of flucloxacillin (h) Time after first dose of flucloxacillin (h)

Figure 4. Illustration of the effect of the covariates eGFR (mL/min) and serum albumin concentration (g/L) on the concentration–time curve of flu-
cloxacillin as assessed by Monte Carlo simulations of the first 24 h of treatment of a virtual critically ill patient, with all median characteristics of the
population, but with two different eGFR values and two different serum albumin concentrations. Both total and unbound flucloxacillin concentrations
were simulated for two different IV dosing regimens: (a) total flucloxacillin concentrations after 1 g q6h; (b) unbound flucloxacillin concentrations
after 1 g q6h; (c) total flucloxacillin concentrations after 2 g q4h; and (d) unbound flucloxacillin concentrations after 2 g q4h. Alb, serum albumin
concentration.

flucloxacillin in critically ill patients are described with the aim of but also via tubular secretion and non-renal mechanisms, where
optimizing flucloxacillin dosing regimens. In a previous study per- non-renal mechanisms account for approximately 30% of total
formed at the Royal Brisbane and Women’s Hospital,5 in which 10 CL;7,16,17 and (ii) eGFR calculated by the Chronic Kidney Disease
of the patients from this current study were included, patients with Epidemiology Collaboration (CKD-EPI) equation is not validated in
hypoalbuminaemia (serum albumin concentration 32 g/L) and critically ill patients and it is known that serum creatinine has
without severe renal dysfunction (serum creatinine <170 lmol/L) demerits in this population, considering the rapid (patho)physio-
were recruited to develop a PK model. Only unbound flucloxacillin logical changes in these patients.18 However, our results show that
concentrations were used for that PK model and thus no protein- eGFR was related to unbound flucloxacillin CL, and thus flucloxacil-
binding model was applied. As expected, our PK parameter esti- lin exposure and PTA, and this is an easily accessible parameter.
mates for central volume of distribution (V1), peripheral volume Furthermore, augmented CLCR (>130 mL/min/1.73 m2) is present
of distribution (V2) and intercompartmental clearance (Q) were in approximately 50% of patients admitted to the ICU.19
similar to those reported in the previously performed study. The Therefore, we believe this parameter is of value when optimizing
PK parameter estimate for CL was lower in the current study, dosing regimens for the individual critically ill patient.
which can be explained by the fact that patients with severe renal We demonstrated that, due to non-linear protein binding
dysfunction were excluded from the previously published study, across the observed concentration range, total flucloxacillin con-
but not from the current study. In a recently published report centrations are not representative of unbound concentrations.
describing the total and unbound PK of flucloxacillin in non- This is in line with the finding of a previous study, where significant
critically ill patients, a two-compartment model with non-linear differences between predicted (from total concentrations) and
protein binding was also found to best describe the data, with a measured unbound concentrations for flucloxacillin were
parameter estimate for Bmax in the same range.15 reported.20 Most assays used in TDM programmes and PK studies
We showed that a lower eGFR was related to a lower flucloxa- of critically ill patients measure total b-lactam antibiotic concen-
cillin CL; BPV of flucloxacillin CL decreased from 88.1% to 71.6% trations.21,22 In order to individualize dosing regimens, published
upon inclusion of eGFR. The large remaining BPV can be explained plasma protein-binding percentages derived from studies per-
by: (i) flucloxacillin is not only eliminated by glomerular filtration, formed in non-critically ill patients are used to calculate unbound

2646
Flucloxacillin dose optimization JAC
(a) (b)

Patients reaching 100% f T>MIC (%)

Patients reaching 100% f T>MIC (%)
100
100

80 80

60 60

40 40
eGFR 33 mL/min

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

eGFR 33 mL/min
20 eGFR 96 mL/min 20
eGFR 96 mL/min
eGFR 153 mL/min eGFR 153 mL/min
0 0
0.0625 0.125 0.25 0.5 1 2 4 0.0625 0.125 0.25 0.5 1 2 4

MIC (mg/L) MIC (mg/L)

(c) (d)
Patients reaching 100% f T>MIC (%)

100

Patients reaching 100% f T>MIC (%)

