International Journal of Antimicrobial Agents 40 (2012) 313–322

Contents lists available at SciVerse ScienceDirect

International Journal of Antimicrobial Agents

journal homepage: http://www.elsevier.com/locate/ijantimicag

Comparison of antimicrobial pharmacokinetic/pharmacodynamic breakpoints

with EUCAST and CLSI clinical breakpoints for Gram-positive bacteria
Eduardo Asín a , Arantxazu Isla a , Andrés Canut b , Alicia Rodríguez Gascón a,∗
a
Pharmacokinetics, Nanotechnology and Gene Therapy Group, Faculty of Pharmacy, University of the Basque Country, 01006 Vitoria-Gasteiz, Spain
b
Microbiology Unit, Hospital Universitario de Álava, 01004 Vitoria-Gasteiz, Spain

a r t i c l e i n f o a b s t r a c t

Article history: This study compared the susceptibility breakpoints based on pharmacokinetic/pharmacodynamic
Received 7 February 2012 (PK/PD) models and Monte Carlo simulation with those defined by the Clinical and Laboratory Stan-
Accepted 8 June 2012 dards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) for
antibiotics used for the treatment of infections caused by Gram-positive bacteria. A secondary objective
Keywords: was to evaluate the probability of achieving the PK/PD target associated with the success of antimicrobial
Monte Carlo simulation
therapy. A 10 000-subject Monte Carlo simulation was executed to evaluate 13 antimicrobials (47 intra-
Minimum inhibitory concentration (MIC)
venous dosing regimens). Susceptibility data were extracted from the British Society for Antimicrobial
Pharmacokinetics/pharmacodynamics
Staphylococci
Chemotherapy database for bacteraemia isolates. The probability of target attainment and the cumula-
Enterococci tive fraction of response (CFR) were calculated. No antibiotic was predicted to be effective (CFR ≥ 90%)
Streptococci against all microorganisms. The PK/PD susceptibility breakpoints were also estimated and were com-
pared with CLSI and EUCAST breakpoints. The percentages of strains affected by breakpoint discrepancies
were calculated. In the case of ␤-lactams, breakpoint discrepancies affected <15% of strains. However,
higher differences were detected for low doses of vancomycin, daptomycin and linezolid, with PK/PD
breakpoints being lower than those defined by the CLSI and EUCAST. If this occurs, an isolate will be
considered susceptible based on CLSI and EUCAST breakpoints although the PK/PD analysis predicts fail-
ure, which may explain treatment failures reported in the literature. This study reinforces the idea of
considering not only the antimicrobial activity but also the dosing regimen to increase the probability of
clinical success of an antimicrobial treatment.
© 2012 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

1. Introduction To maximise the likelihood of a favourable clini-

The resistance rates of Gram-positive bacteria, in par-
ticular meticillin-resistant Staphylococcus aureus (MRSA),
vancomycin-intermediate S. aureus, vancomycin-resistant S.
aureus, vancomycin-resistant enterococci (VRE) and multidrug-
resistant Streptococcus pneumoniae, are a major public health
problem worldwide. Among the interventions to mitigate the
current and future impact of antimicrobial resistance, the devel-
opment of new generations of antimicrobials is one of the most
accepted. Another recognised strategy to diminish antibiotic
resistance is optimisation of the dosing regimen of available
antimicrobials [1].
∗ Corresponding author. Present address: Laboratory of Pharmacy and Pharma-
ceutical Technology, Faculty of Pharmacy, University of the Basque Country, Paseo
de la Universidad 7, 01006 Vitoria-Gasteiz, Spain. Tel.: +34 945 01 30 94;
fax: +34 945 01 30 40.
E-mail address: alicia.rodriguez@ehu.es (A. Rodríguez Gascón).
cal/microbiological response as well as to minimise the probability
of exposure-related toxicity, pharmacokinetic/pharmacodynamic
(PK/PD) modelling represents a very useful tool for dose decision-
making. Use of Monte Carlo simulation provides an estimate
of an antibiotic dosing regimen’s probability of achieving the
targeted pharmacodynamic exposure, given uncertainty in patient
pharmacokinetics and the minimum inhibitory concentration
(MIC) distribution of the bacterial population [2]. Based on Monte
Carlo simulation, we showed differences in the probability of
success of several dosing regimens of vancomycin, linezolid,
daptomycin and tigecycline for the treatment of MRSA infections
in four Western European countries owing to differences in
susceptibility patterns [3]. Another application of PK/PD analysis
and Monte Carlo simulation is the establishment of breakpoints
based on the likelihood of obtaining a targeted exposure [4].
Pharmacodynamics is considered by regulatory agencies for the
development of susceptibility breakpoints to be used by clinical
microbiology laboratories to categorise organisms as susceptible
or resistant. However, divergences between the probability of
pharmacodynamic target attainment and current susceptibility

0924-8579/$ – see front matter © 2012 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.
http://dx.doi.org/10.1016/j.ijantimicag.2012.06.005
314 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322

Table 1
Pharmacokinetic parameters for each antimicrobial agent from published studies among healthy adult volunteers (mean ± standard deviation).

Antimicrobial agent and dosing regimen Administration CLt (mL/min) Vss (L/kg) AUCSS (mg h/L) PB (%)

Amoxicillin [9]
1 g q6h 0.5-h infusion 237.20 ± 44.17 0.31 ± 0.09 – 20
1 g q8h
2 g q6h
2 g q8h

Cloxacillin [10]
1 g q6h Bolus and 1-h infusion 152.8 ± 29.1 0.10 ± 0.02 – 78
1 g q4h
2 g q6h
2 g q4h

Piperacillin/tazobactam [11]
4 g q6h 0.5-h infusion 170.4 ± 35.3 0.15 ± 0.02 – 30
4 g q8h

Cefotaxime [12]
1 g q6h Bolus, 0.5-h and 1-h infusion 275.33 ± 50.11 0.207 ± 0.035 – 38
2 g q6h

Cefepime [22]
1 g q8h Bolus 125 ± 21 0.25 ± 0.04 – 20
1 g q12h
2 g q8h 143 ± 25 0.23 ± 0.05 – 20
2 g q12h

Ertapenem [22]
1 g q12h 0.5-h infusion 29.5 ± 3.4 0.12 ± 0.02 – 95
1 g q24h

Imipenem [22]
500 mg q8h 0.5-h infusion 175 ± 23 0.22 ± 0.05 – 8.7
500 mg q6h
1 g q8h
1 g q6h

Meropenem [22]
500 mg q8h 0.5-h infusion 240 ± 30 0.27 ± 0.04 – 8
500 mg q6h
1 g q8h
1 g q6h

Levofloxacin [22]
500 mg q24h 1-h infusion – – 54.6 ± 11.1 –

Vancomycin [14]
1 g q12h 1-h infusion 77 ± 22 – – –
1 g q8h
1.5 g q8h
1.5 g q6h
2 g q12h

Daptomycin [15,16]
4 mg/kga q24h 0.5-h infusion – – 494.0 ± 75.0 –
6 mg/kga q24h – – 631.8 ± 12.3 –
8 mg/kga q24h – – 858.2 ± 24.9 –
10 mg/kga q24h – – 1038.8 ± 17.2 –
12 mg/kga q24h – – 1277.4 ± 19.8 –

