REPORT FROM THE WORKFORCE ON EVIDENCE-BASED MEDICINE
The Society of Thoracic Surgeons Practice
Guideline Series: Antibiotic Prophylaxis in
Cardiac Surgery, Part II: Antibiotic Choice*
Richard Engelman, MD, David Shahian, MD, Richard Shemin, MD,
T. Sloane Guy, MD, Dale Bratzler, DO, MPH, Fred Edwards, MD,
Marshall Jacobs, MD, Hiran Fernando, MD, and Charles Bridges, MD, ScD
Baystate Medical Center, Springfield, Massachusetts; Tufts University School of Medicine, Boston, Massachusetts; Boston Medical
Center, Boston, Massachusetts; University of California, San Francisco, California; Oklahoma Foundation for Medical Quality,
Oklahoma City, Oklahoma; University of Florida, Shands Jacksonville, Jacksonville, Florida; St. Christopher’s Hospital for
Children, Philadelphia, Pennsylvania; and University of Pennsylvania Medical Center, Philadelphia, Pennsylvania
tional publications in so far as they compare different
antibiotic regimens involving comparable duration of
The importance of prophylactic antibiotics for cardiac
multidose antibiotic administration.
surgery has been clearly demonstrated in a number of
The most pertinent report appeared in 2004 and was
placebo-controlled studies completed nearly 30 years ago
a very complete meta-analysis of seven randomized
Surgical site infections (SSIs) and particularly ster-
nal and mediastinal infections have implications for
trials, comparing the incidence of SSIs in patients receiv-
significantly increasing both morbidity and mortality, as
ing either glycopeptide prophylaxis (vancomycin or
well as their associated costs in both man-hours and
teicoplanin) or a
␤-lactam. Five of the seven trials used a
multidose regimen and two invoked, in one of their trial
Part I of this evidence-based guideline series (The
groups, the single preoperative administration of a long-
Society of Thoracic Surgeons Practice Guideline Series: Anti-
acting agent. In both of these latter reports, the single-
biotic Prophylaxis in Cardiac Surgery, Part I: Duration, pub-
dose agent was either less effective or not significantly
lished in the January 2006 issue of the Annals of Thoracic
different from the multidose antibiotic In this
Surgery) recommended that the duration for routine post-
international, multi-institutional meta-analysis involving
operative administration of prophylactic antibiotics be no
␤-lactams were at least as effective as
longer than 48 hours This initial Guideline did not
glycopeptides for the overall prevention of SSIs. How-
define the choice of antibiotic to be recommended, its
ever, only one institution defined their site as having a
dose, or frequency of administration. Those subjects are
high incidence of MRSA (more than 2.5 new cases of
the basis for this report.
MRSA infection or colonization per 100 admissions) and that may limit the degree to which these findings can
be generalized to current practice in which MRSA is
II. Choice of Primary Prophylactic Antibiotic
much more prevalent. Notwithstanding this caveat, it
Cephalosporin or Glycopeptide
appeared that prophylaxis with glycopeptides such as
vancomycin was less effective in preventing infection by
CLASS I RECOMMENDATION. A
␤-lactam antibiotic is indicated
methicillin-sensitive organisms, while such prophylaxis
as a single antibiotic of choice for standard cardiac
was more effective in preventing infection by methicillin-
surgical prophylaxis in populations that do not have a
high incidence of methicillin-resistant Staphylococcus
aureus (MRSA [Level of Evidence A; see Appendix]).
Distinguishing Between Cephalosporins
There are numerous publications concerned with the
CLASS IIA RECOMMENDATION. Based on availability and cost,
optimal prophylactic antibiotic recommended for cardiac
it is reasonable to use cefazolin (a first-generation agent)
surgery, but many of these protocols are comparing not
as the cephalosporin for standard cardiac surgical pro-
only two or more antibiotic regimens but also two differ-
phylaxis in view of the fact that most randomized trials
ent dosing programs, for example, single dose versus
could not discriminate between cephalosporins (Level of
multidose, which was addressed in the previous Guide-
line. This second published Guideline will address addi-
The next issue to be addressed concerns the choice of
␤-lactam, remembering that there are first- through
*For the full text of the STS Guideline on Antibiotic Prophylaxis in
fourth-generation agents presently available, which have
Cardiac Surgery, as well as other titles in the STS Practice Guideline
differing half-lives, pharmacodynamics, and pharmaco-
S e r i e s , v i s i t at the official STS website
kinetics. It can be stated as fact that the later generation
cephalosporins have better gram-negative and less
Address correspondence to Dr Engelman, Baystate Medical Center,
gram-positive coverage. In that our predominant organ-
Division of Cardiac Surgery, 759 Chestnut St, Springfield, MA 01199;
ism for cardiac surgical infections is a Staphylococcus sp,
2007 by The Society of Thoracic Surgeons
Ann Thorac Surg 2007;83:1569 –76 • 0003-4975/07/$32.00
Published by Elsevier Inc
ENGELMAN ET AL
Ann Thorac Surg
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
the earlier generation cephalosporins are likely to be
65% It has been estimated that colonization with
preferred for prophylaxis. In fact, published data would
methicillin-resistant organisms, often asymptomatic, oc-
support that conclusion
curs in 4% to 8% of ICU patients, 0.18% to 7.2% of
In 1987, a randomized trial of more than 1,000 cardiac
inpatients, and 1.3% to 2% of persons in the community
surgical patients was reported comparing multidose ce-
In one urban hospital, the incidence of MRSA
fazolin, a first-generation cephalosporin, with multidose
among newly admitted patients was 7.3%, which is
cefamandole, a second-generation cephalosporin, and
higher than the 1.3% to 5.3% prevalence in previous
found cefamandole to have a lower sternal infection rate
reports This alarming incidence of colonization has
This study, however, introduced a second agent,
led to a strong recommendation for active surveillance at
gentamicin, as an additional single-dose prophylactic
the time of hospital admission At least one third
drug, in half the patients in each cephalosporin group.
of MRSA-colonized patients will have a healthcare-
That led to the comparative analysis being less than clear
related MRSA infection, which is nearly 10 times the risk
cut in defining an optimal cephalosporin. A more defin-
of noncolonized patients In a study by Lin and
itive randomized double-blind study comparing individ-
associates at a hospital with a high incidence of
ual cephalosporins in 1,641 patients from Johns Hopkins
MRSA, 65% of post-sternotomy staphylococcal infections
Hospital between 1987 and 1990 was reported in 1993
were due to methicillin-resistant organisms
The incidence of all surgical site infections was 8.4%
Some studies suggest that patients with post-
with cefamandole prophylaxis, 8.4% cefazolin, and 9.0%
sternotomy MRSA/MRCNS infections have a less favor-
with cefuroxime (clearly not significant). The relative
able prognosis compared with those having methicillin-
incidence comparing cephalosporins and differentiating
sensitive (MSSA) organisms. For example, in the study of
between deep and superficial infection was also not
Mekontso-Dessap and colleagues overall mortality
significantly different between the groups (specifically,
was 53.3% for MRSA post-sternotomy infections versus
deep sternal infection
ϭ 0.6% cefamandole, 1% cefazolin,
19.2% for MSSA infections, with corresponding 3-year
and 1.5% cefuroxime). A 1992 meta-analysis cited in
actuarial survival rates of 26% versus 79%. Methicillin-
the Hopkins report includes some with inherent flaws
resistant S aureus was the only independent predictor of
but still supports the conclusion that there is no cepha-
overall mortality, and MRSA infections had a higher
losporin regimen that is clearly superior in affecting a
incidence of mediastinitis-related death and treatment
lower infection rate.
failure compared with MSSA. In a study of SSIs composed
of largely cardiac and orthopedic procedures, Engemann
and associates found a mortality rate of 20.7% for MRSA
III. Issues Surrounding Staphylococcal Infection
versus 6.7% for MSSA, and most deaths in the cardiac
Reasons for Concern in Cardiac Surgical Patients
group were due to post-sternotomy mediastinitis. The costs
directly attributable to methicillin resistance were $13,901
Surgical site infections of the sternal wound and under-
per case of staphylococcal infection.
lying mediastinum occur in 0.4% to 4% of cardiac surgical
procedures, with more than 50% due to S aureus or
Potential (Nonallergic) Indications for Primary or
coagulase-negative S epidermidis These infections
Adjuvant Glycopeptide (Vancomycin) Prophylaxis
have profound short- and long-term implications. In-
CLASS IIB RECOMMENDATION. In the setting of either a pre-
hospital mortality rates of 10% to more than 20% have
sumed or known staphylococcal colonization, the institu-
commonly been reported, and a 10-year follow-up study
tional presence of a “high incidence” of MRSA, patients
of such patients by the Northern New England Cardio-
susceptible to colonization (hospitalized longer than 3
vascular Disease Study Group demonstrated a marked
days, transfer from other inpatient facility, already re-
negative impact not only on acute but also on long-term
ceiving antibiotics), or an operation for a patient having
survival Hollenbeak and colleagues found a
prosthetic valve or vascular graft insertion, it would be
1-year mortality rate of 22% for coronary artery bypass
reasonable to combine the
␤-lactam (cefazolin) with a
graft surgery (CABG) patients with deep chest surgical
glycopeptide (vancomycin) for prophylaxis, with the re-
site infections versus 0.6% for uninfected patients (p
striction to limit vancomycin to only one or two doses
0.0001). Deep chest infection resulted in 20 additional
(Level of Evidence C).
hospital days per patient (p
ϭ 0.0001) and added an
The progressive emergence of methicillin-resistant
average of $18,938 in hospital costs. Patients who died as
staphylococcal organisms within hospitals and the com-
a result of their infection incurred average costs that were
munity, as well as the possibly more serious course of
$60,547 more than infected patients who lived.
such infections in the cardiac surgery patient, has led
The choice of a prophylactic antibiotic has become
some to recommend more aggressive use of prophylactic
increasingly controversial with the emergence of MRSA
vancomycin, even for patients with no history of penicil-
and methicillin-resistant coagulase-negative Staphylococcus
lin or cephalosporin allergy For example, it is argued
(MRCNS). According to the National Nosocomial Infec-
that patients having surgery in institutions with a high
tion Surveillance System Report, the median percentage
incidence of methicillin resistance would be better
of MRSA isolates from intensive care unit (ICU) and
served by receiving vancomycin, although it is unclear as
non-ICU patients in hospitals surveyed exceeded 40%,
to what constitutes a high incidence Other poten-
and the median percentage of MRCNS isolates exceeded
tial candidates for vancomycin prophylaxis might include
Ann Thorac Surg
ENGELMAN ET AL
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
patients who are at higher risk for preoperative MRSA
appear most reasonable to employ a cephalosporin as the
colonization, patients at higher risk for post-sternotomy
primary prophylactic agent for the usual 24 to 48 hours,
infection in general, and patients with specific risk factors
and only to use vancomycin selectively as an adjuvant
for MRSA post-sternotomy infection Active surveil-
agent, typically a single dose preoperatively (together
lance of admitted patients for staphylococcal colonization
with the first dose of cephalosporin) with at most one
is desirable but results for cardiac surgery
additional dose in valve or vascular implant patients or in
patients would generally not be available at the time of
all patients in highly selected environments (eg, where
surgery except in those institutions where rapid polymer-
MRSA colonization is likely or documented or where
ase chain reaction (PCR) testing is available. Finally, it
there is a high prevalence of MRSA isolates from infec-
has been suggested, but not generally accepted, that
tions). This should provide a reasonable compromise
because of the devastating consequences of prosthetic
between the goal of providing the broadest spectrum
valve or vascular graft infection with methicillin-resistant
prophylaxis at the time when it is likely to be most
organisms, these patients should also routinely receive
effective, and the competing desire to restrict usage of
vancomycin in order to minimize the emergence of
There are observational and randomized trial data
supporting the use of vancomycin prophylaxis for
Vancomycin as the Sole Prophylactic Antibiotic
cardiac surgery, as well as the results of a sophisticated
decision analytic model Using the best available
CLASS IIB RECOMMENDATION. Because vancomycin is an
clinical and microbiological data from the literature,
agent that has no effect on gram-negative flora, its use-
Zanetti and colleagues estimated that routine
fulness as an exclusive agent in cardiac surgical prophy-
vancomycin use in a cohort of 10,000 CABG patients
laxis is not recommended (Level of Evidence C).
would result in 29 fewer deep chest infections, 58 fewer
DISCUSSION. For situations in which vancomycin is be-
superficial infections, 3 fewer deaths, lower direct medi-
lieved to be indicated as prophylaxis for cardiac surgery,
cal costs over 3 months, and a net $1,170,000 cost saving
␤-lactam allergy, should it be used as a
compared with routine cefazolin. Sensitivity analysis
single agent or combined with another antimicrobial?
indicated that cephazolin was more effective or less costly
Overall, vancomycin has a narrower antimicrobial spec-
only when MRSA represented fewer than 3% of all
trum, inferior tissue and bone penetration, less desirable
staphylococcal isolates in a hospital, which would be
pharmacokinetics, and slower bactericidal killing com-
unusual in contemporary practice. Based on 366,000
pared with cephalosporins and the
CABG procedures annually in the United States, this
incidence of SSI due to methicillin-sensitive organisms
model predicts that vancomycin use would result in 110
has been higher when only vancomycin has been em-
fewer deaths, prevent 3,184 SSIs, and potentially save $43
ployed for prophylaxis Additionally, since some
hospitals report both deep surgical site infections and
One of the most serious objections to increased use of
blood stream infections after cardiac surgery from gram-
vancomycin prophylaxis is concern about the emergence
negative organisms it is recommended that an
of resistant strains of Staphylococcus and Enterococcus
aminoglycoside be added for one preoperative and at
organisms This consideration has prompted the
most one additional postoperative dose to act as a spe-
publication of restrictive guidelines for the use of vanco-
cific gram-negative agent when vancomycin is indicated
mycin or teicoplanin (both glycopeptides), which include
to be the primary prophylactic agent.
a specific recommendation by the CDC against the rou-
Mupirocin for Preoperative Therapy to Eliminate
tine use of vancomycin for prophylaxis However, it
Staphylococcal Nasal Colonization
should be noted that antibiotic resistance may also de-
␤-lactam antimicrobials. Furthermore, the
CLASS I RECOMMENDATION. Routine mupirocin administra-
duration of vancomycin administration as a primary or
tion is recommended for all patients undergoing cardiac
adjuvant prophylactic agent, as opposed to its use for
surgical procedures in the absence of a documented
established post-sternotomy infections, must also be con-
negative testing for staphylococcal colonization (Level of
sidered. In terms of the emergence of drug-resistant
organisms, which is worse— using short-duration pro-
DISCUSSION. Mupirocin is a patient self-administered top-
phylactic vancomycin in a larger number of patients,
ical antibiotic that is highly effective in eradicating nasal
possibly preventing some clinical infections due to me-
S aureus, including methicillin-resistant strains of Staphy-
thicillin-resistant organisms; or using a cephalosporin
loccocus. It is a naturally occurring antibiotic produced by
after which a serious SSI is more likely to involve MRSA
a fermentation of Pseudomonas bacteria mixed in a non-
or MRCNS, thus committing such patients to weeks or
irritating paraffin composition. Its specific mechanism of
months of continuous vancomycin therapy
action is to bind to isoleucyl-transfer RNA synthetase
This is a central question that as yet has not been
and disrupt cell function It is reportedly more than
resolved and would require research not likely to be
90% effective in eradicating nasal colonization of Staphy-
performed. Thus, this particular question cannot be ad-
lococcus for as long as 1 year Short-term therapy (a
dressed by randomized trials.
5-day course) has been shown to be highly effective
Unless there is demonstrated penicillin or
Correlation of nasal or hand colonization and infection in
allergy (see Section V, “Allergy to Penicillin”), it would
the same patient by the same phage type of Staphyloccocus
ENGELMAN ET AL
Ann Thorac Surg
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
has been shown to be near 90% Recent reports of
2. When the surgical incision remains open in the
both randomized and nonrandomized trials in cardiac
operating room, to patients with normal renal func-
surgical patients, one a meta-analysis, supports its rou-
tion, a second dose of 1 g should be administered
tine use in prophylaxis
every 3 to 4 hours. If it is apparent that cardiopul-
Resistance to mupirocin ointment has become a con-
monary bypass will be discontinued within 4 hours,
cern for infectious disease specialists, but such resistance
it is appropriate to delay until perfusion is complete
is largely found after prolonged treatment periods when
to maximize effective blood levels (Class I, Level of
used to treat either large open wounds or dermatitis.
There have been no reports of high-level drug resistant
3. In patients for whom vancomycin is an appropriate
strains developing after a short course of treatment such
prophylactic antibiotic for cardiac surgery, a dose of
as proposed for preoperative prophylaxis despite 4 years
1 to 1.5 g or a weight-adjusted dose of 15 mg/kg
of surveillance in one hospital using this approach rou-
administered intravenously slowly over 1 hour,
tinely in both orthopedic and vascular surgery In
with completion within 1 hour of the skin incision,
fact, many, if not most, cardiac surgical programs have
is recommended (Class I, Level of Evidence A). A
instituted a routine protocol for intranasal mupirocin
second dose of vancomycin of 7.5 mg/kg may be
beginning at least the day before operation (sooner, if
considered during cardiopulmonary bypass, al-
elective operation) and continuing for 2 to 5 days after
though its usefulness is not well established (Class
surgery. Recently, a PCR rapid analysis for Staphyloccocus
IIb, Level of Evidence C).
sp has become available in some hospitals, with addi-
4. For patients who receive an aminoglycoside (usu-
tional institutions gaining access to the technology on a
ally gentamicin, 4 mg/kg) in addition to vancomy-
regular basis. A report has just been published for a
cin before cardiac surgery, the initial dose should
PCR-based mupirocin study performed at the Cleveland
be administered within 1 hour of the skin incision
Clinic. In this study, screening for nasal carriage of S
(Class I, Level of Evidence C). Redosing an amino-
aureus (both MRSA and MSSA) was routinely performed
glycoside during cardiopulmonary bypass is not
before cardiac surgery. There were 6,334 patients
indicated and may be harmful (Class III, Level of
screened over 21 months, and 1,342 were found to have
colonization (21%), which is the identical incidence re-
There is a considerable body of evidence supporting
ported in a second study as well The administration
the need for the timely administration of preoperative
of mupirocin was reserved for these colonized patients,
antibiotics, which means administration within 1 hour of
and while the mupirocin use in the cardiac surgical
the skin incision These data accrue from numer-
population declined significantly (by nearly 80%), there
ous animal and clinical studies and are broadly applica-
was no demonstrable difference between carriers and
ble to all procedures for which prophylactic antibiotics
noncarriers in the overall incidence of infection or in the
are administered In spite of the relative paucity
incidence of infection caused by S aureus. It was con-
of controlled randomized or large-scale retrospective
cluded that the effect of mupirocin on colonized patients
resulted in appropriately reducing the Staphyloccal in-
studies to address this issue specifically in cardiac sur-
fection incidence to nullify the influence of colonization.
gery, the timing of the administration of the prophylactic
Because, inherently, one would not recommend use of
antibiotic is quite important to the cardiac surgical com-
any agent that is not useful for treatment, limiting mupi-
munity. Cardiopulmonary bypass (CPB) is a technique
rocin prophylaxis to colonized patients would appear to
that is nearly exclusively used by cardiac surgeons, and it
be a sensible approach. However, access to the PCR test
has profound effects on the volume of distribution, and
is required. Because mupirocin is self-administered, the
elimination kinetics of a variety of drugs including the
patient must be informed about the need for the treat-
commonly used prophylactic antibiotics such as cepha-
ment and the technique of insertion. In the absence of
losporins, vancomycin, and aminoglycosides
access to PCR testing, routine prophylaxis with mupiro-
Certain drugs, including opiates, nitrates and vancomy-
cin is recommended.
cin also have been shown to be sequestered in the
components of the heart lung machine, decreasing bio-
logical availability both during and after the completion
IV. Guidelines for Appropriate Dosing of
of CPB Therefore, appropriate perioperative
dosing of antibiotics during cardiac surgery presents
unique challenges, particularly since tissue levels, specif-
ically in bone and sternal fat, are likely more relevant
1. In patients for whom cefazolin is the appropriate
than the more commonly measured serum concentra-
prophylactic antibiotic for cardiac surgery, admin-
tions. In fact, cefazolin tissue concentrations during sur-
istration within 60 minutes of the skin incision is
gery are clearly correlated with body weight (increased
indicated (Class I, Level of Evidence A). The pre-
body mass index correlates with decreased tissue levels)
operative prophylactic dose of cefazolin for a pa-
such that therapeutic tissue levels may not be achieved in
tient of greater than 60 kg body weight is recom-
the morbidly obese patient even with 2 g administered
mended to be 2 g (Class I, Level of Evidence B).
Ann Thorac Surg
ENGELMAN ET AL
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
Several studies have investigated intraoperative van-
V. Guidelines for Prophylactic Antibiotics in
comycin cephalosporin and aminoglyco-
side pharmacokinetics. After a single preopera-
Allergy to Penicillin
tive dose of vancomycin, typically administered over 1
hour, immediately before the skin incision serum con-
centrations averaged 18 to 66 mg/L after a dose of either
1. In patients with a history of an immunoglobulin-E
1 g or a weight-adjusted dose of 15 mg/kg All of
(IgE)–mediated reaction to penicillin or cephalo-
these studies also documented an 11% to 41% abrupt
sporin (anaphylaxis, hives, or angioedema), vanco-
decrease in serum vancomycin concentration after the
mycin should be given preoperatively and for no
initiation of cardiopulmonary bypass due primarily to
more than 48 hours. Alternatively, skin testing may
dilution in direct proportion to the pump prime volume.
be performed in these patients and, if negative, a
During cardiopulmonary bypass, there is a progressive
cephalosporin regimen administered (Class I, Level
decline in serum concentrations due to a combination of
of Evidence A).
renal clearance and sequestration in the heart lung
2. For patients with a history of a non-IgE mediated
machine After a single preoperative dose, the
reaction to penicillin (such as a simple rash) or an
serum level in each of the reported studies remains
unclear history either vancomycin or a cephalosporin
above the minimal inhibitory concentration (MIC) for
is recommended for prophylaxis with the under-
90% of both methicillin-sensitive and methicillin-
standing that these patients have a low incidence of
resistant S aureus (1 mg/L) and coagulase-negative Staph-
significant allergic reactions to cephalosporins (Class
ylococcus (2 mg/L) throughout the procedure with an
I, Level of Evidence B).
average bypass time of approximately 1 to 2 hours
3. The addition of an aminoglycoside or other gram-
There is incomplete recovery of serum levels
negative bacterial coverage to a vancomycin antibi-
after bypass, however, owing to vancomycin sequestra-
otic regimen may be reasonable, but its efficacy is not
tion in the heart-lung machine, alterations in protein
well established (Class IIb, Level of Evidence C).
binding, and persistent changes in the volume of distri-
In patients with a history suggestive of an IgE-
bution after bypass. Similarly, studies have shown that
mediated reaction to penicillin (anaphylaxis, hives, or
aminoglycosides first- and second-generation ceph-
angioedema), indiscriminate use of a cephalosporin for
alosporins have a similar (as much as 50%)
surgical prophylaxis in cardiovascular surgery is not
reduction in serum concentration after the initiation of
advised Early studies established a cross-reactivity
rate between penicillin and cephalosporins at approxi-
As a result of the reduction in the levels of cefazolin
mately 20% More recent data including those ceph-
and vancomycin immediately after and during CPB, two
alosporins in current clinical use suggests a cross-
studies evaluated the efficacy of administering a second
reactivity rate of less than 2%
dose of cefazolin or a second dose of vancomycin after
As many as 20% of the general population are labeled
the initiation of cardiopulmonary bypass Both
penicillin-allergic. Fewer than half of these will have a
studies found that with the second dosing regimen, the
history suggesting an IgE-mediated reaction to penicillin.
serum levels were above the MIC for both S aureus and
Of these, fewer than 20% will have a positive penicillin
coagulase-negative Staphylococci throughout the proce-
skin-test Those patients with nonsuggestive or un-
dure. The two-dose regimen of vancomycin resulted in
known histories have a penicillin skin-test positivity rate
higher serum levels but no significant difference in ster-
of less than 2% Among all patients labeled penicil-
nal bone, fat, myocardial, or pericardial tissue levels
lin-allergic, the frequency of serious reactions to cepha-
losporin administration is less than 1%
It is now firmly established with good documentation
With regard to choice of alternative prophylaxis in the
from both clinical and experimental studies that read-
presence of allergy, vancomycin appears to be best owing to
ministration of a prophylactic antibiotic during surgery
its gram-positive coverage and, particularly, coverage of
should be within two half-lives of the antibiotic, exclusive
methicillin-resistant S aureus. There are concerns over lack
of any influence of the effects of cardiopulmonary bypass
of gram-negative coverage with vancomycin relative to
Cefazolin has a half-life of approximately 1.8
cephalosporins. For this reason, an aminoglycoside, usually
hours, and therefore it is recommended that there should
gentamicin, should be added. It must be recognized, how-
be additional dosing during surgery every 3 to 4 hours
ever, that gentamicin is associated with nephrotoxicity and
when an operation is proceeding with an open wound
ototoxicity, and excretion is delayed after cardiopulmonary
beyond that period. The major consideration for defining
bypass Therefore, a single dose, or at most two doses,
the appropriate pharmacodynamics of antimicrobials is
of no more than 4 mg/kg is recommended There is no
to maintain the serum level of any antibiotic used above
study directly comparing vancomycin and vancomycin plus
the MIC for the infecting pathogens, presumed in cardiac
an aminoglycoside. A single study from 1987 compared
surgery to be Staphylococcus sp, while the operative
gentamicin plus a
␤-lactam with the latter alone and found
wound remains open. This typically dictates readminis-
no benefit to the combination therapy, compounded by the
tration approximately every two serum half-lives of each
appearance of resistant gram-negative organisms only in
antibiotic considered appropriate
patients receiving gentamicin
ENGELMAN ET AL
Ann Thorac Surg
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
The use of vancomycin as an alternative to cephalo-
concerns over antibiotic penetration into this area and
sporins is not entirely benign. Vancomycin commonly
resultant infection with S aureus. As in the case of
causes histamine release and cutaneous reactions when
intravenous vancomycin, there is concern over the pro-
administered too rapidly. It is also associated with nephro-
motion of resistant organisms.
toxicity when used in combination with other nephrotoxins
A review of the literature on the use of topical vanco-
and can rarely case anaphylaxis In one study
mycin revealed a single randomized controlled trial from
116 patients (106 adults and 10 children) undergoing cardiac
1989 comparing patients treated with either vancomycin-
surgery procedures were given vancomycin. Thirty-one
thrombin–powdered gelatin paste (223 patients) versus
patients (27%) had an adverse reaction including hypoten-
treatment with thrombin-powdered gelatin paste alone
sion (25%). Maculopapular edema was seen in 6% and was
(193 patients) This group from the University of
associated with hypotension (Red Man’s syndrome) in 5
Massachusetts reported a sternal infection rate of 0.45%
patients and bronchospasm in 1 patient.
(1 patient) in the treatment group and 3.5% (7 patients) in
One group used mathematical modeling to predict the
the control group (p
ϭ 0.02). Multivariate testing was
most cost-effective strategy for antimicrobial prophylaxis
performed. The use of topical vancomycin and shorter
of cardiovascular surgery patients labeled penicillin-
operative times predicted reduced infection rates. An-
allergic The strategy of giving vancomycin to all
other reported study of 4 patients in whom serum levels
patients labeled with a penicillin allergy was found to be
were measured after topical application without systemic
the most expensive but was associated with the lowest
administration found levels of vancomycin in 1 patient as
rate of serious allergic reaction. Giving cefazolin to all
high as 4.4 mg/L 3 to 4 hours after topical application of
such patients was the cheapest, but it was associated with
1 g vancomycin powder to the sternum, which is signif-
the highest rate of allergic reaction. While giving vanco-
icantly lower than would be seen with systemic admin-
mycin to patients with positive skin tests improved cost
istration (18 to 66 mg/L)
effectiveness, it was thought to be impractical on a
The topical application of gentamicin has also been
routine preoperative basis. Therefore, this group adopted
studied. In a randomized trial of 2,000 cardiac surgery
a policy of using vancomycin in place of cephalosporins
patients, Friberg and coworkers compared prophy-
in patients with a history suggesting a previous IgE-
laxis with intravenous isoxazolyl-penicillin alone versus
mediated reaction to penicillin.
this drug plus topical application of collagen-gentamicin
An aminoglycoside is often added to vancomycin for
sponges at sternotomy closure. The topical antibiotic
cardiac surgery in penicillin-allergic patients owing to its
group had an incidence of wound infection at 4.3% and
enhanced gram-negative coverage as well as its coverage
the control group, at 9% (p
Ͻ 0.001). The same author
of methacillin-sensitive S aureus. However, in a 1987
examined serum versus local wound fluid concentrations
study from St. Thomas Hospital in Nashville, the only
of gentamicin in 101 patients receiving topical gentamicin
patients with resistant gram-negative sternal infections
during cardiac surgery and found extremely high concen-
were those who received gentamicin along with either
trations (median 304 mg/L) in local wound fluid but very
cefazolin or cefamandole
low serum concentrations (peak median 2.05 mg/L)
Eklund and associates recently reported a random-
Specific Issues Regarding Gram-Negative Infections
ized controlled trial comparing topical gentamicin-collagen
ϭ 272) and no topical antibiotics (n
during coronary artery bypass graft surgery. Both study
1. For institutions with an outbreak of gram-negative
groups received standard intravenous prophylaxis consist-
deep wound infections due to a specific pathogen, it
ing of cefuroxime (85%) or cefuroxime and vancomycin
is reasonable to employ a first-generation cephalo-
(14%). The sternal wound infection rate was 4.0% in the
sporin for routine prophylaxis (
Յ48 hours) supple-
topical gentamicin group and 5.9% in the control group.
mented with an appropriate antibiotic to which the
Deep mediastinal infections were seen in 1.1% of the topical
offending organisms are sensitive (Class IIa, Level
antibiotic group and 1.9% in the control group. The authors
of Evidence C).
concluded that a slight reduction in infection was seen but
2. In patients with renal dysfunction requiring gram-
that the population was too small to draw a definitive
negative prophylaxis to supplement a cephalosporin
conclusion. While the use of topical antibiotics is controver-
or vancomycin as the primary antibiotic, it is reason-
sial and they are not used by most cardiac surgeons, the
able to use either one dose of an aminoglycoside or an
existing studies demonstrate a reduction in the wound
antibiotic such as levofloxacin with a low incidence of
infection rate. More study is warranted before topical anti-
renal toxicity (Class IIa, Level of Evidence C).
biotics can be recommended as standard prophylaxis.
Topical Application of Antibiotics
CLASS IIB RECOMMENDATION. Topical antibiotics may be con-
sidered for antibiotic prophylaxis in cardiac surgery
The primary prophylactic antibiotic for adult cardiac
(Level of Evidence B).
surgery is recommended to be a first-generation cepha-
Some cardiac surgeons have used topical antibiotics,
losporin, which is usually cefazolin. The most frequent
usually vancomycin or gentamicin, applied to the cut
organism cultured in cardiac SSI is Staphylococcus, and
sternal edges. There is some appeal to this concept given
colonization is considered the major factor in wound
Ann Thorac Surg
ENGELMAN ET AL
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
contamination. For this reason, until rapid screening tests
prophylaxis in cardiac operations. J Thorac Cardiovasc Surg
for S aureus colonization are widely available, mupirocin
is recommended as a routine prophylactic measure. In
15. Kreter B, Woods M. Antibiotic prophylaxis for cardiothoracic
operations. J Thorac Cardiovasc Surg 1992;104:590 –9.
patients considered at high risk for a staphylococcal
16. Massias L, Dubois C, DeLentdecker P, Brodaty O, Fischler
infection, vancomycin (one preoperative with or without
M. Penetration of vancomycin in uninfected sternal bone.
one additional dose) may be reasonable as an adjuvant
Antimicrob Agent Chemother 1992;36:2539 – 41.
agent to the cephalosporin. For patients who are consid-
17. Abboud CS, Wey SB, Baltar VT. Risk factors for mediastinitis
after cardiac surgery. Ann Thorac Surg 2004;77:676 – 83.
␤-lactam or penicillin allergic, vancomycin is rec-
18. Dodds A, Carroll DN, Engemann JJ, et al. Risk factors for
ommended as the primary prophylactic antibiotic with
postoperative mediastinitis due to methicillin-resistant
additional gram-negative coverage. Topical antibiotics
Staphylococcus aureas. Clin Infect Dis 2004;38:1555– 60.
may be useful, but the evidence to support their efficacy
19. Lin CH, Hsu RB, Chang SC, Lin FY, Chu SH. Poststernotomy
mediastinitis due to methicillin-resistant Staphylococcus au-
is limited to three randomized trials.
reus endemic in hospital. Clin Infect Dis 2003;37:679 – 84.
20. Martorell C, Engelman R, Brown CA. Surgical site infections
in cardiac surgery: an 11-year perspective. Am J Infect
Control 2004;32:63– 8.
1. Austin TW, Coles JC, Burnett R, Goldbach M. Aortocoronary
21. Upton A, Roberts SA, Milsom P, Morris AJ. Staphylococcal
bypass procedures and sternotomy infections: a study of
post-sternotomy mediastinitis: five year audit. Aust NZ
antistaphylococcal prophylaxis. Can J Surg 1980;23:483–5.
J Surg 2005;75:198 –203.
22. Mekontso-Dessap A, Honore S, Kirsch M, et al. Blood
2. Fekety FR, Cluff LE, Sabiston DC, Seidl LG, Smith JW,
neutrophil bactericidal activity against methicillin-resistant
Thoburn R. A study of antibiotic prophylaxis in cardiac
and methicillin-sensitive Staphlococcus aureus during car-
surgery. J Thorac Cardiovasc Surg 1969;57:757– 63.
diac surgery. Shock 2005;24:109 –13.
3. Fong IW, Baker CB, McKee DC. The value of prophylactic
23. Braxton JH, Marrin CA, McGrath PD, et al. 10-year follow-up
antibiotics in aorta-coronary bypass operations. J Thorac
of patients with and without mediastinitis. Semin Thorac
Cardiovasc Surg 1979;78:908 –13.
Cardiovasc Surg 2004;16:70 – 6.
4. Goodman JS, Schaffner W, Collins HA, Battersby EJ, Koenig
24. Hollenbeak CS, Murphy DM, Koenig S, Woodward RS,
MG. Infection after cardiovascular surgery. N Engl J Med
Dunagan WC, Fraser VJ. The clinical and economic impact
of deep chest surgical site infections following coronary
5. Engemann JJ, Carmeli Y, Cosgrove SE, et al. Adverse clinical
artery bypass graft surgery. Chest 2000;118:397– 402.
and economic outcomes attributable to methicillin resis-
25. National Nosocomial Infections Surveillance (NNIS) System
tance among patients with Staphylococcus aureus surgical
Report. Data summary from January 1992–June 2004, issued
site infection. Clin Infect Dis 2003;36:592– 8.
October 2004. Am J Infect Control 2004;32:470 – 85.
6. Zanetti G, Goldie SJ, Platt R. Clinical consequences and cost
26. Davis KA, Stewart JJ, Crouch HK, Florez CE, Hospenthal DR.
of limiting use of vancomycin for perioperative prophylaxis:
Methicillin-resistant Staphylococcus aureus (MRSA) nares
example of coronary artery bypass surgery. Emerg Infect Dis
colonization at hospital admission and its effects on subse-
quent MRSA infection. Clin Infect Dis 2004;39:776 – 82.
7. Edwards FH, Engelman RM, Houck P, Shahian D, Bridges C.
27. Hidron AI, Kourbatova EV, Halvosa JS, et al. Risk factors for
The Society of Thoracic Surgeons practice guideline series:
colonization with methicillin-resistant Staphylococcus au-
antibiotic prophylaxis in cardiac surgery, part I: duration.
reus (MRSA) in patients admitted to an urban hospital:
Ann Thorac Surg 2006;81:397– 404.
emergence of community-associated MRSA nasal carriage.
8. Bolon MK, Morlote M, Weber SG, Koplan B, Carmeli Y,
Clin Infect Dis 2005;41:159 – 66.
Wright SB. Glycopeptides are no more effective than beta-
28. Karchmer TB. Prevention of health care-associated methicil-
lactam agents for prevention of surgical site infection after
lin-resistant Staphylococcus aureus infections: adapting to a
cardiac surgery: a meta analysis. Clin Infect Dis 2004;38:
changing epidemiology. Clin Infect Dis 2005;41:167–9.
29. Chenoweth CE, DePestel DD, Prager RL. Are cephalospo-
9. Saginur R, Croteau D, Bergeron MG. Comparative efficacy of
rins adequate for antimicrobial prophylaxis for cardiac sur-
teicoplanin and cefazolin for cardiac operation prophylaxis
gery involving implants? Clin Infect Dis 2005;41:122–3.
in 3027 patients. The ESPRIT Group. J Thorac Cardiovasc
30. Zanetti G, Platt R. Antibiotic prophylaxis for cardiac surgery:
Surg 2000;120:1120 –30.
does the past predict the future? Clin Infect Dis 2004;38:
10. Wilson AP, Treasure T, Gruneberg RN, Sturridge MF, Ross
1364 – 6.
DN. Antibiotic prophylaxis in cardiac surgery: a prospective
31. Crabtree TD, Codd JE, