The Journal of Clinical

Pharmacology
http://www.jclinpharm.org

Influence of Renal Function on the Pharmacokinetics of Piperacillin/Tazobactam in Intensive Care Unit

Patients During Continuous Venovenous Hemofiltration
Alazne Arzuaga, Javier Maynar, Alicia R. Gascón, Arantxazu Isla, Esther Corral, Fernando Fonseca, José Ángel
Sánchez-Izquierdo, Jordi Rello, Andrés Canut and José Luis Pedraz
J. Clin. Pharmacol. 2005; 45; 168
DOI: 10.1177/0091270004269796

The online version of this article can be found at:

http://www.jclinpharm.org/cgi/content/abstract/45/2/168

Published by:

http://www.sagepublications.com

On behalf of:
American College of Clinical Pharmacology

Additional services and information for The Journal of Clinical Pharmacology can be found at:

Email Alerts: http://www.jclinpharm.org/cgi/alerts

Subscriptions: http://www.jclinpharm.org/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from http://www.jclinpharm.org by Luciana Oliveira on April 21, 2009


ARTICLE
ARZUAGACARE
10.1177/0091270004269796
PHARMACOKINETICS
CRITICAL ET AL OF PIPERACILLIN/TAZOBACTAM DURING CVVH
Influence of Renal Function on the
Pharmacokinetics of Piperacillin/Tazobactam
in Intensive Care Unit Patients During
Continuous Venovenous Hemofiltration
Alazne Arzuaga, MSc, Javier Maynar, MD, Alicia R. Gascón, PharmD, Arantxazu Isla, MSc,
Esther Corral, MD, Fernando Fonseca, MD, José Ángel Sánchez-Izquierdo, PhD,
Jordi Rello, PhD, Andrés Canut, PhD, and José Luis Pedraz, PharmD
The pharmacokinetics of piperacillin/tazobactam (4 g/0.5 g
every 6 or 8 hours, by 20-minute intravenous infusion) were
studied in 14 patients with acute renal failure who underwent
continuous venovenous hemofiltration with AN69 mem-
branes. Patients were grouped according to severity (CLCR
10 mL/min, 10 < CLCR 50 mL/min, and CLCR > 50 mL/min). A
noncompartmental analysis was performed. The sieving co-
efficient (0.78 ± 0.28) was similar to the unbound fraction
(0.65 ± 0.24) for tazobactam, but it was significantly different
(0.34 ± 0.25) from the unbound fraction (0.78 0.14) for
piperacillin. Extracorporeal clearance was 37.0% ± 28.8%,
12.7% ± 12.6%, and 2.8% ± 3.2% for piperacillin in each
group and 62.5% ± 44.9%, 35.4% ± 17.0%, and 13.1% ± 8.0%
for tazobactam. No patients presented tazobactam accumu-
A cute renal failure (ARF) is a frequent complication
in critically ill patients with sepsis. Continuous
venovenous hemofiltration (CVVH) is becoming the
core of the supportive care of these patients.1 Sepsis is a
From the Laboratory of Pharmacy and Pharmaceutical Technology, Faculty
of Pharmacy, University of the Basque Country, Vitoria-Gasteiz, Spain (A.
Arzuaga, Dr Gascón, A. Isla, Dr Pedraz); Intensive Care Unit, Santiago
Apóstol Hospital, Vitoria-Gasteiz, Spain (Dr Maynar, Dr Corral, Dr
Fonesca); Intensive Care Unit, Doce de Octubre Hospital, Madrid, Spain
(Dr Sánchez-Izquierdo); Intensive Care Unit, Joan XXIII University Hospital,
University Rovira & Virgili, Tarragona, Spain (Dr Rello); and Microbiology
Unit, Santiago Apóstol Hospital, Vitoria-Gasteiz, Spain (Dr Canut). This
project was supported by the Basque Government (PI-1999-34). Submit-
ted for publication May 27, 2004; revised version accepted August 3,
2004. Address for reprints: Dr J. L. Pedraz, Laboratorio de Farmacia y
Tecnología Farmacéutica, Facultad de Farmacia, Paseo de la Universidad
no. 7, 01006 Vitoria-Gasteiz, Spain.
DOI: 10.1177/0091270004269796
168 • J Clin Pharmacol 2005;45:168-176
CRITICAL CARE
lation. In patients with CLCR < 50 mL/min, t ss
(%) > MIC90 values
were 100% for a panel of 19 pathogens, but in those with CLCR
> 50 mL/min, t ss
(%) > MIC90 indexes were 55.5% and 16.6% for
pathogens with MIC90 values of 32 and 64. The extracor-
poreal clearance of piperacillin/tazobactam is clinically sig-
nificant in patients with CLCR > 50 mL/min, in which the risk
of underdosing and clinical failure is important and extra
doses are required.
Keywords: continuous venovenous hemofiltration (CVVH);
appropriate antimicrobial therapy; piperacillin;
tazobactam; pharmacokinetics; peritonitis;
sepsis
Journal of Clinical Pharmacology, 2005;45:168-176
©2005 the American College of Clinical Pharmacology
contributor to the development of ARF, and most
patients with CVVH require proper antibiotic adminis-
tration to ensure a good outcome.2,3 Continuous veno-
venous hemofiltration maintains fluid and electrolyte
balance by removing fluid and solutes, but it also re-
moves drugs, and thus it complicates dosing regimens.
Some authors have reported the usefulness of this ther-
apy in critically ill patients without renal function im-
pairment: trauma patients,4 cardiovascular surgery pa-
tients,5 and septic patients.6,7
The elimination of antimicrobials depends primar-
ily on the size of the pores in the filter used, the molec-
ular size, the level of protein binding, and adsorption to
the filter.8,9 Accumulation of the drug may be associ-
ated with serious adverse effects. On the other hand,
subtherapeutic levels may be associated with the emer-
gence of bacterial resistance. Therefore, appropriate
 PHARMACOKINETICS OF PIPERACILLIN/TAZOBACTAM DURING CVVH