100

80 80

60 60

40 40

eGFR 33 mL/min eGFR 33 mL/min

20 20 eGFR 96 mL/min
eGFR 96 mL/min
eGFR 153 mL/min eGFR 153 mL/min
0 0
0.0625 0.125 0.25 0.5 1 2 4 0.0625 0.125 0.25 0.5 1 2 4

MIC (mg/L) MIC (mg/L)

Figure 5. Monte Carlo simulations (n = 1000) and PTA for achieving 100%fT>MIC at t = 24 h for various eGFRs calculated by the CKD-EPI equation, for
four different IV flucloxacillin dosing regimens administered to critically ill patients: (a) 1 g q6h; (b) 1 g q4h; (c) 2 g q6h; and (d) 2 g q4h. The ECOFF of
cloxacillin (as a surrogate for flucloxacillin) for S. aureus is 0.5 mg/L, according to EUCAST.14

concentrations from the measured total concentrations.20,23 Our different countries. Since there is no defined therapeutic window
results show that these assays may not be suitable in clinical prac- for flucloxacillin and TDM is not routinely applied, we chose to com-
tice, where the target is defined as the unbound concentration, i.e. pare the population-predicted concentrations (PRED) from the
fT>MIC. An explanation for the non-linear association between un- model with and without the covariates. There are no standardized
bound and total concentrations is that when the concentration of requirements for bias and precision when applying a PK model to
flucloxacillin in plasma increases, binding sites on proteins are in- an external dataset, therefore we used a statistically significant
creasingly saturated, resulting in a higher fraction of unbound drug endpoint to evaluate the predictive performance. For both data-
in plasma, i.e. saturable protein binding.24 In our data, differences sets, we found a significant improvement in predictive perform-
in the unbound fraction of flucloxacillin could be partly explained ance when the covariates were added to the model. This confirms
by differences in serum albumin concentrations; serum albumin that the identified covariates, serum albumin concentration and
concentrations were associated with Bmax. A lower serum albumin eGFR, are of relevance for flucloxacillin exposure in critically ill
concentration was related to lower protein binding (and a higher patients.
unbound fraction) and resulted in a lower exposure of total fluclox- Our results show that large differences in PTA are encountered
acillin, but not unbound flucloxacillin. This finding is similar to what for different eGFRs and target MICs, indicating that it is of import-
was observed in non-critically ill patients.15 Consequently, patients ance to combine information on the targeted pathogen (MIC) with
with low serum albumin concentrations are most likely to have information on the patient (e.g. renal function) to devise a rational
lower flucloxacillin protein binding than the values observed in the dosing regimen. In general, when aiming for an effectiveness tar-
literature. This is of special importance in the ICU, where 40%–50% get of 100%fT>MIC,8 a dosing regimen of 2 g q4h is adequate for
of the patients have low serum albumin concentrations (<25 g/ patients with an eGFR of <96 mL/min and where no MIC of the tar-
L).25 In clinical practice, this means that when only total flucloxa- geted pathogen is available. For patients with an eGFR of 33 mL/
cillin concentrations are measured and converted into unbound min, a dosing regimen of 1 g q6h should be adequate. For patients
concentrations using protein-binding values observed in the litera- with an eGFR of 96 mL/min and who are infected with moderate-
ture, i.e. 95%,7 in order to evaluate the attainment of the PK/PD ly susceptible pathogens, TDM using unbound flucloxacillin con-
target of 100%fT>MIC, there is a risk of underestimating unbound centrations is advised to devise an optimal dosing regimen. For
concentrations. As a consequence, unnecessarily high flucloxacillin patients infected with MSSA, currently used intermittent dosing
doses may be selected when dosing regimens are based on meas- regimens are unlikely to result in an acceptable PTA, particularly in
ured total concentrations. patients without renal impairment. For these patients, measure-
We validated the impact of eGFR and serum albumin concen- ment of the pathogen MIC and TDM using unbound concentrations
trations on the flucloxacillin PK in two external datasets from two is advised.