Tigecycline [17]
50 mg q12h 0.5-h infusion – – 6.14 ± 0.76 –

Linezolid [18]
600 mg q12h 0.5-h infusion – – 179.4 ± 62.0 –

CLt , total body clearance; Vss , apparent volume of distribution at steady-state; AUCss , area under the antimicrobial concentration–time curve for 24 h; PB, protein binding;
qxh, every x h.
a
Dose calculated for a standard weight (70 kg).

percentages based on breakpoints defined by the Clinical and
Laboratory Standards Institute (CLSI) and the European Commit-
tee on Antimicrobial Susceptibility Testing (EUCAST) have been
identified [5]. These divergences have contributed to revising the
breakpoints by the CLSI or EUCAST. For instance, the CLSI decreased
susceptibility breakpoints of Enterobacteriaceae to cephalosporins
[6]. The CLSI also revised the penicillin susceptibility breakpoint
for pneumococcal infection outside of the central nervous system
and moved it from 0.06 mg/L to 2 mg/L [7].
Breakpoint divergences have been studied less extensively in
Gram-positive bacteria, with the exception of glycopeptides [8] or
the abovementioned penicillin and pneumococcal infection. The
main objective of this study was to compare the susceptibility
breakpoints from a PK/PD perspective with the breakpoints defined
by the CLSI or EUCAST for Gram-positive bacteria. Detection of
divergences in the breakpoints could be useful to explain failures in
the treatment of infections by microorganisms considered as sus-
ceptible to the antibiotics used to eradicate the infection process.
 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322 315

A secondary objective of this study was to evaluate the probabil- Table 2

Pharmacokinetic/pharmacodynamic (PK/PD) index and target magnitude for each
ity of obtaining the PK/PD target with 47 dosing regimens of 13
antimicrobial agent.
antibiotics used against Gram-positive bacteria.
Antimicrobial agent PK/PD index Magnitude

2. Methods Amoxicillin [22] f%T>MIC 50

Cloxacillin [22] f%T>MIC 50
Piperacillin/tazobactam [22] f%T>MIC 50
2.1. Antimicrobials
Cefotaxime [22] f%T>MIC 50
Cefepime [22] f%T>MIC 50
Thirteen antimicrobials [47 intravenous (i.v.) dosing regimens] Ertapenem [22] f%T>MIC 30
were chosen based on their routine use for treatment of Gram- Imipenem [22] f%T>MIC 30
positive infections. Table 1 lists all the dosing regimens studied. Meropenem [22] f%T>MIC 30
Levofloxacin [20] AUCSS /MIC 125
Pharmacokinetic parameters were obtained from published
Vancomycin [3] AUCSS /MIC 400
pharmacokinetic studies. A search of the parameters was per- Daptomycin [3] AUCSS /MIC 666
formed using the US National Library of Medicine’s ISI Web Tigecycline [3] AUCSS /MIC 18
Of Knowledge combining the Medical Subject Heading (MeSH) Linezolid [3] AUCSS /MIC 100
‘pharmacokinetics’ or ‘population pharmacokinetics’ with each f%T>MIC , percentage of time that the antimicrobial free serum concentration
antimicrobial’s International Non-proprietary Name (INN). Selec- remained above the MIC; MIC, minimum inhibitory concentration; AUCSS /MIC, ratio
tion of the different studies depended on the number of of the area under the antimicrobial concentration–time curve for 24 h divided by the
MIC.
participants, human race (Caucasian people) and age, year of publi-
cation and pharmacokinetic model. Just one study was selected for
each antimicrobial, except for daptomycin for which two studies
(%), Vss is the antimicrobial’s apparent volume of distribution at
were selected, reporting pharmacokinetic data for all regimens to
steady-state (L), CLt is the total body clearance (L/h),  is the dos-
be studied. All parameters were expressed as mean and standard
ing interval (h) and n is the dose number (fixed at 10). The subject
deviation (S.D.) (Table 1).
weight was fixed at 70 kg when necessary.
• For i.v. infusion-administered ␤-lactams:
2.2. Microbiology

MIC distributions for Enterococcus spp. (Enterococcus faecalis MIC − fCmin,ss

t1 = × tinf
and Enterococcus faecium), Staphylococcus spp. [meticillin-resistant fCmax,ss − fCmin,ss
coagulase-negative staphylococci (MR-CoNS), MRSA, meticillin-  fC  Vss
max,ss
susceptible coagulase-negative staphylococci (MS-CoNS) and t2 = Ln ×
MIC CLt
meticillin-susceptible S. aureus (MSSA)] and Streptococcus spp.
(Streptococcus pyogenes, Streptococcus agalactiae, Streptococcus 100
f %T>MIC = [(t2 + tinf ) − t1 ] ×
anginosus, Streptococcus oralis and S. pneumoniae) were extracted 
from the British Society for Antimicrobial Chemotherapy (BSAC)
database corresponding to bacteraemia isolates collected at hos- where fCmin,ss is the minimum serum concentration of unbound
pitals from the UK and Ireland for every antimicrobial agent. All drug (mg/L) at steady-state, fCmax,ss is the maximum serum con-
data were collected during 2009, except in the following cases: centration of unbound drug (mg/L) at steady-state, tinf is the
ertapenem (2007) and meropenem (2008) against enterococci and infusion time (h), t1 corresponds to the time at which the con-
streptococci. centration reached the MIC during the infusion phase and t2
corresponds to the post-infusion time at which the serum con-
2.3. Pharmacokinetic/pharmacodynamic analyses and Monte centration equalled the MIC in the elimination phase.
Carlo simulation • For the rest of the antimicrobial agents:

2.3.1. Probability of target attainment (PTA) estimation

A 10 000-subject Monte Carlo simulation was conducted
for each antibiotic using Oracle® Crystal Ball Fusion Edition
v.11.1.1.1.00 (Oracle USA Inc., Redwood City, CA). The PTA [19]
is understood as the probability that a specific value of a PK/PD
index associated with the efficacy of the antimicrobial treatment is
achieved at a certain MIC. Table 2 shows the magnitude or value of
the PK/PD indices associated with the success of therapy for each
antimicrobial. The PTA was calculated over a range of doubling MICs
between 0.002 mg/L and 512 mg/L using the following equations.
• For i.v. bolus-administered ␤-lactams:
 