Table I Clinical Characteristics of the Patients Included in the Study, Grouped by Renal Function
Case Weight,
Group Number Age, y Gender kg Diagnoses AP II SOFA pH TP Alb TB

I 1 48 Male 70 Trauma, VAP 18 12 7.39 4.8 2.1 1.0

2 67 Male 76 Peritonitis 21 10 7.40 3.7 1.5 1.0
3 69 Female 68 UTI sepsis 36 15 7.36 5.4 2.8 4.0
4 37 Male 73 Crohn’s disease,
sepsis 10 17 7.15 4.8 1.7 2.7
Mean 55.3 71.8 21.0 13.5 7.33 4.7 2.0 2.0
Standard deviation 15.4 3.5 11.0 3.1 0.12 0.7 0.6 1.7
II 5 72 Male 77 Peritonitis 21 11 7.39 3.8 2.3 1.9
6 72 Female 70 Peritonitis 25 12 7.50 3.9 2.5 0.7
7 72 Male 71 Peritonitis 32 14 7.29 3.6 1.8 3.4
8 56 Female 86 Necrotizing
fascitis 18 9 7.40 5.1 2.7 0.9
9 50 Male 65 Abdominal
surgery, VAP 25 9 7.45 6.5 3.1 0.7
Mean 64.4 73.8 24.0 11.0 7.41 4.6 2.5 0.8
Standard deviation 10.6 8.0 5.3 2.1 0.08 1.2 0.5 0.1
III 10 77 Male 74 Peritonitis,
VAP, AMI 23 10 7.40 4.7 2.2 1.6
11 48 Male 80 Trauma, VAP 15 8 7.37 6.5 1.9 0.5
12 18 Male 80 Trauma, MODS 19 8 7.33 4.6 1.8 2.2
13 49 Male 90 Trauma, UTI 22 8 7.35 5.0 1.7 6.7
14 57 Male 78 Peritonitis, VAP 24 11 7.31 4.2 2.5 0.4
Mean 49.8 80.4 21.0 9.0 7.35 5.0 2.0 2.2
Standard deviation 21.3 5.9 3.6 1.4 0.03 0.9 0.3 2.5
All patients
Mean 56.6 75.6 22.0 11.0 7.36 4.8 2.2 1.9
Standard Deviation 16.4 6.9 3.5 2.8 0.08 0.9 0.5 1.9
Group I: CLCR ≤ 10 mL/min; Group II: 10 < CLCR ≤ 50 mL/min; Group III: CLCR > 50 mL/min. AP II, Acute Physiology and Chronic Health Evaluation II; SOFA, Se-
quential Organ Failure Assessment; TP, total plasma protein concentration (g/dL); Alb, plasma albumin concentration (g/dL); TB, total plasma bilirubin concen-
tration (mg/L); VAP, ventilator-associated pneumonia; UTI, urinary tract infection; AMI, acute myocardial infarction; MODS, multiple organ dysfunction
syndrome.

management of critically ill infected patients is manda-
tory to avoid treatment failures and to improve sur-
vival. Appropriate management is based on in vitro
sensitivities, penetration to the infection site, correct
dosing of antibiotics, and avoidance of delay therapy
onset.10,11
Piperacillin/tazobactam is a β-lactam/β-lactamase
inhibitor combination with a broad spectrum of
antibacterial activity that is frequently prescribed in
the intensive care unit (ICU), particularly for sepsis,
pneumonia, and intra-abdominal infections.12,13 Both
components are particularly suitable for coadmini-
stration because it has been reported that they have
broadly similar pharmacokinetic profiles in adults and
children.14 The physicochemical properties (water sol-
ubility, protein binding, and molecular size) of both
drugs are highly comparable,15 and it is anticipated that
the elimination kinetics during CVVH will be similar.
Van der Werf et al15 reported that the fixed combination
of piperacillin/tazobactam during CVVH in anuric pa-
tients results in some accumulation of tazobactam.
Our objective was to study the pharmacokinetics of
piperacillin and tazobactam during CVVH in ICU pa-
tients with various degrees of renal impairment. Al-
though previous studies15-18 have explored this issue,
all of them have been performed in patients with acute
renal failure and anuria.
MATERIAL AND METHODS
Patients
Fourteen patients treated with piperacillin/tazobactam
and CVVH were included in the study. The study pro-
tocol was approved by the Medical Ethical Committee

