Full text of DOI:10.1592/phco.20.6.596.35153

Comparison of Five ␤-Lactam Antibiotics Against
Common Nosocomial Pathogens Using the Time above
MIC at Different Creatinine Clearances — A Comment
Robert P. Rapp, Pharm.D., FCCP, and George A. Davis, Pharm.D.
(Pharmacotherapy 2000;20(6):596–598)
We read with interest the recent article by Kays
entitled “Comparison of five ␤-lactam antibiotics
against common nosocomial pathogens using the
time above MIC at different creatinine
clearances”1 and found several items that
confused us after reading the article.
use an antibiotic when the microbiology
laboratory reports the isolate as “not susceptible,”
we would ask that the author explain what we
believe to be misleading data on the percentage of
time above the MIC (Tables 3–5). In addition, all
of the drugs used in the study have published
clinical data that support their clinical success in
treating a variety of nosocomial infections.
Our first concern is that we found Table 2 to be
very confusing because many of the “weighted
geometric mean MIC90” values are higher than
the National Committee for Clinical Laboratory
Standards (NCCLS) susceptibility breakpoints.
Although the author states this fact in the
discussion section and states that because of this,
combination therapy should be considered when
Pseudomonas aeruginosa is a potential pathogen,
we simply do not understand why the study was
done in this manner because the data is simply
misleading, at best. While this is true for all of
the drugs for P. aeruginosa, it is also true that
many of the other values are also above the
NCCLS breakpoints for susceptibility. Examples
include an MIC90 for cefepime of 19.80 µg/ml
and for piperacillin-tazobactam 99.29 µg/ml (we
assume this refers to the piperacillin
component), whereas the NCCLS susceptibility
breakpoint for cefepime is less than or equal to 8
µg/ml and for piperacillin-tazobactam less than
or equal to 64 µg/ml.
Further, in reading Table 2, we find that one of
the MICs for cefepime, five of the MICs for
ceftazidime, all of the MICs for piperacillin, six of
the MICs for piperacillin-tazobactam, and one of
the MICs for imipenem are above NCCLS
susceptibility breakpoints. Since one does not
From the College of Pharmacy, University of Kentucky,
Lexington, Kentucky (both authors).
Address correspondence to Robert P. Rapp, Pharm.D.,
FCCP, College of Pharmacy, University of Kentucky, 800
Rose Street, Lexington, KY 40536-0082.
Our second concern is that although most of
the dosages selected are reasonable for
nosocomial infections, clearly there are problems
in the package insert for several infections. For
piperacillin-tazobactam, the package insert for
nosocomial pneumonia calls for a dosage of
3.375 g every 4 hours combined with an
aminoglycoside in patients with normal renal
function. An alternative dosage that appears to
be widely used is 4.5 g intravenously every 6
hours. In our opinion, these are the piperacillin-
tazobactam dosages that should have been
studied. Could the author explain how the
dosages were chosen? For imipenem, for severe
life-threatening infections, dosages of up to 3
g/day can be used for patients with normal renal
function. Even for cefepime, the dosage in the
febrile neutropenic patient is 2 g intravenously
every 8 hours. We believe that across the board,
maximum dosages for nosocomial infections
included in the Food and Drug Administration’s
approved package insert should have been used
to evaluate each drug.
We believe that the two points we have
outlined above make the tables misleading for a
number of drugs that have excellent track records
in treating serious infections, and we have
difficulty in understanding what is accomplished
by the Kays article. Would it have been more
reasonable to choose adequate dosages for all

ALTERNATIVE VIEWPOINTS Rapp and Davis
597
drugs evaluated and, even more important, only
use the NCCLS susceptibility breakpoints or
MIC90s for susceptible isolates in the evaluation?
some of the drugs. Why is this confusing? There
are numerous published reports where the MIC90
for a given drug is greater than its susceptibility
3
breakpoint.2,
In fact, whenever the
susceptibility of an organism to a drug is reported
to be less than 90%, the MIC90 for the drug is
greater than its susceptibility breakpoint. In one
particular study,2 for example, the MIC90 for
piperacillin-tazobactam against Enterobacter
species was reported to be greater than 256
µg/ml, while the susceptibility was reported to be
70.8%. The data in Table 2 are not misleading;
they are, quite simply, the geometric mean MIC90
values for the drugs versus the organisms
evaluated. The only thing that may be confusing
is that the values do not follow the traditional 2-
fold dilution scheme because they are mean
MIC90 values. Drs. Rapp and Davis seem to
criticize the fact that only one MIC90 for
cefepime and imipenem is above their respective
susceptibility breakpoints, while five, six, and
eight MIC90 values are above the susceptibility
breakpoints for ceftazidime, piperacillin-
tazobactam, and piperacillin, respectively.
