Clinical Therapeutics/Volume 35, Number 10, 2013

Ceftazidime Dosage Recommendations in Burn Patients:

From a Population Pharmacokinetic Approach to Clinical
Practice via Monte Carlo Simulations
Jean-Marie Conil, MD, PhD1; Bernard Georges, MD, PhD1; François Ravat, MD2;
Stéphanie Ruiz, MD1; Thierry Seguin, MD1; David Metsu, PharmD3;
Olivier Fourcade, MD, PhD1; and Sylvie Saivin, PharmD, MD, PhD3
1
Anesthésie Réanimation, CHU Rangueil, Toulouse, France; 2Centre des brûlés, Centre hospitalier, Lyon,
France; and 3Laboratoire de Pharmacocinétique et Toxicologie Clinique, Institut Fédératif de Biologie,
Toulouse, France

ABSTRACT the need of a reappraisal of ceftazidime’s use in burn

Background: Ceftazidime dosage regimen recom- patients. Doses between 3 and 16 g/d are proposed,
mendations based on pharmacokinetic/pharmacody- taking into account the pathogens’ MICs. However,
namic approaches are not available for burn patients. for sepsis caused by a pathogen with an MIC Z8 mg/L,
Objective: The goal of this study was to propose a an insufficient percentage of burn patients will reach
continuous dosage regimen of ceftazidime in burn the therapeutic target with the recommended dosages.
patients, taking into account different MICs and (Clin Ther. 2013;35:1603–1612) & 2013 Elsevier HS
pharmacokinetic covariates. Journals, Inc. All rights reserved.
Methods: The population pharmacokinetic analysis Key words: burn, ceftazidime, dosage regimen,
was conducted by using software dedicated to the Monte Carlo simulations, pharmacokinetics.
analysis of nonlinear mixed effects models. The
population pharmacokinetic model was first devel-
oped and validated in 70 adult burn patients. Taking
INTRODUCTION
into account various MICs of pathogens, 3 Monte
Severe Pseudomonas aeruginosa infection can jeopard-
Carlo simulation trials were conducted by using target
ize the prognosis for burn patients and necessitates rapid
concentration intervals (10–100, 20–100, and 40–
antibiotic therapy. Among the antibiotics available,
100 mg/L). The recommended dosages were defined
ceftazidime may be an appealing choice because of its
as the minimum dose leading to the highest percentage
lower risk of resistance compared with most β-lactams.1,2
of patients whose ceftazidime concentrations were
Ceftazidime pharmacokinetics are defined by high
included in the target interval.
renal excretion and a volume of distribution similar to
Results: Serum creatinine and age were identified
that of the extracellular space.3 It is mainly eliminated
as covariates of ceftazidime clearance. Age was also
by glomerular filtration, and 88% of the dose is
involved in volume of distribution. The simulations
recovered in the urine over 24 hours. Consequently,
showed that a dose of 6 g/d did not allow achievement
changes in renal function in the hypermetabolic phase
of the target interval in most patients. Regardless of
after a burn injury affect the pharmacokinetics of
dosage regimen, age, and serum creatinine, the mean
ceftazidime,4,5 and a 2-g dose every 8 hours may be
percentage of patients reaching the 10- to 100-mg/L
inadequate in burn or intensive care unit (ICU) patients.6
and the 20- to 100-mg/L target intervals were 99.4%
For β-lactams, bacterial killing is primarily related
(0.3%) and 96.1% (0.8%), respectively. For the 40- to
to time that concentrations in the tissues and plasma
100-mg/L target interval, this percentage was only
76.4% (2.1%) (range, 65%–80%).
Accepted for publication August 23, 2013.
Conclusions: Age and serum creatinine level can be http://dx.doi.org/10.1016/j.clinthera.2013.08.014
used at the bedside to determine the initial doses of 0149-2918/$ - see front matter
ceftazidime. These Monte Carlo simulations highlight & 2013 Elsevier HS Journals, Inc. All rights reserved.

October 2013 1603

 Clinical Therapeutics
exceed a certain threshold.7 Dosing regimens are
therefore being re-evaluated to keep plasma concen-
trations above certain thresholds over the whole
treatment period, regardless of the administrat-
ion schedule.8 Because ceftazidime exhibits time-
dependent killing of gram-negative bacteria in vitro
or in critically ill patients, studies involving continu-
ous administration of cephalosporin confirm that the
steady-state blood concentration should be in excess
of the bacterial MIC.7,9 Continuous infusion opti-
mizes the pharmacodynamics of β-lactams by provid-
ing adequate antibacterial activity over the 24-hour
dosing period with a reduction in the total daily dose
of the antimicrobial agent.10–12
Some studies have reported an influence of cova-
riates such as age, creatinine, or creatinine clearance
on ceftazidime disposition in burn patients.4,6,13 Two
studies used a population approach5,14 without any
validation process. Therefore, a recommended cefta-
zidime dosage regimen based on a pharmacokinetic/
pharmacodynamic approach is not actually deter-
mined in such singular burn patients. The aim of the
present study was to propose an a priori–required
dosage of ceftazidime given by continuous infusion to
this population. Its estimation is based on clinical and
biological parameters used in our practice at the
bedside of burn patients.
MATERIALS AND METHODS
Two prospective, open-label, randomized population
studies of ceftazidime’s pharmacokinetics were con-
ducted in burn patients, adhering to the agreement of
the Toulouse and Lyon Ethical Committees. The
clinical parts of the studies were performed in the
burn patient units of the University Hospital in
Toulouse-Rangueil (n = 50) and of the Hospital “St
Joseph et St Luc” in Lyon (n = 20). The Toulouse
population’s data have been partially used in previous
studies.4–6 Written informed consent was obtained
from the patients or their relatives.
All the ceftazidime plasma concentration data were
obtained and analyzed in the Laboratoire de Pharma-
cocinétique et Toxicologie Clinique of the Purpan
Hospital in Toulouse, France.
Patients
Patients were studied during the secondary phase of
their burn injuries. The following demographic, clin-
ical, and biological parameters were collected as
1604
possible covariates: age, sex, weight, height, mechan-
ical ventilation, serum creatinine, proteins, and blood
urea nitrogen. Specific burn indices such as Baux,
UBS, Tobiasen, and burn area were also recorded. The
glomerular filtration rate was estimated from creati-
nine clearance, calculated by using the Cockcroft-
Gault method21 and the Modification of Diet in
Renal Disease (MDRD).22
Drug Administration
Antibiotics were prescribed for local infections or
for sepsis either empirically when a patient’s condition
required immediate treatment or after a bacteriologic
evaluation. Before inclusion, patients were random-
ized to treatment according to the mode of admin-
istration: continuous infusion of ceftazidime with a
loading dose or discontinuous administration. Seventy
patients entered into these studies between 2003 and
2009, and they received ceftazidime by 3 modes of
administration. One group received a ceftazidime dose
of 6 g/d in 3 separate 20-minute infusions of 2 g each
every 8 hours (n = 25). The second group was given
6 doses of 1 g each every 4 hours (n = 25). A trough
concentration was measured at the end of day 1; the
frequency and/or the dose of ceftazidime was then
adjusted to obtain a target trough concentration if
necessary. The new dosing schedule could be increas-
ed to 1 g 8 times a day (ie, every 3 hours), or reduced to
1 g 4 times a day (ie, every 6 hours) or to 1 g 3 times a
day (ie, every 8 hours). The third group received 8 g on
day 1, a 2-g loading dose over 30 minutes followed by
6 g/d continuously administered (n = 20). This dosage
was initially adapted in patients with an impaired renal
function according to the usual recommendations
(summary of product characteristics).
Blood Sampling and Measurements
In the fractionated administration groups, blood
was sampled in dry tubes at 24, 24.33, 48, and 48.33
hours after the beginning of the treatment. In the
continuous infusion group, patients were randomly
distributed in 3 subgroups, and blood was sampled at
0, 0.25, 1.5, 6, and 24 hours or 0.5, 2, 12, and 24
hours or 1, 4, 18, and 24 hours. When the dosing
regimen was changed for monitoring reasons, further
samples were taken to determine the new ceftazidime
concentrations. A total of 286 measurable serum
concentration were available.
Volume 35 Number 10
 J.-M. Conil et al.

Ceftazidime Assay Secondarily, the influence of each covariate pre-

Total serum concentrations of ceftazidime were
measured by using a validated reversed-phase HPLC
method as previously described.5
Estimation of Population Parameters
The population pharmacokinetic analysis was con-
ducted on a full population analysis by using Monolix,
a free software dedicated to the analysis of nonlinear
mixed effects models. A Markov chain Monte Carlo
process was then used for the simulation step. A
1- versus 2-compartment model was evaluated to
describe the pharmacokinetics. Different error models
were tested to describe the interindividual variability.
The choice of the model was based on the lower value
of the objective function and the higher maximum of
likelihood. The pharmaco-statistical model was fitted
to the data to obtain the population parameters (mean
and variance of each parameter), in terms of total
body clearance, distributional clearance in liters per
hour, and the central and peripheral volumes of
distribution in liters.
sented in Table I was examined in the structural
model, with a 0.05 level of significance of the
objective function. The resulting pharmaco-statistical
model was refined by independently deleting each
covariate with 0.001 as the level of significance of
the objective function. To evaluate the predictive
performance of this model, we compared the meas-
ured concentrations (Cobs) with the predicted concen-
trations (Cpred). In a second step, the predictive
performance of our model was evaluated by the
generation of 1000 Monte Carlo–simulated concen-
tration sets from the qualification population. Nor-
malized distribution prediction error (npde) was then
evaluated.
Simulations
The final model was used to perform 1000 Monte
Carlo simulations in typical burn patients with cova-
riates ranging in intervals of the studied population.
These typical patients were defined by the combina-
tion of the two selected covariates of the model: (1)

Table I. Characteristics of the burn patients (n ¼ 70).

Variable Mean (SD) Range Code

Quantitative
Age, y 49 (19) 15–90 AGE
Weight, kg 74 (16) 43–120 WT
Height, cm 172 (8) 155–185 HEIG
Burned surface area (% of the total body surface) 32 (22) 10–98 BSA
Baux index 82 (24) 30–145 BAUX
UBS index 94 (76) 10–320 UBS
Tobiasen index 8 (3) 4–15 TOBI
Serum creatinine, mmol/L 77 (30) 24–234 CREATININE
Cockcroft creatinine clearance, mL/min/1.73 m2 118 (60) 33–394 CLCR
MDRD, mL/min/1.73 m2 114 (49) 29–295 MDRD
Proteinemia, g/L 57 (8) 38–75 PROT
Blood urea nitrogen, mmol/L 8 (4) 2–25 BUN
Qualitative variables
Sex 50, 20 SEX
Men ¼ 0; women ¼ 1
Mechanical ventilation 39, 31 VENT
Without ¼ 0; with ¼ 1

MDRD ¼ Modification of the Diet in Renal Disease.

October 2013 1605

 Clinical Therapeutics
serum creatinine from 30 to 160 mmol/L, with a step
of 10 mmol/L (12 cases); and (2) age from 20 to 90
years, with a step of 10 years (8 cases). Ninety-six
(12  8) typical patients were studied, and 96,000
simulations were thus performed for each tested
dosage.
Ceftazidime concentrations were simulated over 72
hours for dosage regimens, including a 2-g loading
dose followed by doses ranging from 3 to 16 g/d as a
continuous infusion with a 2-g step between 4 and
16 g/d.
Interpretation
The concentrations generated by the previous
Monte Carlo process were used to explore the appro-
priateness of different dosage regimens in burn pa-
tients by comparing these simulated concentrations
with a target interval. The concentrations have been
arbitrarily simulated at 72 hours, ensuring they are at
steady state regardless of the patients’ ceftazidime t½
values.
The lower limit of the target interval was defined as
a steady-state concentration 5-fold higher than the
MIC of the strain.7,15,16 The most frequently observed
MIC for Pseudomonas aeruginosa was 1 to 2 mg/L;
therefore, the lower limit of 10 mg/L was studied. For
strains with an MIC of 4 mg/L, the lower limit of
20 mg/L was also tested. The MIC 3-fold in empirical
situations is defined as 8 mg/L by the European
Committee on Antimicrobial Susceptibility Testing;
it leads to a lower limit of the target interval equal to
40 mg/L.7,9,17
Conversely, the upper limit of the target interval
was considered to be 100 mg/L to prevent adverse
events, regardless of the given dose.
For each typical patient of the simulated popula-
tion, the target dosage regimen was defined as the
minimal dose leading to the highest percentile of
patients whose ceftazidime concentrations were in-
cluded between the chosen target and 100 mg/L. A
dose was considered as recommendable when 490%
of the population was in the target interval.
RESULTS
Patients
Seventy burn patients, with a total of 286 meas-
urable serum ceftazidime concentrations, were in-
cluded in the study. Table I summarizes patient
characteristics. Mean age of the patients w
1606
49 (19) years; they were badly injured (UBS, 94 [76];
Tobiasen, 8 [3]), with mechanical ventilation required
in 44% of cases. Mean serum creatinine was in the
normal range (77 [30] mmol/L) as the mean Cockcroft
creatinine clearance (118 [60] mL/min) and as the
mean MDRD clearance (114 [49] mL/min). These
parameters showed considerable interindividual
variability (%CV, 39%, 50%, and 43%, respectively).
Model Validation
The open 2-compartment pharmacokinetic model
with first-order elimination was chosen to describe
the concentration–time data for ceftazidime in serum.
A proportional error model was the most accurate
for the residual variability. For the clearance, age,
urea, serum creatinine, and glomerular filtration rate
estimated by using the Cockcroft and MDRD equa-
tions, ventilation, Baux and Tobiasen indices, statisti-
cally decreased the objective function and the inter-
individual variability. For the central and peripheral
volumes of distribution, the statistically significant
covariates were, respectively, age, urea, creatinine
clearance, ventilation status, and Tobiasen index.
After independent deletion, the model was:
TVCL ¼ 21:2 – ð0:00683  creatinineÞ
– ð0:0115  ageÞ
TVV1 ¼ 13:4 þ ð0:00669  ageÞ
TVQ ¼ 1:41
TVV2 ¼ 28:9
where TVCL, TVV1, TVQ, and TVV2 are typical
values, respectively, for ceftazidime clearance (liters
per hour), central volume of distribution (liters),
intercompartmental clearance (liters per hour), and
peripheral volume of distribution (liters).
The linear regression between the individual
observed concentrations (Cobs) and the individual
predicted concentrations (IPred) determined by the
conditional mode was as follows:
CIpred ðmg=LÞ ¼ 0:8454ðCobs Þ þ 5:9318ðr ¼ 0:9335Þ:
The objective function was reduced from 2274 to
2215 (P o 0.0001; ddf ¼ 3). The ceftazidime clea-
rance interindividual variability was reduced from
65% to 52% and the central volume of distribution
from 44% to 33%.
The mean of the individual patients’ pharmacokinetic
parameters calculated by taking into account serum
creatinine level and age are shown in Table II. The
npde qualification (Figure 1) did not reveal any bias for
Volume 35 Number 10
 J.-M. Conil et al.