2647
Jager et al.

Our study has several limitations. First, the limited number of

patients and total and unbound flucloxacillin concentrations that Supplementary data
were used for model building has likely resulted in a limited power Figures S1 to S3 and Appendices 1 to 3 are available as Supplementary data
to identify covariates that have a significant impact. This might, for at JAC Online.
example, explain why no statistically significant relationship was
found between renal replacement therapy (RRT) (n = 4 patients)
and flucloxacillin. However, we believe that with this dataset we References
were able to identify the most clinically relevant covariates, which 1 Vincent J, Rello J, Marshall J et al. International study of the preva-

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

was also confirmed in two external patient cohorts. Also, although lence and outcomes of infection in intensive care units. JAMA 2009;
our dataset consisted of a limited number of patients, it is the 302: 2323–9.
largest ICU patient dataset used for population PK modelling of 2 Kollef MH, Sherman G, Ward S et al. Inadequate antimicrobial treatment
flucloxacillin to date. Second, our PK model is solely based on of infections: a risk factor for hospital mortality among critically ill patients.
patients receiving intermittent, and not continuous, infusion of Chest 1999; 115: 462–74.
flucloxacillin. This is due to the nature of this retrospective co- 3 MacArthur RD, Miller M, Albertson T et al. Adequacy of early empiric anti-
hort in a centre where intermittent infusion is current standard biotic treatment and survival in severe sepsis: experience from the MONARCS
practice in the ICU. Third, all patients in the model development trial. Clin Infect Dis 2004; 38: 284–8.
and both external datasets had a serum albumin concentration 4 Kumar A, Roberts D, Wood K et al. Duration of hypotension before initiation
below 35 g/L. Therefore, extrapolation of this model to patients of effective antimicrobial therapy is the critical determinant of survival in
human septic shock. Crit Care Med 2006; 34: 1589–96.
with higher serum albumin concentrations should be per-
formed with caution. 5 Ulldemolins M, Roberts JA, Wallis SC et al. Flucloxacillin dosing
in critically ill patients with hypoalbuminaemia: special emphasis
Nevertheless, our data represent an important step forward,
on unbound pharmacokinetics. J Antimicrob Chemother 2010; 65:
as this study is the first to develop and validate a PK model
1771–8.
incorporating both total and unbound flucloxacillin. This
6 Roberts JA, Abdul-Aziz MH, Lipman J et al. Individualised antibiotic dosing
resulted in several relevant findings that are easily applicable to
for patients who are critically ill: challenges and potential solutions. Lancet
the optimization of flucloxacillin dosing regimens in daily clinic- Infect Dis 2014; 14: 498–509.
al practice. 7 Sweetman SC, Flucloxacillin. In: Martindale: The Complete Drug Reference.
36th edn. Pharmaceutical Press, 2009; 277.
Conclusions 8 Jager NGL, van Hest RM, Lipman J et al. Therapeutic drug monitoring of
We showed that both eGFR and serum albumin concentration anti-infective agents in critically ill patients. Expert Rev Clin Pharmacol 2016;
have a significant impact on flucloxacillin exposure and PK/PD tar- 9: 961–79.
get attainment and should be taken into account when devising a 9 Wong G, Briscoe S, McWhinney B et al. Therapeutic drug monitoring of
b-lactam antibiotics in the critically ill: direct measurement of unbound
rational dosing regimen for critically ill patients. For patients with
drug concentrations to achieve appropriate drug exposures. J Antimicrob
high CLCR and infected with moderately susceptible pathogens,
Chemother 2018; 73: 3087–94.
TDM is advised, as a risk of underexposure exists. Dose adjust-
10 McWhinney BC, Wallis SC, Hillister T et al. Analysis of 12 b-lactam antibiot-
ments should be based on unbound concentrations, due to the
ics in human plasma by HPLC with ultraviolet detection. J Chromatogr B
non-linear protein binding of flucloxacillin and the high prevalence Analyt Technol Biomed Life Sci 2010; 878: 2039–43.
of hypoalbuminaemia in this patient population.
11 Sheiner LB, Beal SL. Some suggestions for measuring predictive perform-
ance. J Pharmacokinet Biopharm 1981; 9: 503–12.
12 Sutherland R, Croydon EA et al. Flucloxacillin, a new isoxazolyl penicillin,
Funding compared with oxacillin, cloxacillin, and dicloxacillin. Br Med J 1970; 4:
This study was supported by internal funding. 455–60.
13 EUCAST. Antimicrobial wild type distributions of microorganisms, cloxacil-
lin/Staphylococcus aureus. 2019. https://mic.eucast.org/Eucast2/Search
Transparency declarations Controller/search.jsp?action=performSearch&BeginIndex=0&Micdif=mic&
R.M.vH. has provided consultancies for Nordic Pharma. R.J.M.B. has served NumberIndex=50&Antib=-1&Specium=14
as a consultants to Astellas Pharma, Inc., F2G, Amplyx, Gilead Sciences,
14 Toutain PL, Bousquet-Melou A. Free drug fraction vs. free drug concentra-
Merck Sharp & Dohme Corp., and Pfizer, Inc., and has received unrestricted
tion: a matter of frequent confusion. J Vet Pharmacol Ther 2002; 25: 460–3.
and research grants from Astellas Pharma, Inc., Gilead Sciences, Merck
Sharp & Dohme Corp., and Pfizer, Inc. All contracts were through 15 Wilkes S, van Berlo I, Ten Oever J et al. Population pharmacokinetic model-
Radboudumc, and all payments were invoiced by Radboudumc. J.L. has ling of total and unbound flucloxacillin in non-critically ill patients to devise a
received honoraria from Pfizer and MSD. J.A.R. wishes to acknowledge fund- rational continuous dosing regimen. Int J Antimicrob Agents 2019; 53: 310–7.
ing from the Australian National Health and Medical Research Council for 16 Landersdorfer CB, Kirkpatrick CMJ, Kinzig-Schippers M et al. Population
Centre of Research Excellence (APP1099452) and a Practitioner Fellowship pharmacokinetics at two dose levels and pharmacodynamic profiling of flu-
(APP1117065). J.A.R. has provided lectures/consultancies or had grant cloxacillin. Antimicrob Agents Chemother2007; 51: 3290–7.
funding provided to his institution from MSD, Astellas, BioMérieux, 17 Thijssen HH, Wolters J. The metabolic disposition of flucloxacillin in
Accelerate Diagnostics and Cardeas Pharma. All other authors: none to patients with impaired kidney function. Eur J Clin Pharmacol 1982; 22:
declare. 429–34.