However, these data simply demonstrate the
superior in vitro activity of cefepime and
imipenem for the organisms studied.
Since the study used MIC90 values, Drs. Rapp
and Davis believe that the time above the MIC90
data listed in Tables 3–5 are misleading.
However, a cursory review of the literature will
identify many articles that report the time above
the MIC90 for a ␤-lactam against a given
pathogen.4–9 Once again, it seems prudent to use
an antimicrobial empirically that will provide
coverage for the largest percentage of isolates
likely to be encountered clinically. If this strategy
is employed, cefepime and imipenem will
provide a desirable time above MIC (T>MIC) for
more isolates than ceftazidime, piperacillin-
tazobactam, and piperacillin. These data should
not be misinterpreted to suggest that ceftazidime,
piperacillin-tazobactam, and piperacillin are not
useful agents. Many studies support the clinical
utility of these three agents; however, one must
remember that all isolates recovered from
patients in a clinical study have to be susceptible
to the antibiotic being studied. Patients with
organisms that are resistant to a study drug are
excluded from efficacy analyses, but we cannot
exclude these patients in our clinical practices.
The antibiotic dosages were selected using
recommendations from the package inserts of
each agent. For the initial treatment of patients
with nosocomial pneumonia, the piperacillin-
Reference
1. Kays MB. Comparison of five ␤-lactam antibiotics against
common nosocomial pathogens using the time above MIC at
different creatinine clearances. Pharmacotherapy 1999;19(12):
1392–9.
Author’s Reply
After reading the comments of Drs. Rapp and
Davis, I can certainly appreciate their confusion
regarding my article,1 considering they missed
the main points of the analysis. As they have
requested, I offer the following information in an
attempt to clarify their confusion.
My first suggestion would be to look at the
MIC and pharmacodynamic data from the
standpoint of empiric antimicrobial therapy,
before the results of culture and susceptibility
tests are known. The goals of empiric therapy
are to provide antimicrobial coverage for the
most likely bacterial pathogens causing a given
infection and to cover the largest possible
percentage of the organisms. To wit, the MIC90
has traditionally been used as a measure of in
vitro potency since this is the concentration of
drug that inhibits 90% of isolates in a given
population of organisms. Therefore, my
evaluation was performed using an MIC90 to
determine which ␤-lactam provided desirable
pharmacodynamics for the greatest percentage of
isolates for nosocomial pathogens likely to be
encountered in a clinical situation. Since there
are a wide range of MIC90 values reported for a
given drug-organism combination, an average
(geometric mean) MIC90 was calculated using
published studies, and the MIC90 from each
study was weighted for the number of isolates
tested. While Drs. Rapp and Davis are correct in
that one does not use an antibiotic when the
microbiology laboratory reports the isolate as
“not susceptible,” antimicrobial therapy is
usually started before culture and susceptibility
results are known, and I would prefer to
empirically use an agent that covers 90% or more
of isolates for a given organism, as opposed to
60–70%.
Drs. Rapp and Davis state that they are
confused by the data in Table 2, which lists the
weighted geometric mean MIC90 values for the
five ␤-lactams, because the MIC90 is greater than
the established susceptibility breakpoints for

598
PHARMACOTHERAPY Volume 20, Number 5, 2000
tazobactam package insert suggests a dosage of
3.375 g every 4 hours plus an aminoglycoside;
however, this recommendation is only for
patients with creatinine clearances of 90
ml/minute or greater. The recommended dosage
for piperacillin-tazobactam is 3.375 g every 6
hours for all indications, including nosocomial
pneumonia, in patients with creatinine clearances
of 40–90 ml/minute. Using my data, piperacillin-
tazobactam 3.375 g every 4 hours provides a
desirable T>MIC for Staphylococcus aureus only
at a creatinine clearance of 120 ml/minute, S.