CRITICAL CARE 169

 ARZUAGA ET AL
of the 2 participating hospitals (Santiago Apostol Hos-
pital and Doce de Octubre Hospital) and was per-
formed in accordance with the ethical standards de-
tailed in the 1964 Declaration of Helsinki. Patient
characteristics and diagnoses are presented in Table I.
Procedures
Vascular access was obtained with 13.5-FG dual-
lumen catheters (Niagara, Bard, Ontario, Canada). A
Prisma (HOSPAL, Lyon, France) machine was used for
CVVH with a 0.9-m2 AN69 acrylonitrile and sodium
methyl sulfonate copolymer filter (PRISMA M100
HOSPAL, Lyon, France). The blood flow was kept be-
tween 150 and 220 mL/min, and the ultrafiltrate flow
varied depending on the renal function (Table II). Re-
placement fluid was delivered prefilter. Prefilter blood
and ultrafiltrate samples were collected at 0, 0.3, 0.5,
0.75, 1, 3, 6, and 8 hours (in case of administration ev-
ery 8 hours) after the administration of Tazocel. Time
0 was considered just before the beginning of the 20-
minute infusion. Blood samples were obtained using
EDTA as anticoagulant. The blood specimens were
centrifuged for 10 minutes at 1000g, and the plasma
and ultrafiltrate samples were stored at –80°C until
drug analysis. Trough (just before the beginning of the
next administration) and peak (at the end of the infu-
sion) samples were also obtained for several days. Pro-
tein binding of both compounds was measured by
ultrafiltration using Sartorius Centrisart I filters (cutoff
10.000) (Sartorius AG, Goettingen, Germany).
Drug Assay
Determination of piperacillin and tazobactam concen-
trations in plasma and ultrafiltrate fluid was performed
by high-performance liquid chromatography (HPLC)
with a Waters (Milford, Mass) apparatus coupled to a
spectophotometric detector. All the analytical methods
were conveniently validated.19,20 The assay was linear
over the concentration range from 5 to 500 µg/mL for
plasma samples and from 1 to 100 µg/mL for
ultrafiltrate. The intraday and interday coefficients of
variation ranged from 2.71% to 10.83% for plasma
samples and from 1.30% to 4.60% for ultrafiltrate sam-
ples at the 3 concentrations tested (8, 80, and 400 µg/
mL for plasma and 5, 40, and 75 µg/mL for ultrafiltrate).
The bias at these concentrations ranged from 0.69% to
14.7% in plasma samples and from 0.20% to 9.33% in
ultrafiltrate fluid. The limit of quantification was con-
sidered the lowest level included in the calibration
curve (5 µg/mL in plasma and 1 µg/mL in ultrafiltrate),
in which measures of the intraday coefficient of varia-
170 • J Clin Pharmacol 2005;45:168-176
tion and bias were 8.96% and 0.77% for plasma
samples and 3.15% and 4.85% for ultrafiltrate sam-
ples, respectively. No interfering peaks were detected
with the assay. For tazobactam, the assay was linear
over the concentration range from 3 to 50 µg/mL. The
intraday and interday coefficients of variation ranged
from 1.31% to 1.89% in plasma samples and from
0.39% to 2.85% in ultrafiltrate samples at the 3 concen-
trations tested (8, 20, and 40 µg/mL for all samples).
The bias at these concentrations ranged from 0.10% to
7.79% in plasma and from 0.05% to 14.28% in ultra-
filtrate fluid. The limit of quantification was consid-
ered the lowest level included in the calibration curve
(3 µg/mL), in which measures of the intraday coeffi-
cient of variation and bias were 3.32% and 1.70% for
plasma samples and 3.66% and 2.26% for ultrafiltrate
samples, respectively. No interfering peaks were
detected with the assay.
Pharmacokinetic and Statistical Analysis
Plasma and ultrafiltrate concentrations of piperacillin
and tazobactam were plotted against time, and individ-
ual pharmacokinetic parameters were determined
from plasma levels according to a noncompartmental
analysis by using WinNonlin version 1.1 (Pharsight
Corporation, Mountain View, Calif). The plasma and
ultrafiltrate areas under the curve (AUCs) were deter-
mined from the first to the last data point by the linear
trapezoidal method. The terminal elimination rate
constant was determined via log-linear regression
analysis using the terminal portion of the plasma drug
concentration versus time curves. Half-life was derived
from this rate constant as follows: t1/2 = ln(2)/ke. The to-
tal body clearance (CL) was obtained by the equation
CL = dose/AUC. The total mean residence time (MRT)
was calculated as MRT = AUMC/AUC, where AUMC is
the area under the moment curve, and the volume of
distribution at steady-state (Vss) was obtained by Vss =
MRT × CL.
The sieving coefficient (Sc) was calculated as Sc =
AUCHF/AUCP, where AUCHF is the area under the
ultrafiltrate versus time curve, and AUCP is the area un-
der the plasma (collected in the prefilter port) versus
time curve.
The hemofiltration clearance (CLHF) of tazobactam
and piperacillin was calculated with the following
equation: CLHF = AUCHF × QHF/AUCP, where QHF is the
ultrafiltrate flow rate. The total amount of the drug
eliminated by hemofiltration (XHF) was calculated as
XHF = AUCHF × QHF.
The Mann-Whitney U-test was used to analyze the
data of the pharmacokinetic parameters and the siev-
 PHARMACOKINETICS OF PIPERACILLIN/TAZOBACTAM DURING CVVH

Table II Values of Pharmacokinetic Parameters of Piperacillin and Tazobactam for Patients Undergoing Con-
tinuous Venovenous Hemofiltration Following the Administration of 4 g of Piperacillin (PIP) Plus 0.5 g of
Tazobactam (TZ)
< 10 mL/min 10-50 mL/min > 50 mL/min
Creatinine Clearance (Group I) (Group II) (Group III) Mean ⴞ SD