D × (1 − PB) × (1 − exp(−n × (CLt /Vss ) × ))
f %T>MIC = Ln
Vss × MIC × (1 − exp(−(CLt /Vss ) × ))
Vss 100
× ×
CLt  knowledge of the clinical pathogen’s susceptibility and was calcu-
where f%T>MIC is the proportion of time that the free serum con-
centration remains above the MIC at steady-state (%), Ln is the
natural logarithm, D is the dose of antibiotic administered (mg),
PB is the fraction of drug bound to proteins in human serum
AUCss D × 24
=
MIC MIC ×  × CLt
where AUCss is the area under the serum concentration–time
curve at steady-state over 24 h (mg h/L) [19].
Logarithmic transformation was applied to the mean and S.D.
of all pharmacokinetic parameters to normalise their distributions,
whereas protein binding was included following a uniform distri-
bution (±10%).
2.3.2. Calculation of the cumulative fraction of response (CFR)
CFR [19] is defined as the expected population PTA for a specific
drug dose and a specific population of microorganisms. It allowed
us to calculate the probability of success for a treatment without
lated taking into account the PTA for each MIC and the bacterial
population MIC distribution. A CFR ≥ 80% but <90% was associated
with moderate probabilities of success, whereas a CFR ≥ 90% was
considered as optimal against that bacterial population [2].
316 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322
2.3.3. Comparison of susceptibility breakpoints
PK/PD breakpoints were set at the highest MIC value with bac-
tericidal target attainments ≥90% as this is the accepted target
attainment cut-off currently used by the CLSI when determining
MIC breakpoints [21]. Then, PK/PD breakpoints were compared
with CLSI (M100-S22) and EUCAST (version 2.0) breakpoints for
2012 to evaluate the impact of discrepancies on the interpretation
of global organism susceptibilities. Breakpoint discrepancies were
considered as modest if they affected <10% of strains or large if they
affected ≥15% of the strains [22].
3. Results
3.1. Probability of target attainment and cumulative fraction of
response
Table 3 shows the PTA (%) at MIC values ranging from 0.002 mg/L
to 512 mg/L for each regimen calculated by PK/PD analysis and
Monte Carlo simulation. Table 4 shows the CFR values (%) for
all antibiotic regimens against Gram-positive cocci. High prob-
abilities of success against every Gram-positive cocci were not
obtained with any of the dosing regimens studied. Regarding ente-
rococci, only CFR values ≥90% against E. faecalis were obtained
with piperacillin/tazobactam [4 g every 6 h (q6h)], imipenem and
meropenem (≥500 mg q6h), daptomycin [≥8 mg/kg every 24 h
(q24h)] and tigecycline. For E. faecium, CFR ≥ 90% was only achieved
with tigecycline.
Regarding staphylococci, high probabilities of success were
obtained with daptomycin (doses ≥8 mg/kg q24h) and van-
comycin (1.5 g q6h). Cloxacillin (except bolus 1 g q6h) and
piperacillin/tazobactam were predicted to achieve a CFR ≥ 90%
against meticillin-susceptible staphylococci, whereas tigecycline
provided a CFR ≥ 90% only against MSSA. A CFR ≥ 90% was also
achieved with linezolid against MR-CoNS, but the success proba-
bilities were lower for the rest of the staphylococci.
With regard to streptococci, all ␤-lactam regimens, vancomycin
[≥1 g every 8 h (q8h)], daptomycin (≥8 mg/kg q24h) and tigecycline
provided CFR values ≥90%. We should highlight that CFR values
<90% were obtained with linezolid for every Gram-positive cocci
except MR-CoNS.
3.2. Breakpoint comparisons
Table 5 shows the PK/PD breakpoints calculated and the break-
points published by EUCAST and the CLSI. It is important to take
into account that PK/PD breakpoints are regimen-dependent and
species-independent; therefore, they are expressed as a range from
the lowest to the highest breakpoint achieved with the different
dosing regimens of each antimicrobial agent. When comparing
breakpoints, differences between those obtained by PK/PD anal-
ysis and those published by EUCAST and the CLSI, when detected,
were generally within 1 or 2 doubling dilutions.
With regard to enterococci and staphylococci, when discrepan-
cies were observed, PK/PD breakpoints were generally lower than
published breakpoints. No differences were found for imipenem
and meropenem against enterococci and staphylococci. Regarding
streptococci, PK/PD breakpoints were generally similar to those
published by EUCAST and the CLSI. However, PK/PD breakpoints
were higher for cefepime, imipenem and meropenem and were
lower for levofloxacin and linezolid.
Table 6 depicts the PTA (%) at EUCAST and CLSI breakpoints.
When there was coincidence between EUCAST and CLSI pub-
lished breakpoints and PK/PD breakpoints, the success likelihood
is high (PTA ≥ 90%) at committee breakpoints. However, when
CLSI or EUCAST breakpoints were higher than those calculated
by PK/PD analysis, low PTA values were observed at breakpoints
set by international committees. This is the case for levofloxacin
against S. pneumoniae (PK/PD breakpoint, 0.25 mg/L; EUCAST
and CLSI breakpoint, 2 mg/L; PTA 0%). Another antibiotic with
important differences in the PTA at the breakpoint given by the
CLSI and EUCAST is linezolid against staphylococci (PK/PD break-
point, 1 mg/L; EUCAST and CLSI breakpoint, 4 mg/L; PTA 1%). More
differences in the PTAs at the different breakpoints are shown in
Table 6.
3.3. Impact of divergent breakpoints on susceptibility
interpretations
When differences in the PTA at the PK/PD breakpoint and
the EUCAST and CLSI breakpoint were detected, the percentage
of strains affected by these divergences was calculated. Table 7
presents the cumulative frequency distribution for enterococci,
staphylococci and streptococci obtained from the BSAC database
and the percentage of strains affected by discrepancies in the
breakpoints. As PK/PD breakpoints depend on administered dose,
discrepancies have been calculated for each antimicrobial dosing
regimen. Table 7 only includes cases in which divergences affected
≥15% of the strains. Cloxacillin was the only ␤-lactam present-
ing significant percentages of affected strains (≥15%), although not
with all dosing regimens. Breakpoint discrepancies affected signif-
icant percentages also for vancomycin, even at high doses (≥1.5 g
q8h), although these percentages varied depending on the dose
and the microorganism. For instance, differences in the breakpoints
found with vancomycin 1 g every 12 h affected almost 100% of
strains for some staphylococci, and lower percentages, but also
significant, for the rest of the cocci. In contrast, the differences
in breakpoints with vancomycin 1.5 g q6h affected 24% of the E.
faecalis bacterial population, but no significant percentages for the
rest of the cocci. Breakpoint differences for high doses (≥8 mg/kg
q24h) of daptomycin against enterococci affected 83% of the strains
of E. faecium; however, only 4% of the E. faecalis strains were
affected. Daptomycin also presented discrepancies in the break-
points against almost all cocci, mainly observed with the low
doses (≤6 mg/kg q24h). Linezolid presented discrepancies between
breakpoints for the majority of the cocci, affecting a high proportion
of the strains (Table 7).
4. Discussion
Gram-positive cocci such as enterococci, staphylococci
and streptococci are a common aetiology of nosocomial and
community-acquired infections. The main problem influencing
antimicrobial therapy against these bacteria is the emergence of
resistant strains such as VRE, MRSA or MR-CoNS. Inappropriate use
of antimicrobials could be responsible for a higher risk of treatment
failure, higher mortality, increased cost, increased toxicity and
the increasing emergence of resistance. Therefore, selection of
an agent with activity against the pathogen responsible for the
infection and use of a dosing regimen that guarantees an optimal
concentration at the infection site are essential for correct therapy.
Knowledge of how the dosing regimen affects the exposure of
the microorganisms presenting a certain MIC to the antimicrobial
agent is of great relevance for optimal selection [23]. In this regard,
Monte Carlo simulation is an advanced statistical modelling tool
that allows estimating the influence of the dosing regimen on the
probability of attaining the PK/PD target associated with antibiotic
efficacy considering the variability in pharmacokinetic parameters
and the natural MIC distribution within a bacterial population [24].
In this study, we have estimated using Monte Carlo simu-
lation the CFR of antibiotics considered clinically useful for the
Table 3
Probability of target attainment (PTA) (%) at each value of minimum inhibitory concentration (MIC).a
Antimicrobial agent and dosing regimen MIC (mg/L)

0.002 0.004 0.008 0.015 0.03 0.06 0.125 0.25 0.5 1 2 4 8 16 32 64 128 256 512
Amoxicillin
1 g q6h 100 100 100 100 100 100 100 100 100 99 96 77 16 0 0 0 0 0 0
1 g q8h 100 100 100 100 100 100 100 99 97 90 71 29 1 0 0 0 0 0 0
2 g q6h 100 100 100 100 100 100 100 100 100 100 99 95 77 16 0 0 0 0 0
2 g q8h 100 100 100 100 100 100 100 100 99 97 90 72 30 1 0 0 0 0 0
Cloxacillin
1 g q6h (bolus) 100 100 100 99 99 97 91 78 57 29 6 0 0 0 0 0 0 0 0
1 g q4h (bolus) 100 100 100 100 100 100 100 99 96 85 55 16 1 0 0 0 0 0 0
2 g q6h (bolus) 100 100 100 100 99 99 96 90 77 57 29 7 0 0 0 0 0 0 0
2 g q4h (bolus) 100 100 100 100 100 100 100 100 99 97 86 56 16 1 0 0 0 0 0
1 g q6h (1-h inf.) 100 100 100 100 100 100 98 95 83 56 20 2 0 0 0 0 0 0 0
1 g q4h (1-h inf.) 100 100 100 100 100 100 100 100 100 99 86 40 2 0 0 0 0 0 0
2 g q6h (1-h inf.) 100 100 100 100 100 100 100 99 95 84 56 19 1 0 0 0 0 0 0
2 g q4h (1-h inf.) 100 100 100 100 100 100 100 100 100 100 99 86 39 2 0 0 0 0 0

E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322

Piperacillin/tazobactam
4 g q6h 100 100 100 100 100 100 100 100 100 100 100 98 89 59 13 0 0 0 0
4 g q8h 100 100 100 100 100 100 100 100 99 97 91 75 45 13 1 0 0 0 0
Cefotaxime
1 g q6h (bolus) 100 100 100 100 100 100 99 96 89 68 32 4 0 0 0 0 0 0 0
2 g q6h (bolus) 100 100 100 100 100 100 100 99 96 88 67 31 4 0 0 0 0 0 0
1 g q6h (0.5-h inf.) 100 100 100 100 100 100 100 98 94 79 44 8 0 0 0 0 0 0 0
2 g q6h (0.5-h inf.) 100 100 100 100 100 100 100 100 99 94 80 46 8 0 0 0 0 0 0
1 g q6h (1-h inf.) 100 100 100 100 100 100 100 100 98 91 61 14 0 0 0 0 0 0 0
2 g q6h (1-h inf.) 100 100 100 100 100 100 100 100 100 98 91 62 15 0 0 0 0 0 0
Cefepime
1 g q12h 100 100 100 100 100 100 100 100 100 97 84 39 2 0 0 0 0 0 0
1 g q8h 100 100 100 100 100 100 100 100 100 100 100 97 56 1 0 0 0 0 0
2 g q12h 100 100 100 100 100 100 100 99 98 93 79 49 13 0 0 0 0 0 0
2 g q8h 100 100 100 100 100 100 100 100 100 100 99 96 78 25 0 0 0 0 0
Ertapenem
1 g q12h 100 100 100 100 100 99 98 97 94 88 77 52 6 0 0 0 0 0 0
1 g q24h 100 100 100 100 99 99 97 93 87 74 48 6 0 0 0 0 0 0 0
Imipenem
500 mg q8h 100 100 100 100 100 100 100 100 100 100 100 92 6 0 0 0 0 0 0
500 mg q6h 100 100 100 100 100 100 100 100 100 100 100 100 67 0 0 0 0 0 0
1 g q8h 100 100 100 100 100 100 100 100 100 100 100 100 92 6 0 0 0 0 0
1 g q6h 100 100 100 100 100 100 100 100 100 100 100 100 100 67 0 0 0 0 0
Meropenem
500 mg q8h 100 100 100 100 100 100 100 100 100 100 99 4 0 0 0 0 0 0 0
500 mg q6h 100 100 100 100 100 100 100 100 100 100 100 97 2 0 0 0 0 0 0
1 g q8h 100 100 100 100 100 100 100 100 100 100 100 99 41 0 0 0 0 0 0
1 g q6h 100 100 100 100 100 100 100 100 100 100 100 100 97 1 0 0 0 0 0
Levofloxacin
500 mg q24h 100 100 100 100 100 100 100 100 22 0 0 0 0 0 0 0 0 0 0
Vancomycin
1 g q12h 100 100 100 100 100 100 100 100 100 66 2 0 0 0 0 0 0 0 0
1 g q8h 100 100 100 100 100 100 100 100 100 97 27 0 0 0 0 0 0 0 0
1.5 g q8h 100 100 100 100 100 100 100 100 100 100 81 5 0 0 0 0 0 0 0
1.5 g q6h 100 100 100 100 100 100 100 100 100 100 97 27 0 0 0 0 0 0 0
2 g q12h 100 100 100 100 100 100 100 100 100 100 66 2 0 0 0 0 0 0 0
Daptomycin
4 mg/kg q24h 100 100 100 100 100 100 100 100 99 2 0 0 0 0 0 0 0 0 0
6 mg/kg q24h 100 100 100 100 100 100 100 100 100 0 0 0 0 0 0 0 0 0 0
8 mg/kg q24h 100 100 100 100 100 100 100 100 100 100 0 0 0 0 0 0 0 0 0
10 mg/kg q24h 100 100 100 100 100 100 100 100 100 100 0 0 0 0 0 0 0 0 0
12 mg/kg q24h 100 100 100 100 100 100 100 100 100 100 0 0 0 0 0 0 0 0 0
Tigecycline
50 mg q12h 100 100 100 100 100 100 100 99 0 0 0 0 0 0 0 0 0 0 0
Linezolid
600 mg q12h 100 100 100 100 100 100 100 100 100 94 32 1 0 0 0 0 0 0 0
qxh, every x h; inf., infusion.
a
Grey shading indicates PTA ≥ 90%.

317
Table 4

318
Cumulative fractions of response (CFR) (%) calculated for all dosing regimens.a
Antimicrobial agent and dosing regimen Enterococcus spp. Staphylococcus spp. Streptococcus spp.

E. faecalis E. faecium MR-CoNS MRSA MS-CoNS MSSA S. pyogenes S. agalactiae S. anginosus S. oralis S. pneumoniae
Amoxicillin
1 g q6h NA NA – – NA NA 100 100 100 97 100
1 g q8h NA NA – – NA NA 100 99 100 93 99
2 g q6h NA NA – – NA NA 100 100 100 99 100
2 g q8h NA NA – – NA NA 100 100 100 96 100
Cloxacillin
1 g q6h (bolus) NA NA – – 87 72 – – – – –
1 g q4h (bolus) NA NA – – 99 98 – – – – –
2 g q6h (bolus) NA NA – – 94 86 – – – – –
2 g q4h (bolus) NA NA – – 100 99 – – – – –
1 g q6h (1-h inf.) NA NA – – 96 91 – – – – –
1 g q4h (1-h inf.) NA NA – – 100 100 – – – – –
2 g q6h (1-h inf.) NA NA – – 98 97 – – – – –
2 g q4h (1-h inf.) NA NA – – 100 100 – – – – –

E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322

Piperacillin/tazobactam
4 g q6h 97 2 – – 100 99 100 100 100 99 100
4 g q8h 75 0 – – 97 92 100 100 100 96 99
Cefotaxime
1 g q6h (bolus) NA NA – – NA NA 100 100 96 92 96
2 g q6h (bolus) NA NA – – NA NA 100 100 99 97 99
1 g q6h (0.5-h inf.) NA NA – – NA NA 100 100 99 96 100
2 g q6h (0.5-h inf.) NA NA – – NA NA 100 100 100 99 100
1 g q6h (1-h inf.) NA NA – – NA NA 100 100 100 98 100
2 g q6h (1-h inf.) NA NA – – NA NA 100 100 100 100 100
Cefepime
1 g q8h NA NA NA NA NA NA NA NA NA NA NA
1 g q12h NA NA NA NA NA NA NA NA NA NA NA
2 g q8h NA NA NA NA NA NA NA NA NA NA NA
2 g q12h NA NA NA NA NA NA NA NA NA NA NA
Ertapenem
1 g q12h 23 0 – – NA NA 100 100 98 99 99
1 g q24h 7 0 – – NA NA 100 99 96 97 98
Imipenem
500 mg q8h 100 3 – – NA NA 100 100 100 100 100
500 mg q6h 100 3 – – NA NA 100 100 100 100 100
1 g q8h 100 3 – – NA NA 100 100 100 100 100
1 g q6h 100 3 – – NA NA 100 100 100 100 100
Meropenem
500 mg q8h 76 3 – – NA NA 100 100 100 100 100
500 mg q6h 93 3 – – NA NA 100 100 100 100 100
1 g q8h 95 3 – – NA NA 100 100 100 100 100
1 g q6h 97 3 – – NA NA 100 100 100 100 100
Levofloxacin
500 mg q24h NA NA NA NA NA NA NA NA NA NA NA
Vancomycin
1 g q12h 17 38 6 55 8 29 99 77 69 89 95
1 g q8h 37 57 29 83 33 56 100 97 97 99 99
1.5 g q8h 64 64 76 96 78 89 100 100 100 100 100
1.5 g q6h 78 65 91 99 93 98 100 100 100 100 100
2 g q12h 57 62 63 93 65 80 100 100 100 100 100
Daptomycin
4 mg/kg q24h 38 4 72 77 69 86 100 99 85 71 100
6 mg/kg q24h 38 4 72 77 69 86 100 99 84 71 100
8 mg/kg q24h 96 18 99 100 98 100 100 100 100 96 100
10 mg/kg q24h 96 17 99 100 98 100 100 100 100 96 100
12 mg/kg q24h 96 17 99 100 98 100 100 100 100 95 100
Tigecycline
50 mg q12h 100 100 41 84 70 97 100 100 100 100 100
Linezolid
600 mg q12h 36 38 92 34 85 35 34 36 75 82 60
MR-CoNS, meticillin-resistant coagulase-negative staphylococci; MRSA, meticillin-resistant Staphylococcus aureus; MS-CoNS, meticillin-susceptible coagulase-negative staphylococci; MSSA, meticillin-susceptible S. aureus; qxh,
every x h; NA, data not available; –, microorganism is a poor target for therapy with the drug; inf., infusion.
a
Dark grey shading indicates CFR ≥ 90%; light grey shading indicates 80% ≤ CFR < 90%.
 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322 319

Table 5
Comparison of the pharmacokinetic/pharmacodynamic (PK/PD), European Committee on Antimicrobial Susceptibility Testing (EUCAST) and Clinical and Laboratory Standards
Institute (CLSI) breakpoints (mg/L) for Gram-positive cocci.

Antimicrobial agent PK/PDa Enterococcus Staphylococcus ␤-Haemolytic Other streptococci Streptococcus

streptococci pneumoniae

EUCAST CLSI EUCAST CLSI EUCAST CLSI EUCAST CLSI EUCAST CLSI

Amoxicillin 1–4 4 NA NA 4 NA NA 0.5 NA NA 2

Cloxacillin 0.125–2 – – 2b 2b NA – – – – NA
Piperacillin/tazobactam 2–4 NA NA NA 8 NA NA NA NA NA NA
Cefotaxime 0.25–2 – – NA 8 NA 0.5 0.5 1 0.5 1
Cefepime 1–4 – – NA 8 NA 0.5 0.5 1 1 1
Ertapenem 0.25–0.5 – – NA 2 NA 1 0.5 1 0.5 1
Imipenem 4–8 4 NA NA 4 NA NA 2 NA 2 0.125
Meropenem 2–8 – – NA 4 NA 0.5 2 0.5 2 0.25
Levofloxacin 0.25 – 2 1 1 1 2 – 2 2 2
Vancomycin 0.5–2 4 4 2c 2c 2 1 2 1 2 1
Daptomycin 0.5–1 IE 4 1 1 1 1 – 1 IE NA
Tigecycline 0.25 0.25 NA 0.5 NA 0.25 NA IE NA IE NA
Linezolid 1 4 2 4 4 2 2 IE 2 2 2

NA, no breakpoint available; –, susceptibility testing is not recommended as the species is a poor target for therapy with the drug; IE, insufficient evidence that the species
in question is a good target for therapy with the drug.
a
PK/PD breakpoints are regimen-dependent and species-independent, so they are expressed as a range from lowest to highest breakpoints achieved with the various
regimens for each antimicrobial agent.
b
If coagulase-negative staphylococci, 0.25 mg/L.
c
If coagulase-negative staphylococci, 4 mg/L.

treatment of infections caused by Gram-positive cocci (Section
3.1). This parameter is useful for empirical therapy when the MIC
value of the pathogen is unknown. No treatment provided high
CFR values (≥90%) against all microorganisms. Tigecycline was a
good option under the suspicion of an enterococcal infection. For
infections by staphylococci, high doses of vancomycin and dapto-
mycin were found to be the best options; and for the empirical
treatment of infections due to streptococci, all antibiotics studied,
except linezolid, provided high success probabilities.
Monte Carlo simulation can be used to develop interpretive sus-
ceptibility criteria based on PK/PD principles (PK/PD breakpoints).
A susceptibility breakpoint is defined as a cut-off MIC that allows
categorising a bacterial population based on its susceptibility and,
therefore, establishing the antimicrobial dose [25]. Susceptibility
breakpoints are typically set by the integration of a variety of micro-
biological, PK/PD and clinical data [5,13]. Although the CLSI and
EUCAST now use PK/PD simulations for estimation of the break-
points of new antibiotics, only some of the CLSI breakpoints of
older antimicrobials have been reviewed, most of them for Gram-
negative bacteria. However, in Europe, harmonisation of existing
breakpoints seems to be more important.
In this study, we evaluated the discrepancies between the PK/PD
breakpoints and those of EUCAST and CLSI (Section 3.2). Most
detected discrepancies involved 1 or 2 MIC dilutions. Although the
PK/PD simulations support the conservation of existing breakpoints
for the ␤-lactams, higher differences in the breakpoints of other
antimicrobials (levofloxacin, vancomycin, daptomycin or linezolid)
were detected (Table 7). When large discrepancies were observed,
the PK/PD breakpoints were generally lower than those defined by
the CLSI or EUCAST. If this occurs, an isolate will be considered
as susceptible based on CLSI and EUCAST breakpoints although
the PK/PD analysis predicts failure. This situation may explain
clinical failures of treatments with antimicrobials against microor-
ganisms considered susceptible. Gallagher et al. [26] reported a
20% microbiological failure with daptomycin (6 mg/kg q24h) for
the treatment of VRE bacteraemia, although susceptibility testing
revealed daptomycin-susceptible VRE. The breakpoint defined by
the CLSI for daptomycin against VRE is 4 mg/L, but the PK/PD break-
point estimated in the current study for 6 mg/kg q24h daptomycin
was 0.5 mg/L, with estimated differences in susceptibility of 63%
and 96% for E. faecalis and E. faecium, respectively (Table 7). The
authors suggested that the efficacy rate would have been higher if
all patients had received doses >6 mg/kg q24h. It is important to
note that PK/PD breakpoints increase as the dose increases.
Clinical failure cases with linezolid have also been reported
in the literature. Administration of linezolid to treat a
patient with post-neurosurgical meningitis by E. faecium in a
UK hospital resulted in failure despite susceptibility testing that
revealed a linezolid-susceptible isolate (MIC = 2 mg/L) [27]. For
linezolid, discrepancies between the breakpoints defined by the
CLSI (2 mg/L) and EUCAST (4 mg/L) and the PK/PD breakpoint
estimated in this study (1 mg/L) affected up to 91% of the strains.
Another treatment failure case with linezolid affected a patient
with pneumonia caused by MRSA [28]; linezolid was adminis-
tered following vancomycin and gentamicin failure. Susceptibility
testing carried out in bronchoalveolar fluid revealed susceptibility
to linezolid, with a MIC < 2 mg/L (EUCAST and CLSI breakpoints,
4 mg/L, whereas the PK/PD breakpoint estimated in this study
is 1 mg/L). As shown previously, high probabilities of treatment
failure with linezolid for isolates with a MIC of 2 mg/L were
predicted in this study (Section 3.1).
In a previous study [22], discrepancies between breakpoints
defined by EUCAST and the CLSI and PK/PD breakpoints were
also detected against Gram-negative aerobic bacteria. As shown
for Gram-positive cocci, large discrepancies in breakpoints may
explain therapy failures with antibiotics used for the treatment
of patients infected by Gram-negative pathogens. Bhat et al. [13],
based on pharmacodynamic grounds, suggested failure for current
cefepime breakpoints to predict clinical outcomes of bacteraemia
caused by Gram-negative organisms (8 mg/L) and proposed to
lower the breakpoints. In fact, in a study by Frei et al. [22],
large divergences in percentage susceptibility of cefepime were
reported, which is consistent with the thesis put forward by Bhat
et al.
It is important to keep in mind that PK/PD breakpoints are based
on simulations and that a number of considerations are assumed.
Frei et al. [22] explained these limitations in a previous publication
about PK/PD analysis and Monte Carlo simulation. (i) Pharmacoki-
netic information is often obtained from healthy subjects rather
than patients and it is known that pathophysiological conditions
affect the distribution and elimination of antimicrobials [29]. (ii)
Pharmacokinetic equations used to estimate drug exposure are
usually simple models that facilitate calculations. (iii) The simu-
lation in this paper is based on serum pharmacokinetics; therefore,
320 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322

Table 6
Probability of target attainment (PTA) (%) at European Committee on Antimicrobial Susceptibility Testing (EUCAST) and Clinical and Laboratory Standards Institute (CLSI)
published breakpoints.a

Antimicrobial agent and dosing regimen Enterococcus Staphylococcus ␤-Haemolytic Other streptococci Streptococcus
streptococci pneumoniae

EUCAST CLSI EUCAST CLSI EUCAST CLSI EUCAST CLSI EUCAST CLSI

Amoxicillin
1 g q6h 77 NA NA 77 NA NA 100 NA NA 96
1 g q8h 29 NA NA 29 NA NA 97 NA NA 71
2 g q6h 95 NA NA 95 NA NA 100 NA NA 99
2 g q8h 72 NA NA 72 NA NA 99 NA NA 90
Cloxacillinb
1 g q6h (bolus) – – 6 6 NA – – – – NA
1 g q4h (bolus) – – 55 55 NA – – – – NA
2 g q6h (bolus) – – 29 29 NA – – – – NA
2 g q4h (bolus) – – 86 86 NA – – – – NA
1 g q6h (1-h inf.) – – 20 20 NA – – – – NA
1 g q4h (1-h inf.) – – 86 86 NA – – – – NA
2 g q6h (1-h inf.) – – 56 56 NA – – – – NA
2 g q4h (1-h inf.) – – 99 99 NA – – – – NA
Piperacillin/tazobactam
4 g q6h NA NA NA 89 NA NA NA NA NA NA
4 g q8h NA NA NA 45 NA NA NA NA NA NA
Cefotaxime
1 g q6h (bolus) – – NA 0 NA 89 89 68 89 68
2 g q6h (bolus) – – NA 4 NA 96 96 88 96 88
1 g q6h (0.5-h inf.) – – NA 0 NA 94 94 79 94 79
2 g q6h (0.5-h inf.) – – NA 8 NA 99 99 94 99 94
1 g q6h (1-h inf.) – – NA 0 NA 98 98 91 98 91
2 g q6h (1-h inf.) – – NA 15 NA 100 100 98 100 98
Cefepime
1 g q12h – – NA 2 NA 100 100 97 97 97
1 g q8h – – NA 56 NA 100 100 100 100 100
2 g q12h – – NA 13 NA 98 98 93 93 93
2 g q8h – – NA 78 NA 100 100 100 100 100
Ertapenem
1 g q12h – – NA 77 NA 88 94 88 94 88
1 g q24h – – NA 48 NA 74 87 74 87 74
Imipenem
500 mg q8h 92 NA NA 92 NA NA 100 NA 100 100
500 mg q6h 100 NA NA 100 NA NA 100 NA 100 100
1 g q8h 100 NA NA 100 NA NA 100 NA 100 100
1 g q6h 100 NA NA 100 NA NA 100 NA 100 100
Meropenem
500 mg q8h – – NA 4 NA 100 99 100 99 100
500 mg q6h – – NA 97 NA 100 100 100 100 100
1 g q8h – – NA 99 NA 100 100 100 100 100
1 g q6h – – NA 100 NA 100 100 100 100 100
Levofloxacin
500 mg q24h – 0 0 0 0 0 – 0 0 0
Vancomycinc
1 g q12h 0 0 2 2 2 66 2 66 2 66
1 g q8h 0 0 27 27 27 97 27 97 27 97
1.5 g q8h 5 5 81 81 81 100 81 100 81 100
1.5 g q6h 27 27 97 97 97 100 97 100 97 100
2 g q12h 2 2 66 66 66 100 66 100 66 100
Daptomycin
4 mg/kg q24h IE 0 2 2 2 2 – 2 IE NA
6 mg/kg q24h IE 0 0 0 0 0 – 0 IE NA
8 mg/kg q24h IE 0 100 100 100 100 – 100 IE NA
10 mg/kg q24h IE 0 100 100 100 100 – 100 IE NA
12 mg/kg q24h IE 0 100 100 100 100 – 100 IE NA
Tigecycline
50 mg q12h 99 NA 0 NA 99 NA IE NA IE NA
Linezolid
600 mg q12h 1 32 1 1 32 32 IE 32 32 32