aureus and Escherichia coli at a creatinine
clearance of 100 ml/minute, and S. aureus, E. coli,
and Klebsiella pneumoniae at a creatinine
clearance of 90 ml/minute. Piperacillin-
tazobactam 4.5 g every 6 hours will provide a
desirable T>MIC for S. aureus only at a creatinine
clearance of 90 ml/minute or greater. In
addition, piperacillin-tazobactam 3.375 g every 4
hours costs ~ $81/day at our institution, whereas
cefepime 2 g every 12 hours costs ~ $47/day and
provides coverage for more nosocomial
pathogens. Although imipenem dosages up to 4
g/day have been recommended, the frequency of
side effects, especially seizures, is higher at this
dosage.10, 11 Therefore, imipenem 500 mg every 6
hours was evaluated, since this is a commonly
used dosage regimen.
NCCLS susceptibility breakpoints or MIC90s for
susceptible isolates in the evaluation. My answer
would be absolutely not. Bacteria don’t care what
the NCCLS breakpoint is for an antimicrobial
and, since we encounter resistance bacteria on a
daily basis, to perform an analysis using only
susceptible isolates would indeed be misleading.
Michael B. Kays, Pharm.D.
References
1. Kays MB. Comparison of five ␤-lactam antibiotics against
common nosocomial pathogens using the time above MIC at
different creatinine clearances. Pharmacotherapy
1999;19(12):1392–9.
2. Jones RN, Pfaller MA, Doern GV, et al. Antimicrobial activity
and spectrum investigation of eight broad-spectrum ␤-lactam
drugs: a 1997 surveillance trial in 102 medical centers in the
United States. Diagn Microbiol Infect Dis 1998;30:215–28.
3. Murray PR, Cantrell HF, Lankford RB, et al. Multicenter
evaluation of the in vitro activity of piperacillin-tazobactam
compared with eleven selected ␤-lactam antibiotics and
ciprofloxacin against more than 42,000 aerobic gram-positive
and gram-negative bacteria. Diagn Microbiol Infect Dis
1994;19:111–20.
4. Drusano GL, Ryan PA, Standiford HC, Moody MR, Schimpff
SC. Integration of selected pharmacologic and microbiologic
properties of three new ␤-lactam antibiotics: a hypothesis for
rational comparison. Rev Infect Dis 1984;6:357–63.
5. Nicolau DP, Quintiliani R. Choosing between the new
cephalosporin antibiotics. A pharmacodynamic approach.
PharmacoEconomics 1994;5(suppl 2):34–9.
6. Craig WA. The pharmacokinetics of cefpirome—rationale for a
twelve-hour dosing regimen. Scand J Infect Dis 1993;
91(suppl):33–40.
In closing, Drs. Rapp and Davis ask if it would
have been more reasonable to choose adequate
dosages for all drugs evaluated. I submit that
dosages that are used clinically were evaluated.
Cefepime 2 g every 12 hours provides the same
coverage for P. aeruginosa as ceftazidime 2 g every
8 hours, while cefepime also provides excellent
coverage for Group I ␤-lactamase-producing
organisms. Cefepime also provides coverage for
more organisms than piperacillin-tazobactam,
regardless of the dosage of piperacillin-
tazobactam. Drs. Rapp and Davis also ask if it
would have been more reasonable to only use the
7. Klepser ME, Marangos MN, Patel KB, Nicolau DP, Quintiliani
R, Nightingale CH. Clinical pharmacokinetics of newer
cephalosporins. Clin Pharmacokinet 1995;28:361–84.
8. Craig WA. Antimicrobial resistance issues of the future. Diagn
Microbiol Infect Dis 1996;25:213–7.
9. Craig WA, Andes D. Pharmacokinetics and pharmacodynamics
of antibiotics in otitis media. Pediatr Infect Dis J
1996;15:255–9.
10. Freifeld AG, Walsh T, Marshall D, et al. Monotherapy for fever
and neutropenia in cancer patients: a randomized comparison
of ceftazidime versus imipenem. J Clin Oncol 1995;13:165–76.
11. Winston DJ, Ho WG, Bruckner DA, Champlin RE. Beta-lactam
antibiotic therapy in febrile granulocytopenic patients: a
randomized trial comparing cefoperazone plus piperacillin,
ceftazidime plus piperacillin, and imipenem alone. Ann Intern
Med 1991;115:849–59.