CLCR, mL/min 8.67 ± 2.31 25.20 ± 7.73 82.40 ± 20.03

QUF, mL/min 27.1 ± 7.8 30.3 ± 4.3 20.0 ± 7.5
EC CLCR, mL/min 27.0 ± 5.4 25.6 ± 3.6 18.6 ± 7.1
α
PIP 72.3 ± 11.2 76.2 ± 18.3 86.4 ± 9.8 78.8 ± 14.2
TZ 63.2 ± 35.8 61.7 ± 19.9 70.5 ± 19.6 65.3 ± 23.5
Sc
PIP 0.42 ± 0.25 0.38 ± 0.37 0.23 ± 0.07 0.34 ± 0.25
TZ 0.76 ± 0.26 0.73 ± 0.32 0.86 ± 0.30 0.78 ± 0.28
Cmax, mg/L
PIP 365.6 ± 232.3 244.5 ± 122.1 160.6 ± 93.2
TZ 38.4 ± 13.4 31.5 ± 5.1 15.7 ± 6.6 a,b
t1/2, h
PIP 7.8 ± 4.2 4.2 ± 2.3 2.6 ± 0.8a
TZ 7.9 ± 3.0 4.1 ± 0.9a 5.0 ± 3.9a
CL, mL/min
PIP 50.0 ± 53.0 90.6 ± 29.9 265.2 ± 152.2a,b
TZ 50.4 ± 38.3 68.2 ± 26.2 180.1 ± 73.9a,b
CLHF, mL/min
PIP 11.45 ± 6.5 12.2 ± 13.2 4.8 ± 3.3a
TZ 20.9 ± 12.6 21.9 ± 9.6 19.6 ± 15.3
CLHF, %
PIP 37.0 ± 28.8 12.7 ± 12.6 2.8 ± 3.2a,b
TZ 62.5 ± 44.9 35.4 ± 17.0 13.1 ± 8.0
XHF, mg
PIP 1995 ± 1335 1049 ± 841 136 ± 152a,b
TZ 319 ± 184 212 ± 112 74 ± 71a,b
XHF, %
PIP 49.9 ± 33.4 22.6 ± 21.0 3.4 ± 3.8a,b
TZ 63.9 ± 36.8 42.3 ± 22.5 14.7 ± 13.9a,b
AUC0-τ, mgh/L
PIP 76 143 ± 49 748 45 445 ± 25 525 17 328 ± 11 134a,b
TZ 23 218 ± 27 943 5501 ± 1344 2098 ± 1030a,b
Vss, L
PIP 21.0 ± 11.7 26.8 ± 19.8 44.9 ± 20.4
TZ 18.9 ± 7.1 21.6 ± 3.0. 60.3 ± 34.6
Values are mean ± SD. QUF, ultrafiltrate flow rate; EC CLCR, extracorporeal creatinine clearance; α, free drug fraction; Sc, sieving coefficient; Cmax, measured peak
concentration of drug in plasma; t1/2, elimination half-life; CL, total plasma clearance; CLHF, hemofiltration clearance, XHF: amount of drug eliminated by
hemofiltration; AUC0-τ, area under the concentration versus time curve from 0 to τ (dosing interval); Vss, volume of distribution at steady state.
a. Significant differences with respect to group I.
b. Significant differences with respect to group II (P < .05).

ing coefficient among the renally impaired groups us- Antimicrobial Susceptibility Assay
ing SPSS 11.5 for Windows (SPSS, Chicago). Statistical
significance was assessed at P < .05. Broth microdilution tests were performed in custom-
dried 96-well microdilution trays (Sensititre Division,

CRITICAL CARE 171

 ARZUAGA ET AL

Accumed International, Westlake, Ohio). Minimum in-

hibitory concentrations (MICs) were determined in ac-
cordance with the methods of the National Committee
for Clinical Laboratory Standards (NCCLS). 2 1
Piperacillin/tazobactam was tested with a constant in-
hibitor concentration of 4 µg/mL in serial 2-fold dilu-
tions (8-64 µg/mL).
Tests were performed with a panel of 19 clinical
strains, including Pseudomonas aeruginosa (n = 4),
Escherichia coli (n = 3), Enterococcus faecalis (n = 2),
Klebsiella pneumoniae (n = 1), Serratia marcenses (n =
1), Burkholderia cepacia (n = 1), Stenotrophomonas
maltophilia (n = 1), Staphylococcus aureus (n = 1), and
Acinetobacter baumannii (n = 1). The quality control Figure 1. Mean piperacillin plasma levels in patients with different
degree of renal impairment and minimum inhibitory concentration
strains used were E. coli ATCC 25922, E. coli ATCC (MIC90) values for the microorganisms found in patients. Group I:
35218, S. aureus ATCC 29213, E. faecalis ATCC 29212, CLCR 10 mL/min (open triangles); Group II: 10 < CLCR 50 mL/min
and B. fragilis ATCC 25285. (open squares); Group III: CLCR > 50 mL/min (filled circles).