Simulation, Dosages, and Interpretation

Table II. Pharmacokinetics parameters calculated
in the burn population (N ¼ 70). Tables III through V show the most appropriate
dosage regimen for reaching the ceftazidime target
Minimum– concentration intervals, in function of age and serum
Parameter Mean (SD) Maximum %CV creatinine. These tables present the results for different
concentration target intervals required by different MICs.
Cl, L/h 7.6 (2.7) 2.8–14.8 35 The results in Tables III and IV correspond to the
V1, L 18.7 (2.6) 14.4–25.1 14 recommended dosage regimen, which lead to 490%
Q, L/h 2.3 (2.2) 0.6–12.4 95 of the population with a steady-state concentration in the
V2, L 38.6 (23.6) 7.4–117.1 61 defined interval. Table V shows recommended dosages
V1 þ V2, L 57.3 (23.7) 25.2–137.1 41 which can only lead to >70% of the population with a
t½, h 5.7 (2.4) 2.3–13.4 43 steady-state concentration in the defined interval.
As an example, Figure 2 presents the percentage of
Cl ¼ total body clearance; VI ¼ central volume of 50-year-old patients reaching the different steady-state
distribution; Q ¼ distributional clearance; V2 ¼ concentration intervals after various dosage regimens of
peripheral volume of distribution.
ceftazidime as a function of serum creatinine. In such a
typical patient, the percentage of the population reach-
ing a target interval increases with the lowest doses
the prediction, and the normality assumption was not when steady-state concentrations exceed the lower limit
rejected. The expected values were mean ¼ 0, variance ¼ of the interval. This percentage then decreases due to the
1, skewness ¼ 0, and kurtosis ¼ 0. The results of our number of patients with concentrations 4100 mg/L.
qualification process were mean ¼ –0.029, variance ¼ Our results reveal that a recommended dosage
0.938, skewness ¼ 0.037, and kurtosis ¼ 0.095. regimen is easy to determine for the 10- and 20-mg/

PWRES IWRES NPDE

0.7 0.7 0.5
0.6 0.6
0.4
0.5 0.5
0.4 0.4 0.3

0.3 0.3 0.2

0.2 0.2
0.1
0.1 0.1
0 0 0
–5 0 5 –5 0 5 –5 0 5

3 3 3

2 2 2

1 1 1
Y Quantiles

Y Quantiles

Y Quantiles

0 0 0

–1 –1 –1

–2 –2 –2

–3 –3 –3
–5 0 5 –5 0 5 –5 0 5
X Quantiles X Quantiles X Quantiles

Figure 1. Distribution of the population-weighted residuals (PWRES), individual-weighted residuals (IWRES),

and normalized predicted distribution error (NPDE).

October 2013 1607

 Clinical Therapeutics

Table III. Recommended ceftazidime dosage regimen after a 2-g loading dose required to reach a steady-state
concentration between 10 and 100 mg/L in the highest percentage of the typical burn patients as a
function of serum creatinine and age. All results correspond to recommended dosage regimen,
which lead to 490% of the population with a steady-state concentration in the defined interval.

Target Interval: 10–100 mg/L

Creatinine (mmol/L) 20 Years 30 Years 40 Years 50 Years 60 Years 70 Years 80 Years 90 Years

30 10 g 10 g 10 g 8 g 6 g 6 g 6 g 6 g
40 10 g 10 g 8 g 8 g 6 g 6 g 4 g 4 g
50 10 g 10 g 8 g 8 g 6 g 6 g 4 g 4 g
60 10 g 8 g 8 g 6 g 6 g 6 g 4 g 4 g
70 8 g 6 g 6 g 6 g 6 g 4 g 4 g 3 g
80 8 g 6 g 6 g 6 g 4 g 4 g 4 g 3 g
90 8 g 6 g 6 g 6 g 4 g 4 g 4 g 3 g
100 6 g 6 g 4 g 4 g 4 g 4 g 4 g 3 g
120 6 g 6 g 4 g 4 g 4 g 3 g 3 g 3 g
140 6 g 4 g 4 g 4 g 4 g 3 g 3 g 3 g
160 4 g 4 g 4 g 3 g 3 g 3 g 3 g 3 g

L lower limit, with 99.4% (0.3%) and 96.1% (0.8%) DISCUSSION

of the population in the defined interval, respectively. The aim of our present research was to propose the
When the target interval is 40 to 100 mg/L, the mean determination of an a priori–required dosage regimen
percentage of patients reaching the target is 76% (2%) of ceftazidime given by continuous infusion in clinical
(range, 65%–80%). practice. This was performed by using a model taking

Table IV. Recommended ceftazidime dosage regimen after a 2-g loading dose required to reach a steady-
state concentration between 20 and 100 mg/L in the highest percentage of typical burn patients
as a function of serum creatinine and age. All results correspond to recommended dosage
regimen, which lead to 490% of the population with a steady-state concentration in the defined
interval.

Target Interval: 20–100 mg/L

Creatinine (mmol/L) 20 Years 30 Years 40 Years 50 Years 60 Years 70 Years 80 Years 90 Years

30 12 g 12 g 12 g 10 g 10 g 8 g 8 g 8 g
40 12 g 12 g 12 g 10 g 10 g 8 g 8 g 6 g
50 12 g 12 g 10 g 10 g 8 g 8 g 6 g 6 g
60 12 g 12 g 10 g 10 g 8 g 8 g 6 g 6 g
70 12 g 10 g 8 g 8 g 8 g 6 g 6 g 6 g
80 12 g 10 g 8 g 8 g 8 g 6 g 6 g 6 g
90 10 g 10 g 8 g 8 g 6 g 6 g 6 g 4 g
100 10 g 8 g 8 g 6 g 6 g 6 g 6 g 4 g
120 8 g 8 g 6 g 6 g 6 g 4 g 4 g 4 g
140 8 g 6 g 6 g 6 g 4 g 4 g 4 g 3 g
160 6 g 6 g 6 g 4 g 4 g 4 g 3 g 3 g

1608 Volume 35 Number 10

 J.-M. Conil et al.

Table V. Recommended ceftazidime dosage regimen after a 2-g loading dose required to reach a steady-state
concentration between 40 and 100 mg/L in the highest percentage of typical burn patients
as a function of serum creatinine and age. These results correspond to recommended dosage
regimen, but do not lead to 490% of the population with a steady-state concentration in the
defined interval.

Target Interval: 40–100 mg/L

Creatinine (mmol/L) 20 Years 30 Years 40 Years 50 Years 60 Years 70 Years 80 Years 90 Years

30 16 g 16 g 16 g 16 g 14 g 12 g 12 g 10 g
40 16 g 16 g 16 g 14 g 12 g 12 g 10 g 10 g
50 16 g 16 g 16 g 12 g 12 g 10 g 10 g 8 g
60 16 g 16 g 14 g 12 g 12 g 10 g 8 g 8 g
70 16 g 14 g 12 g 12 g 10 g 10 g 8 g 8 g
80 14 g 14 g 12 g 10 g 10 g 8 g 8 g 8 g
90 14 g 14 g 12 g 10 g 10 g 8 g 8 g 6 g
100 14 g 12 g 10 g 10 g 8 g 8 g 8 g 6 g
120 12 g 10 g 10 g 8 g 8 g 6 g 6 g 6 g
140 10 g 10 g 8 g 8 g 6 g 6 g 6 g 4 g
160 10 g 8 g 8 g 6 g 6 g 6 g 4 g 4 g

into account pharmacokinetic and pharmacodynamic
aspects.
From a pharmacokinetic point of view, we chose to
only perform simulation by continuous route even if
our model is based on continuous and discontinuous
administrations. These different administration schemes
allow us to better estimate pharmacokinetic parameters,
particularly the volume of distribution, which could not
be assessed with continuous infusion data alone. By
adopting the Monolix method, we investigated the
quantitative relationships between pharmacokinetic pa-
rameters and physiologic features in burn patients
treated with ceftazidime. The data collected were best
described by using a 2-compartment pharmacokinetic
model with a usual high interindividual variability fitted
by a proportional error model.4,5,14,18,19
After the independent deletion step, age and serum
creatinine were the most significant variables, and the
interindividual variability was significantly reduced as
previously demonstrated.4,14 Age and serum creatinine
fitted the results better than creatinine clearance,
regardless of the estimation formula. Their implication
is also in accordance with the renal elimination of
ceftazidime20 because they are used in all equations to
estimate creatinine clearance or glomerular filtration
rate.21,22 The present population pharmacokinetic
model was built on data from 70 burn patients and
was validated by using data splitting and the more
advanced model evaluation tools (npde).
In this population, the mean ceftazidime clearance
(7.6 [2.7] L/h) was higher than that reported in burn
patients (2.72 L/h)14 or in ICU patients (5.48 [2.20]
L/h). This observed ceftazidime clearance is in agreement
with that reported (6.88 L/h) in a previous study.5
Compared with healthy volunteers,5,13,14 our study
found an increase in the mean total volume of distribu-
tion (57.3 [23.7] L).
From a pharmacodynamic point of view, the lower
limit of the target concentration interval has been defin-
ed as a steady-state concentration 5-fold higher than the
MIC of the strain, as recently suggested.23 There is no
consensus about the number of times the MIC must be
exceeded. Some studies have based their target on 1
MIC.11 Mouton et al showed that administration by
continuous infusion achieving concentrations near the
MIC does not provide bacteriologic advantages com-
pared with bolus dosing. Other in vitro and in vivo
studies7,24 have confirmed that concentrations given by
continuous infusion should be maintained at 4- to 5-
fold the MIC to have a better cure rate. In a clinical
study using continuous infusion of ceftazidime in ICU
patients, Lipman et al17 proposed a target concentra-
tion of 40 mg/L. More recently, Aubert et al9 used the
same target concentration.

October 2013 1609

Clinical Therapeutics

Ceftazidime 3 to 10 g/d after a 2-g loading dose in 50-year-old patients

100

the 10- to 100-mg/L Steady-State

Percentage of Patients Reaching

Concentration Target Interval

90

80 3 g/d
4 g/d
6 g/d
70 8 g/d
10 g/d

60

50
0 20 40 60 80 100 120 140 160 180
Serum Creatinine (µmol/L)

Ceftazidime 3 to 12 g/d after a 2-g loading dose in 50-year-old patients

100
the 20- to 100-mg/L Steady-State
Percentage of Patients Reaching

Concentration Target Interval

90

80 3 g/d
4 g/d
6 g/d
8 g/d
70 10 g/d
12 g/d

60

50
0 20 40 60 80 100 120 140 160 180
Serum Creatinine (µmol/L)

Ceftazidime 6 to 16 g/d after a 2-g loading dose in 50-year-old patients

100
the 40- to 100-mg/L Steady-State
Percentage of Patients Reaching

Concentration Target Interval

90

80 6 g/d
8 g/d
10 g/d
12 g/d
70 14 g/d
16 g/d

60

50
0 20 40 60 80 100 120 140 160 180
Serum Creatinine (µmol/L)

Figure 2. Percentage of 50-year-old patients reaching the (A) 10- to 100-mg/L, (B) 20- to 100-mg/L, and (C)
40- to 100-mg/L steady-state concentration target intervals after a 2-g loading dose followed by a
4- to 16-g/d continuous infusion of ceftazidime in function of serum creatinine.

1610 Volume 35 Number 10


Even if the relationship between ceftazidime con-
centration and toxicity is poorly described,25 the
maximum target concentration has been fixed to
100 mg/L by assimilation to the described toxicity of
cefepime.26 This molecule is chemically and pharmaco-
kinetically very close to ceftazidime.4
We conducted 3 simulation trials taking into
account 3 typical MICs (2, 4, and 8 mg/L). All these
simulations found that in a burn population, a stand-
ard 6-g dose of ceftazidime given as a continuous
infusion in most cases is not sufficient to reach the
defined target concentration and consequently to
ensure efficacy of this β-lactam similar to other clinical
situations. High ceftazidime doses are required due to
the hyperfiltration glomerular rate that is frequently
observed in the youngest patients with low serum
creatinine concentrations.27 This phenomenon had
already been highlighted by Loirat et al28 with the
use of tobramycin in burn patients.
In patients aged o50 years, the minimal required
dosage given by continuous infusion is 4 g/d, and it
could be 8 to 16 g/d for patients with the lowest serum
creatinine; the goal is to reach a minimal target
concentration of 10 to 40 mg/L. In the oldest patients,
whenever serum creatinine is o160 mmol/L, no dose
r3 g/d would lead to the target steady-state concen-
tration. The use of such high doses is not in accord-
ance with the summary of the product characteristics,
but it has been reported that the initial doses of
ceftazidime are insufficient to ensure efficacy against
P aeruginosa in various clinical situations.29,30
For the 2 lowest-target MICs, it is possible to
propose a ceftazidime dose leading to 490% of the
population reaching a steady-state concentration in the
target interval. For MICs equal to 8 mg/L, a maximum
of 76% of the population will reach the target interval
with the recommended doses. In such cases, combina-
tion and/or use of other antibiotics may be discussed.
CONCLUSIONS
Ceftazidime clearance and volume of distribution are
increased in burn patients, exposing them to the risk
of inefficient plasma concentrations. Age and serum
creatinine level significantly influenced the ceftazidime
disposition in these burn patients. These covariates,
which are readily usable at the bedside of burn
patients, were taken into account to propose the
initial doses of ceftazidime. Our Monte Carlo simu-
lations found that there should be a reappraisal of the
October 2013
J.-M. Conil et al.
use of ceftazidime in burn patients. The recommended
dosage regimens are higher than in other patients, and
doses between 3 and 16 g/d are proposed, taking into
account the MICs of pathogens. However, for sepsis
caused by a pathogen with an MIC Z8 mg/L, an
insufficient percentage of burn patients will reach the
therapeutic target with the recommended dosages.
ACKNOWLEDGMENT
Drs. Conil, Ravat, and Saivin were responsible for the
conception, design, pharmacokinetic/dynamic analy-
sis, and drafting of the manuscript. Dr. Georges was
responsible for the collection of the clinical data,
interpretation of data and drafting of the manuscript.
Drs. Ruiz and Seguin were responsible for the collec-
tion of the clinical data. Dr. Metsu was responsible for
the collection of biological data. Dr. Fourcade was
responsible for the coordination of the study and
helped to draft the manuscript.
CONFLICTS OF INTEREST
The authors have indicated that they have no conflicts
of interest regarding the content of this article.
REFERENCES
1. Richard P, Le Floch R, Chamoux C, et al. Pseudomonas
aeruginosa outbreak in a burn unit: role of antimicrobials
in the emergence of multiply resistant strains. J Infect Dis.
1994;170:377–383.
2. Georges B, Conil JM, Dubouix A, et al. Risk of emergence
of Pseudomonas aeruginosa resistance to beta-lactam
antibiotics in intensive care units. Crit Care Med.
2006;34:1636–1641.
3. Rains CP, Bryson HM, Peters DH. Ceftazidime. An update
of its antibacterial activity, pharmacokinetic properties
and therapeutic efficacy. Drugs. 1995;49:577–617.
4. Conil JM, Georges B, Lavit M, et al. Pharmacokinetics of
ceftazidime and cefepime in burn patients: the impor-
tance of age and creatinine clearance. Int J Clin Pharmacol
Ther. 2007;45:529–538.
5. Conil JM, Georges B, Lavit M, et al. A population
pharmacokinetic approach to ceftazidime use in burn
patients: influence of glomerular filtration, gender and
mechanical ventilation. Br J Clin Pharmacol. 2007;64:27–35.
6. Conil JM, Georges B, Fourcade O, et al. Intermittent
administration of ceftazidime to burns patients: influence
of glomerular filtration. Int J Clin Pharmacol Ther.
2007;45:133–142.
7. Mouton JW, den Hollander JG. Killing of Pseudomonas
aeruginosa during continuous and intermittent infusion of
1611
 Clinical Therapeutics
ceftazidime in an in vitro pharma-
cokinetic model. Antimicrob Agents
Chemother. 1994;38:931–936.
8. Lipman J, Wallis SC, Rickard C.
Low plasma cefepime levels in
critically ill septic patients: phar-
macokinetic modeling indicates
improved troughs with revised dos-
ing. Antimicrob Agents Chemother.
1999;43:2559–2561.
9. Aubert G, Carricajo A, Coudrot M,
et al. Prospective determination of
serum ceftazidime concentrations
in intensive care units. Ther Drug
Monit. 2010;32:517–519.
10. Alou L, Aguilar L, Sevillano D,
et al. Is there a pharmacodynamic
need for the use of continuous
versus intermittent infusion with
ceftazidime against Pseudomonas
aeruginosa? An in vitro pharmaco-
dynamic model. J Antimicrob Che-
mother. 2005;55:209–213.
11. Nicolau DP, Nightingale CH,
Banevicius MA, et al. Serum bac-
tericidal activity of ceftazidime:
continuous infusion versus inter-
mittent injections. Antimicrob Agents
Chemother. 1996;40:61–64.
12. Mouton JW, Horrevorts AM,
Mulder PG, et al. Pharmacoki-
netics of ceftazidime in serum
and suction blister fluid during
continuous and intermittent infu-
sions in healthy volunteers. Antimi-
crob Agents Chemother. 1990;34:
2307–2311.
13. Walstad RA, Aanderud L,
Thurmann-Nielsen E. Pharmacoki-
netics and tissue concentrations of
ceftazidime in burn patients. Eur J
Clin Pharmacol. 1988;35:543–549.
14. Dailly E, Pannier M, Jolliet P, Bourin
M. Population pharmacokinetics of
ceftazidime in burn patients. Br J
Clin Pharmacol. 2003;56:629–634.
15. Fantin B, Farinotti R, Thabaut A,
Carbon C. Conditions for the emer-
gence of resistance to cefpirome
and ceftazidime in experimental
endocarditis due to Pseudomonas
aeruginosa. J Antimicrob Chemother.
1994;33:563–569.
1612
16. Cappelletty DM. Evaluation of several
dosing regimens of cefepime, with
various simulations of renal function,
against clinical isolates of Pseudomo-
nas aeruginosa in a pharmacody-
namic infection model. Antimicrob
Agents Chemother. 1999;43:129–133.
17. Lipman J, Gomersall CD, Gin T,
et al. Continuous infusion ceftazi-
dime in intensive care: a random-
ized controlled trial. J Antimicrob
Chemother. 1999;43:309–311.
18. Weinbren MJ. Pharmacokinetics of
antibiotics in burn patients. J Anti-
microb Chemother. 1999;44:319–327.
19. Weinbren MJ. Pharmacokinetics of
antibiotics in burns patients. J
Antimicrob Chemother. 2001;47:720.
20. Ljungberg B, Nilsson-Ehle I. Com-
parative pharmacokinetics of cef-
tazidime in young, healthy and
elderly, acutely ill males. Eur J Clin
Pharmacol. 1988;34:179–186.
21. Cockcroft DW, Gault MH. Prediction
of creatinine clearance from serum
creatinine. Nephron. 1976;16:31–41.
22. Levey AS, Greene T, Kusek JW, Beck
GJ. A simplified equation to predict
glomerular filtration rate from se-
rum creatinine. J Am Soc Nephrol.
2000;11:155A.
23. Sinnollareddy MG, Roberts MS, Lip-
man J, Roberts JA. Beta-lactam
pharmacokinetics and pharmacody-
namics in critically ill patients and
strategies for dose optimization: a
structured review. Clin Exp Pharmacol
Physiol. 2012;39:489–496.
Address correspondence to: Sylvie Saivin, PharmD, MD, PhD, Labora-
toire de Pharmacocinétique et Toxicologie Clinique, Institut Fédératif de
Biologie, 330 Avenue de Grande Bretagne, TSA 40031, 31059 Toulouse
Cedex 9, France. E-mail: saivin.s@chu-toulouse.fr
24. Mouton JW, van Ogtrop ML, An-
des D, Craig WA. Use of pharma-
codynamic indices to predict
efficacy of combination therapy
in vivo. Antimicrob Agents Chemo-
ther. 1999;43:2473–2478.
25. Jackson GD, Berkovic SF. Ceftazi-
dime encephalopathy: absence
status and toxic hallucinations.
J Neurol Neurosurg Psychiatry. 1992;
55:333–334.
26. Jallon P, Fankhauser L, Du Pasqu-
ier R, et al. Severe but reversible
encephalopathy associated with
cefepime. Neurophysiol Clin. 2000;
30:383–386.
27. Conil JM, Georges B, Fourcade O,
et al. Assessment of renal function
in clinical practice at the bedside
of burn patients. Br J Clin Pharma-
col. 2007;63:583–594.
28. Loirat P, Rohan J, Baillet A, et al.
Increased glomerular filtration rate
in patients with major burns and
its effect on the pharmacokinetics
of tobramycin. N Engl J Med.
1978;299:915–919.
29. Taccone FS, Laterre PF, Dugernier
T, et al. Insufficient beta-lactam
concentrations in the early phase
of severe sepsis and septic shock.
Crit Care. 2010;14:R126.
30. Seyler L, Cotton F, Taccone FS,
et al. Recommended beta-lactam
regimens are inadequate in septic
patients treated with continuous
renal replacement therapy. Crit
Care. 2011;15:R137.
Volume 35 Number 10