2648
Flucloxacillin dose optimization JAC
18 Sunder S, Jayaraman R, Mahapatra HS et al. Estimation of renal function 22 El-Najjar N, Hösl J, Holzmann T et al. UPLC–MS/MS method for therapeutic
in the intensive care unit: the covert concepts brought to light. J Intensive drug monitoring of 10 antibiotics used in intensive care units. Drug Test Anal
Care 2014; 2: 31. 2018; 10: 584–91.
19 Claus BOM, Hoste EA, Colpaert K et al. Augmented renal clearance is a 23 Carlier M, Stove V, Wallis SC et al. Assays for therapeutic drug monitoring
common finding with worse clinical outcome in critically ill patients receiving of b-lactam antibiotics: a structured review. Int J Antimicrob Agents 2015; 46:
antimicrobial therapy. J Crit Care 2013; 28: 695–700. 367–75.
20 Wong G, Briscoe S, Adnan S et al. Protein binding of b-lactam antibiotics 24 Zeitlinger MA, Derendorf H, Mouton JW et al. Protein binding: do we ever
in critically ill patients: can we successfully predict unbound concentrations? learn? Antimicrob Agents Chemother 2011; 55: 3067–74.

Downloaded from https://academic.oup.com/jac/article/75/9/2641/5842232 by Universite de Montreal user on 31 October 2023

Antimicrob Agents Chemother 2013; 57: 6165–70. 25 Finfer S, Bellomo R, McEvoy S et al. Effect of baseline serum albumin
21 Pinder N, Brenner T, Swoboda S et al. Therapeutic drug monitoring of concentration on outcome of resuscitation with albumin or saline in patients
b-lactam antibiotics – influence of sample stability on the analysis of pipera- in intensive care units: analysis of data from the saline versus albumin fluid
cillin, meropenem, ceftazidime and flucloxacillin by HPLC-UV. J Pharm evaluation (SAFE) study. BMJ 2006; 333: 1044.
Biomed Anal 2017; 143: 86–93.

2649