RESULTS

Fourteen patients aged 56.6 ± 16.4 years who were ad-

mitted to the intensive care unit, treated with CVVH,
and received tazobactam and piperacillin were en-
rolled in this study. Table I presents patient characteris-
tics and diagnoses. Concomitant drug therapy con-
sisted mainly of midazolam (n = 13 patients), morphine
chloride (n = 11 patients), enoxaparine (n = 9 patients),
amikacin (n = 8 patients), noradrenaline (n = 7 pa-
tients), omeprazole (n = 5 patients), and ranitidine (n =
5 patients).
These 14 patients were grouped into 3 categories ac-
cording to the renal function: 4 patients with severe
failure, CLCR ≤ 10 mL/min; 5 patients with moderate
failure, 10 < CLCR ≤ 50 mL/min; and 5 patients with
mild failure CLCR > 50 mL/min. Patients received
Tazocel (4 g of piperacillin plus 0.5 g of tazobactam) ev-
ery 6 (n = 7) or 8 hours (n = 7) as a 20-minute infusion
(Table I).
After antibiotic onset, a mean of 11 previous doses
was administered before starting sample collection to
achieve steady-state concentrations of the drugs. Fig-
ure 1 shows the mean piperacillin plasma levels in the
patients with different degrees of renal impairment and
the minimum inhibitory concentration (MIC90) values
for the microorganisms tested in these patients. Plasma
levels were above the MIC90 values for all the pathogens
throughout the dose interval in the subset of patients
with a clearance of creatinine under 10 mL/min. Table
II shows the mean pharmacokinetic parameters of
piperacillin in the study population. An increase in the
elimination half-life and a decrease in the total clear-
ance with the degree of renal insufficiency were docu-
Figure 2. Mean tazobactam plasma levels in patients with different
degrees of renal impairment. Group I: CLCR 10 mL/min (open trian-
gles); Group II: 10 < CLCR 50 mL/min (open squares); Group III: CLCR
> 50 mL/min (filled circles).
mented. The contribution of the hemofiltration clear-
ance to the total clearance increased with the degree of
renal insufficiency.
Figure 2 shows the mean plasma concentrations of
tazobactam in the 3 groups of patients considered. Dif-
ferences in the plasma profiles, depending on renal im-
pairment, were similar to those found for piperacillin.
The mean pharmacokinetic parameters obtained for
tazobactam are also presented in Table II. For both
drugs, significant differences were documented in the
majority of the pharmacokinetic parameters when pa-
tients with CLCR > 50 mL/min were compared to pa-
tients with CLCR ≤ 10 mL/min.
In addition, plasma and ultrafiltrate levels of
piperacillin and tazobactam (trough and peak) were
determined at steady state during several dose inter-

172 • J Clin Pharmacol 2005;45:168-176

 PHARMACOKINETICS OF PIPERACILLIN/TAZOBACTAM DURING CVVH
Figure 3. Trough levels of tazobactam measured before the admin-
istration of a new dose during several days after the plasma pharm-
acokinetic study. Group I: CLCR 10 mL/min (open triangles); Group
II: 10 < CLCR 50 mL/min (open squares); Group III: CLCR > 50 mL/
min (filled circles).
vals. Figure 3 shows tazobactam trough levels mea-
sured before the administration of a new dose at steady
state. There was no evidence of accumulation of either
piperacillin or tazobactam after the administration of
multiple doses, even in anuric patients.
Sieving coefficients were 0.34 ± 0.25 and 0.78 ± 0.28
for piperacillin and tazobactam, respectively. Determi-
nations in different subsets of patients are detailed in
Table II. The sieving coefficient of tazobactam (0.78 ±
0.28) is similar to the unbound fraction (0.65 ± 0.24),
but the sieving coefficient of piperacillin (0.34 ± 0.25)
is different from the unbound fraction (0.78 ± 0.14). No
significant differences in the sieving coefficient, de-
pending on renal impairment (neither for piperacillin
nor for tazobactam), were documented.
DISCUSSION
Our study is the first one that provides information
about the pharmacokinetics of piperacillin and tazo-
bactam in critically ill patients with different degrees
of renal impairment undergoing CVVH. The impact of
the extracorporeal removal of both drugs under these
conditions is clearly dependent on renal impairment.
Correct doses of these drugs should take into account
this observation to avoid clinical failures due to
underdosing.11,22
One of the most important factors in the extra-
corporeal elimination of a drug is the sieving coeffi-
cient. This parameter describes the drug fraction elimi-
nated through the membrane and is mainly dependent
on drug protein binding.23-25 Golper23 has shown that
the sieving coefficient of 66 drugs measured during
continuous hemofiltration in humans correlates well
(r = 0.72, P < .001) with the known unbound fraction. In
CRITICAL CARE
our case, the sieving coefficient of tazobactam (0.78 ±
0.28) is similar to the unbound fraction (0.65 ± 0.24),
but the sieving coefficient of piperacillin (0.34 ± 0.25)
is very different from the unbound fraction (0.78 ±
0.14). In a previous study, Capellier et al16 described the
pharmacokinetics of piperacillin in critically ill pa-
tients undergoing CVVH. Although they did not calcu-
late the sieving coefficient, they presented a graphic
with the ultrafiltrate and serum levels, and it is evident
that the sieving coefficient is much lower than the un-
bound fraction. Moreover, in Golper’s list, there are
several examples of drugs with very different values of
the sieving coefficient and the unbound fraction (phos-
phomycin, phenytoin, cyclosporin). Because patients
in our study had multiple organ failure and could have
had the protein binding modified, literature values14,26
may be inappropriate for individual patients. Thus, we
have measured the protein binding in our patients to
compare it with the sieving coefficient. Mueller et al18
have reported a piperacillin sieving coefficient value of
0.84, similar to the unbound fraction, although they
employed continuous venovenous hemodialysis
(CVVHD). It has been reported that amikacin’s sieving
coefficient during slow hemodialysis is smaller than
that obtained by continuous hemofiltration.27,28 Thus,
the sieving coefficient depends on the technique em-
ployed.29 Continuous replacement techniques that in-
volve hemodialysis (CVVHD and CVVHDF) eliminate
low–molecular weight molecules as with these anti-
biotics more efficiently than continuous venovenous
hemofiltration.23,30,31 Other factors, such as the protein
layer in the filter membrane,32 can also affect the hemo-
filtration efficacy. Caution must be taken when using
the unbound fraction instead of the sieving coefficient
to calculate the supplemental dose of a drug during
hemofiltration processes.
The observed sieving coefficient of piperacillin and
tazobactam plus the effluent amount gave a relevant
extracorporeal clearance only in the severe renal im-
pairment group, with more than 25% of total clearance
for both drugs. In the group of patients with moderate
renal impairment, only tazobactam features elimina-
tion with repercussions on the total clearance. Johnson
et al33 had already shown that the pharmacokinetic pa-
rameters of piperacillin and tazobactam are dependent
mainly on renal function. Piperacillin and tazobactam
elimination half-life values progressively increased
with decreasing CLCR, and total plasma clearance de-
creased with decreasing CLCR. For piperacillin and
tazobactam, the estimated values of CL and CLHF (%) in
patients with severe renal failure and their variability
are comparable with values reported for intensive care
patients undergoing CRRT.15,17,18
173
 ARZUAGA ET AL
In terms of tazobactam pharmacokinetics, our find-
ings disagree with a prior study by Van der Werf et al,15
who reported that the administration of a fixed
piperacillin/tazobactam combination during CVVH in
anuric patients resulted in some accumulation of
tazobactam. These authors suggested alternating the
doses of piperacillin alone and piperacillin/
tazobactam, although the toxicity of tazobactam is low.
It is important to consider that the characteristics of the
system, such as the ultrafiltration rates and the filter
membrane, influence the elimination of piperacillin/
tazobactam. In the current study, using a convection
dose in a range from 20 to 35 mL/min, the
accumulation of tazobactam was not observed, even
in those patients with no residual renal function. The
clinical implications of these differences in the out-
comes of patients with severe infections remain uncer-
tain, but it is a practical issue: nowadays, it is not possi-
ble to administer piperacillin alone to our patients.
Besides, even if the tazobactam sieving coefficient
were found to be higher than the piperacillin sieving
coefficient, with tazobactam being eliminated more
efficiently through the hemofiltration membrane,
the piperacillin-tazobactam plasma level relationship
would not be altered, so β-lactamase inhibitor activity
of tazobactam would be good enough, and no changes
in the efficacy due to differences in the sieving coeffi-
cient between piperacillin and tazobactam would be
expected.
The volume of distribution in patients with CLCR >
50 mL/min presented a trend toward being higher than
the one for the group of CLCR ≤ 10 mL/min. Most pa-
tients with CLCR > 50 mL/min were critically ill trauma
patients receiving copious quantities of intravenous
fluid over an extended term of treatment. This can re-
sult in an expanded extracellular compartment, which
could be the reason for the increased distribution vol-
ume. There are several examples of drugs with distri-
bution volume values that increase in critically ill
trauma patients, such as ceftazidime34 and amino-
glycosides.35-38 For piperacillin, Kroh39 reported a dis-
tribution volume greater than the 200% in critically ill
patients. In our study, the distribution volume values
in the 3 groups agree with the different impacts of the
CVVH in the elimination of both drugs. The groups
with a smaller distribution volume have been shown to
have a higher impact on the pharmacokinetic data.
It has been shown that a surrogate marker to predict
the outcome for β-lactam antibiotics is the duration of
time that the plasma concentration exceeds the MIC40
(t ss
(%) > MIC90). For nonsevere infections, it is desirable
to maintain the concentration of these time-killing an-
tibiotics in plasma over the MIC throughout half of the
174 • J Clin Pharmacol 2005;45:168-176
dosing interval. However, studies in neutropenic ani-
mals have shown that the concentration of β-lactam
antibiotics should exceed MICs of pathogens during
the whole dosing interval.41 As the clinical relevance of
a postantibiotic effect is uncertain,41 keeping the
plasma of the β-lactam antibiotic concentration above
the MIC seems to be justified, especially in critically ill
patients.31 In this way, Turnidge42 stated that β-lactam
levels need to exceed the MIC for 90% to 100% of the
dosing interval to be effective against gram-negative
bacilli and streptococci. The t ss (%) > MIC90 values ob-
tained in our study were 100% for all the pathogens in
the patients with creatinine clearance < 10 mL/min. In
the patients who had a creatinine clearance between 10
and 50 mL/min, t ss (%) > MIC90 was 100% for pathogens
with MIC90 ≤ 32, but it decreased to 50% for microor-
ganisms with an MIC90 of 64, such as S. marcenses and
B. cepacia. However, in patients with creatinine clear-
ance > 50 mL/min, as piperacillin elimination was
faster, t ss
(%) > MIC90 was only 55.5% and 16.6% for
pathogens with MIC90 values of 32 and 64, respectively.
Thus, to increase the t ss
(%) > MIC90 index, administration
of the piperacillin/tazobactam combination every 4
hours could be a better dosage regimen in patients pre-
senting CLCR > 50 mL/min.
This study has several limitations. First, penetration
of antibiotics to the target site is variable, and the effi-
cacy may be different in sites with poor drug penetra-
tion. Second, creatinine clearance is variable in criti-
cally ill patients. Thus, frequent monitoring and
further adjustments should be done to note fluctua-
tions, and the calculated dose should take into account
eventual treatment interruptions or the presence of di-
alysis. Obviously, differences between hemofiltration
and hemodialysis may be more pronounced in the
presence of higher dialysate flow rates. Similarly, the
time dependency of the diffusion process can lead to
some inaccurate calculations. Third, in vitro sensitivi-
ties present frequent variations between different insti-
tutions, and these data should not be extrapolated
without taking into account local patterns of sensitiv-
ity. Fourth, the sample size is small, and findings based
on a small sample size should be confirmed by further
studies. Indeed, larger study populations would docu-
ment significant findings in patients with lower
creatinine clearances (e.g., 25 mL/min). Fifth, the case
mix and the volume of distribution may be different in
other series (other modalities of ventilation, different
proportion of shock). Indeed, variations in pharmaco-
kinetic parameters are frequently encountered in
critically ill patients. Thus, one should be cautious
before generalizing the findings to individual patients
in other institutions.
 PHARMACOKINETICS OF PIPERACILLIN/TAZOBACTAM DURING CVVH
In summary, our study indicates that it may not be
correct to assume the equivalence between the un-
bound fraction and the sieving coefficient of
piperacillin. In addition, one should not expect some
accumulation of tazobactam in patients with severe re-
nal impairment and who have undergone CVVH. Our
findings suggest that clearance of piperacillin/
tazobactam is clinically significant in patients under-
going CVVH with a creatinine clearance smaller than
50 mL/min and a convection dose more than 25 mL/
min. In this subset of patients, the risk of underdosing
and clinical failure is important, and it may require the
administration of extra doses of piperacillin/
tazobactam. In critically ill patients with no renal im-
pairment, we have seen how the administration of
piperacillin 4 g every 6 hours did not achieve serum
levels to ensure adequate t ss
(%) > MIC90 values.
This project was supported by the Basque Government (PI-1999-
34). We would also like to thank the Basque Government for the
predoctoral research grant awarded to A. Arzuaga.
REFERENCES
1. Bellomo R, Ronco C. Continuous haemofiltration in the intensive
care unit. Crit Care. 2000;4:339-345.
2. St John RC, Dorinsky PM. Immunologic therapy for ARDS, septic
shock and multiple-organ failure. Chest. 1993;103:932-943.
3. Reetze-Bonorden P, Bohler J, Keller E. Drug dosage in patients dur-
ing continuous renal replacement therapy: pharmacokinetic and
therapeutic considerations. Clin Pharmacokinet. 1993;24:362-379.
4. Sanchez-Izquierdo Riera JA, Alted E, Lozano MJ, Ambros A, Cabal-
lero R. Influence of continuous hemofiltration on the hemodynamics
of trauma patients. Surgery. 1997;122:902-908.
5. Journois D, Israel-Biet D, Pouard P, et al. High-volume, zero-
balanced hemofiltration to reduce delayed inflammatory response to
cardiopulmonary bypass in children. Anesthesiology. 1996;85:965-
976.
6. Oudemans-Van Straaten HM, Bosman RJ, van der Spoel JI,
Zandstra DF. Outcome of critically ill patients treated with intermit-
tent high-volume haemofiltration: a prospective cohort analysis. Int
Care Med. 1999;25:814-821.
7. Honoré PM, Jamez J, Wauthier M, et al. Prospective evaluation of
short-term, high-volume isovolemic hemofiltration on the hemo-
dynamic course and outcome in patients with intractable circulatory
failure resulting from septic shock. Crit Care Med. 2000;28:3581-
3587.
8. Joos B, Schmidli M, Keusch G. Pharmacokinetics of antimicrobial
agents in anuric patients during continuous veno-venous haemo-
filtration. Nephrol Dial Transplant 1996;11:1582-1585.
9. Perry CM, Markham A. Piperacillin/tazobactam: an updated re-
view of its use in the treatment of bacteria infections. Drugs.
1999;57:805-843.
10. Iregui M, Ward S, Sherman G, Fraser VJ, Kollef MH. Clinical im-
portance of the delays in the initiation of appropriate antibiotic treat-
ment for ventilator-associated pneumonia. Chest. 2002;122:262-268.
CRITICAL CARE
11. Karam GH, Niederman MS. How do we achieve adequate therapy
for severe infection? Crit Care Med. 2003;31:648-650.
12. Reed MD, Goldfarb J, Yamashita TS. Single-dose
pharmacokinetics of piperacillin and tazobactam in infants and chil-
dren. Antimicrob Agents Chemother. 1994;38:2817-2826.
13. Rello J, Bodi M, Mariscal D, et al. Microbial testing and outcome of
patients with severe community-acquired pneumonia. Chest.
2003;123:174-180.
14. Sörgel F, Kinzig M. The chemistry, pharmacokinetics and tissue
distribution of piperacillin/tazobactam. J Antimicrob Chemother.
1993;31(suppl A):39-60.
15. Van der Werf T, Mulder POM, Zijlstra JG, Uges DRA, Stegeman
CA. Pharmacokinetics of piperacillin and tazobactam in critically ill
patients with renal failure treated with continuous veno-venous
hemofiltration (CVVH). Int Care Med. 1997;23:873-877.
16. Capellier G, Cornette C, Boillot A, et al. Removal of piperacillin in
critically ill patients undergoing continuous venovenous
hemofiltration. Crit Care Med. 1998;26:88-91.
17. Valtonen M, Tiula E, Takkunen O, Backman JT, Neuvonen PJ.
Elimination of piperacillin/tazobactam combination during continu-
ous venovenous haemofiltration and haemodiafiltration in patients
with acute renal failure. J Antimicrob Chemother. 2001;48:881-885.
18. Mueller SC, Majcher-Peszunska J, Hickstein H, et al. Pharmaco-
kinetics of piperacillin-tazobactam in anuric intensive care patients
during continuous venovenous hemodialysis. Antimicrob Agents
Chemother. 2002;46:1557-1560.
19. Shah VP, Midha KK, Findlay JWA, et al. Bioanalytical method val-
idation: a revisit with a decade of progress. Workshop/conference re-
port. Pharm Res. 2000;17:1551-1557.
20. Food and Drug Administration. Guidance for Industry:
Bioanalytical Methods Validation for Human Studies. Rockville, Md:
Center for Drug Evaluation and Research; 1998.
21. National Committee for Clinical Laboratory Standards. Methods
for Dilution Antimicrobial Susceptibility Tests for Bacteria That
Grow Aerobically; Approved Standard. 5th ed. M7-A5. Wayne, Pa:
NCCLS; 2000.
22. Kollef MH. Antibiotic heterogeneity: should we use it? Crit Care
Med. 2003;31:2074-2076.
23. Golper TA. Drug removal during continuous hemofiltration or
hemodialysis. Contrib Nephrol. 1991;93:110-116.
24. Golper TA, Marx AM. Drug dosing adjustments during continu-
ous renal replacement therapies. Kidney Int. 1998;53:S165-S168.
25. Golper TA. Update on drug sieving coefficients and dosing adjust-
ments during continuous renal replacement therapies. Contrib
Nephrol. 2001;132:349-353.
26. Komuro M, Maeda T, Kakuo H, Matsushita H, Shimada J. Inhibi-
tion of the renal excretion of tazobactam by piperacillin. J Antimicrob
Chemother. 1994;34:555-564.
27. Takagi N, Oda H, Tokita Y, et al. Changes of the serum amikacin
(AMK) level in patients with serious acute renal failure treated by
continuous arteriovenous hemofiltration (CAVH). Artif Organs.
1989;13:238-241.
28. Armendariz E, Chelluri L, Ptachcinski R. Pharmacokinetics of
amikacin during continuous venovenous hemofiltration. Crit Care
Med. 1990;18:675-676.
29. Kihara M, Ikeda Y, Takagi N, et al. Pharmacokinetics of single-
dose intravenous amikacin in critically ill patients undergoing slow
haemodialysis. Int Care Med. 1995;21:348-351.
175
 ARZUAGA ET AL
30. Malone RS, Fish DN, Abraham E, Teitelbaum I. Pharmacokinetics
of cefepime during continuous renal replacement therapy in criti-
cally ill patients. Antimicrob Agents Chemother. 2001;45:3148-3155.
31. Valtonen M, Tiula E, Backman JT, Neuvonen PT. Elimination of
meropenem during continuous veno-venous haemofiltration and
haemodiafiltration in patients with acute renal failure. J Antimicrob
Chemother. 2000;45:701-704.
32. Schetz M, Ferdinande P, Van den Berghe G, Verwaest C, Lauwers
P. Pharmacokinetics of continuous renal replacement therapy. Int
Care Med. 1995;21:612-620.
33. Johnson CA, Halstenson CE, Kelloway JS, et al. Single-dose
pharmacokinetics of piperacillin and tazobactam in patients with re-
nal disease. Clin Pharmacol Ther. 1992;51:32-41.
34. Hanes SD, Wodd GC, Herring V, et al. Intermittent and continuous
infusion for critically ill trauma patients. Am J Surg. 2000;179:436-
440.
35. Reed RL, Ericsson CD, Wu A, Miller-Crotchett P, Fischer RP. The
pharmacokinetics of prophylactic antibiotic in trauma. J Trauma.
1992;32:21-27.
176 • J Clin Pharmacol 2005;45:168-176
36. Townsend PL, Fink MP, Stein KL, Murphy SG. Aminoglycoside
pharmacokinetics: dosage requirements and nephrotoxicity in
trauma patients. Crit Care Med. 1989;17:154-157.
37. Botha FJ, van der Bijl P, Seifart HL, Parkin DP. Fluctuation of the
volume of distribution of amikacin and its effect on once-daily dosage
and clearance in a seriously ill patient. Int Care Med. 1996;22:433-
436.
38. Fernández de Gatta MM, Mendez ME, Romano S, Calvo MV,
Domínguez-Gil A, Lanao JL. Pharmacokinetics of amikacin in inten-
sive care patients. J Clin Pharm Ther. 1996;21:417-421.
39. Kroh UF. Drug administration in critically ill patients with acute
renal failure. New Horizons. 1995;3:748-759.
40. Hyatt JM, McKinnon PS, Zimmer GS, Schentag JJ. The importance
of pharmacokinetic/pharmacodynamic surrogate markers to out-
come: focus on antibacterial agents. Clin Pharmacokinet.
1995;28:143-160.
41. Craig WA, Ebert SC. Continuous infusion of β-lactam antibiotics.
Antimicrob Agents Chemother. 1992;36:2577-2583.
42. Turnidge JD. The pharmacodynamics of β-lactams. Clin Infect Dis.
1998;27:10-22.