NA, no breakpoint available; –, susceptibility testing is not recommended as the species is a poor target for therapy with the drug; IE, insufficient evidence that the species
in question is a good target for therapy with the drug; inf., infusion.
a
Grey shading indicates PTA ≥ 90%.
b
PTA (%) for CLSI and EUCAST breakpoint against coagulase-negative staphylococci were: if bolus administration, 78% (1 g q6h), 99% (1 g q4h), 90% (2 g q6h) and 100% (2 g
q4h); and after 1-h inf. administration, 95% (1 g q6h), 100% (1 g q4h), 99% (2 g q6h) and 100% (2 g q4h).
c
PTA (%) for EUCAST and CLSI breakpoint against coagulase-negative staphylococci were: 0% (1 g q12h), 0% (1 g q8h), 5% (1.5 g q8h), 27% (1.5 g q6h) and 2% (2 g q12h).
 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322 321

Table 7
Cumulative frequency distribution for Gram-positive cocci isolates from the British Society for Antimicrobial Chemotherapy (BSAC) database with the corresponding per-
centage susceptible using pharmacokinetic/pharmacodynamic (PK/PD), European Committee on Antimicrobial Susceptibility Testing (EUCAST) and Clinical and Laboratory
Standards Institute (CLSI) breakpoints.a,b

Antimicrobial agent MIC (mg/L) Divergence in % susceptible

0.125 0.25 0.5 1 2 4 8 >256

Cloxacillin A B
MS-CoNS 75PA 96PBEC 98 98 100 100 100 100 21 –
MSSA 11PA 68PB 97 100 100EC 100 100 100 90 32
Vancomycin D F G
Enterococcus faecalis 0 0 0PD 25PF 72PG 96EC 96 100 96 72 24
Enterococcus faecium 0 0 0PD 57PF 65PG 65EC 65 100 65 – –
MR-CoNS 0 0 0PD 6PF 92PG 100EC 100 100 100 94 –
MRSA 0 0 2PD 80PF 100PGEC 100 100 100 98 20 –
MS-CoNS 0 0 0PD 10PF 94PG 100EC 100 100 100 90 –
MSSA 0 0 0PD 42PF 100PGEC 100 100 100 100 58 –
Streptococcus agalactiae 0 0 35PD 99PFC 100PGE 100 100 100 65 – –
Streptococcus anginosus 0 0 11PD 100PFC 100PGE 100 100 100 90 – –
Streptococcus oralis 0 0 67PD 100PFC 100PGE 100 100 100 33 – –
Streptococcus pneumoniae 0 1 85PD 100PFC 100PGE 100 100 100 15 – –
Daptomycin H I
E. faecalis 2 3 37PH 96PI 99 100C 100 100 63 –
E. faecium 0 0 4PH 17PI 84 100C 100 100 96 83
MR-CoNS 1 8 72PH 99PIEC 100 100 100 100 27 –
MRSA 0 2 78PH 100PIEC 100 100 100 100 22 –
MS-CoNS 0 8 69PH 98PIEC 100 100 100 100 29 –
S. anginosus 0 32 84PH 100PIC 100 100 100 100 16 –
S. oralis 0 18 71PH 96PIC 100 100 100 100 24 –
Linezolid J
E. faecalis 0 0 0 6PJ 100C 100E 100 100 94
E. faecium 0 0 0 9PJ 100C 100E 100 100 91
MRSA 0 0 0 3PJ 100 100EC 100 100 97
MSSA 0 0 0 3PJ 100 100EC 100 100 97
S. pyogenes 0 0 0 3PJ 100EC 100 100 100 97
S. agalactiae 0 0 0 6PJ 100EC 100 100 100 94
S. anginosus 0 0 5 68PJ 100C 100 100 100 32
S. oralis 0 0 16 78PJ 100C 100 100 100 22
S. pneumoniae 0 0 0 46PJ 100EC 100 100 100 55

MIC, minimum inhibitory concentration; MSSA, meticillin-susceptible S. aureus; MS-CoNS, meticillin-susceptible coagulase-negative staphylococci; MR-CoNS, meticillin-
resistant coagulase-negative staphylococci; MRSA, meticillin-resistant Staphylococcus aureus; qxh, every x h.
P: PK/PD breakpoints; E: EUCAST breakpoints; C: CLSI breakpoints.
a
Table only includes cases in which divergence affected ≥15% of the strains; dashes are used when divergence was <15%.
b
The dose-dependent nature of PK/PD breakpoints (P) is shown as different annotations: cloxacillin (A: 1 g q6h bolus, 0.125 mg/L; B: 2 g q6h bolus and 1 g q6h 1-h infusion,
0.25 mg/L); vancomycin (D: 1 g q12h, 0.5 mg/L; F: 1 g q8h, 1.5 g q8h and 2 g q12h, 1 mg/L; G: 1.5 g q6h, 2 mg/L); daptomycin (H: 4 mg/kg and 6 mg/kg q24h, 0.5 mg/L; I: 8, 10
and 12 mg/kg q24h, 1 mg/L); and linezolid (J: 600 mg q12h, 1 mg/L).

these findings are most readily applicable to bloodstream infections
[22]. However, this could result in a reduction of the applicabil-
ity of these findings to other sites of infection such as respiratory
tract infections by S. pneumoniae. And (iv) PK/PD modelling is based
on prior studies that have identified correlations between PK/PD
indices and health outcomes. These relationships have typically
been derived from animal models; however, a good correlation
and validation between animal models and the clinic has been
established [30]. Finally, another limitation is that pathogen sus-
ceptibility could differ hugely depending on geographic region
because of antimicrobial use and local susceptibility.
Despite the above limitations, PK/PD analysis and Monte Carlo
simulations are very useful tools to estimate susceptibility break-
points. However, it is important to take into account that other
parameters should be considered to establish valid breakpoints,
such as clinical and microbiological data and information on resis-
tance mechanisms. In this sense, promotion of prospective trials
that include pharmacokinetic analysis, microbiological susceptibil-
ity and clinical data would help to establish a correlation between
breakpoints and clinical outcome.
5. Conclusion
To guarantee clinical success in empirical therapy, based on
these findings and limitations, we can conclude that tigecycline
resulted in a good option under the suspicion of an enterococcal
infection. For infections by staphylococci, high doses of vancomycin
and daptomycin resulted in the best options; and for the empirical
treatment of infections due to streptococci, all antibiotics stud-
ied, except linezolid, provided high success probabilities. When
comparing the PK/PD breakpoints for numerous antimicrobial
classes against Gram-positive cocci with those defined by interna-
tional committees such as the CLSI and EUCAST, large divergences
were observed for some antibiotics (low doses of vancomycin,
linezolid and daptomycin), which may explain treatment failures
reported in the literature. Finally, this study reinforces the idea of
considering not only the antimicrobial activity but also the dos-
ing regimen to increase the probability of clinical success of an
antimicrobial treatment.
Funding: This work was supported by the Departamento de
Educación, Universidades e Investigación (IT341-10) and by the
Departamento de Sanidad (2009111062), Gobierno Vasco, Spain.
The authors would also like to thank the Departamento de Edu-
cación, Universidades e Investigación, Gobierno Vasco, Spain, for a
research grant awarded to EA.
Competing interests: None declared.
Ethical approval: Not required.
References
[1] Olofsson SK, Cars O. Optimizing drug exposure to minimize selection of antibi-
otic resistance. Clin Infect Dis 2007;45(Suppl. 2):S129–36.
322 E. Asín et al. / International Journal of Antimicrobial Agents 40 (2012) 313–322
[2] Bradley JS, Dudley MN, Drusano GL. Predicting efficacy of antiinfectives
with pharmacodynamics and Monte Carlo simulation. Pediatr Infect Dis J
2003;22:982–92.
[3] Canut A, Isla A, Betriu C, Gascón AR. Pharmacokinetic–pharmacodynamic
evaluation of daptomycin, tigecycline, and linezolid versus vancomycin for
the treatment of MRSA infections in four western European countries. Eur J
Clin Microbiol Infect Dis 2012, http://dx.doi.org/10.1007/s10096-012-1560-7
[Epub ahead of print].
[4] Ambrose PG. Monte Carlo simulation in the evaluation of susceptibility break-
points: predicting the future: insights from the Society of Infectious Diseases
Pharmacists. Pharmacotherapy 2006;26:129–34.
[5] Ambrose PG. Antimicrobial susceptibility breakpoints: PK-PD and susceptibil-
ity breakpoints. Treat Respir Med 2005;4(Suppl. 1):5–11.
[6] DeRyke CA, Kuti JL, Nicolau DP. Reevaluation of current susceptibility break-
points for Gram-negative rods based on pharmacodynamic assessment. Diagn
Microbiol Infect Dis 2007;58:337–44.
[7] Weinstein MP, Klugman KP, Jones RN. Rationale for revised penicillin
susceptibility breakpoints versus Streptococcus pneumoniae: coping with
antimicrobial susceptibility in an era of resistance. Clin Infect Dis 2009;48:
1596–600.
[8] Rybak M, Lomaestro B, Rotschafer JC, Moellering R, Craig W, Billeter M, et al.
Therapeutic monitoring of vancomycin in adult patients: a consensus review
of the American Society of Health-System Pharmacists, the Infectious Diseases
Society of America, and the Society of Infectious Diseases Pharmacists. Am J
Health Syst Pharm 2009;66:82–98.
[9] Staniforth D, Jackson D, Horton R, Davies B. Parenteral augmentin: pharma-
cokinetics. Int J Clin Pharmacol Ther Toxicol 1984;22:430–4.
[10] Nauta EH, Mattie H. Pharmacokinetics of flucloxacillin and cloxacillin in healthy
subjects and patients on chronic intermittent haemodialysis. Br J Clin Pharma-
col 1975;2:111–21.
[11] Kim MK, Capitano B, Mattoes HM, Xuan D, Quintiliani R, Nightingale CH, et al.
Pharmacokinetic and pharmacodynamic evaluation of two dosing regimens for
piperacillin–tazobactam. Pharmacotherapy 2002;22:569–77.
[12] Nix DE, Wilton JH, Hyatt J, Thomas J, Strenkoski-Nix LC, Forrest A, et al. Phar-
macodynamic modeling of the in vivo interaction between cefotaxime and
ofloxacin by using serum ultrafiltrate inhibitory titers. Antimicrob Agents
Chemother 1997;41:1108–14.
[13] Bhat SV, Peleg AY, Lodise Jr TP, Shutt KA, Capitano B, Potoski BA, et al. Failure
of current cefepime breakpoints to predict clinical outcomes of bacteremia
caused by Gram-negative organisms. Antimicrob Agents Chemother 2007;51:
4390–5.
[14] Bauer LA, Black DJ, Lill JS. Vancomycin dosing in morbidly obese patients. Eur J
Clin Pharmacol 1998;54:621–5.
[15] Dvorchik BH, Brazier D, DeBruin MF, Arbeit RD. Daptomycin pharmacokinetics
and safety following administration of escalating doses once daily to healthy
subjects. Antimicrob Agents Chemother 2003;47:1318–23.
[16] Benvenuto M, Benziger DP, Yankelev S, Vigliani G. Pharmacokinetics and
tolerability of daptomycin at doses up to 12 milligrams per kilogram of
body weight once daily in healthy volunteers. Antimicrob Agents Chemother
2006;50:3245–9.
[17] Muralidharan G, Micalizzi M, Speth J, Raible D, Troy S. Pharmacokinetics of tige-
cycline after single and multiple doses in healthy subjects. Antimicrob Agents
Chemother 2005;49:220–9.
[18] MacGowan AP. Pharmacokinetic and pharmacodynamic profile of linezolid in
healthy volunteers and patients with Gram-positive infections. J Antimicrob
Chemother 2003;51(Suppl. 2):ii17–25.
[19] Mouton JW, Dudley MN, Cars O, Derendorf H, Drusano GL. Standardization
of pharmacokinetic/pharmacodynamic (PK/PD) terminology for anti-infective
drugs. Int J Antimicrob Agents 2002;19:355–8.
[20] Benko R, Matuz M, Doro P, Peto Z, Molnar A, Hajdu E, et al. Pharmacokinetics
and pharmacodynamics of levofloxacin in critically ill patients with ventilator-
associated pneumonia. Int J Antimicrob Agents 2007;30:162–8.
[21] Maglio D, Kuti JL, Nicolau DP. Simulation of antibiotic pharmacodynamic expo-
sure for the empiric treatment of nosocomial bloodstream infections: a report
from the OPTAMA program. Clin Ther 2005;27:1032–42.
[22] Frei CR, Wiederhold NP, Burgess DS. Antimicrobial breakpoints for Gram-
negative aerobic bacteria based on pharmacokinetic–pharmacodynamic
models with Monte Carlo simulation. J Antimicrob Chemother 2008;61:621–8.
[23] Liu P, Müller M, Derendorf H. Rational dosing of antibiotics: the use of
plasma concentrations versus tissue concentrations. Int J Antimicrob Agents
2002;19:285–90.
[24] Roberts JA, Kirkpatrick CMJ, Lipman J. Monte Carlo simulations: maximizing
antibiotic pharmacokinetic data to optimize clinical practice for critically ill
patients. J Antimicrob Chemother 2011;66:227–31.
[25] Turnidge J, Paterson DL. Setting and revising antibacterial susceptibility break-
points. Clin Microbiol Rev 2007;20:391–408.
[26] Gallagher JC, Perez ME, Marino EA, LoCastro LG, Abrardo LA, MacDougall C.
Daptomycin therapy for vancomycin-resistant enterococcal bacteremia: a ret-
rospective case series of 30 patients. Pharmacotherapy 2009;29:792–9.
[27] Webster DP, Griffiths S, Bowler IC. Failure of linezolid therapy for post-
neurosurgical meningitis due to Enterococcus faecium. J Antimicrob Chemother
2009;63:622–3.
[28] Saner FH, Heuer M, Rath P, Gensicke J, Radtke A, Drühe N, et al. Successful
salvage therapy with tigecycline after linezolid failure in a liver transplant
recipient with MRSA pneumonia. Liver Transpl 2006;12:1689–92.
[29] Pea F, Viale P. The antimicrobial therapy puzzle: could
pharmacokinetic–pharmacodynamic relationships be helpful in address-
ing the issue of appropriate pneumonia treatment in critically ill patients?
Clin Infect Dis 2006;42:1764–71.
[30] Ambrose PG, Bhavnani SM, Rubino CM, Louie A, Gumbo T, Forrest A, et al.
Pharmacokinetics–pharmacodynamics of antimicrobial therapy: it’s not just
for mice anymore. Clin Infect Dis 2007;44:79–86.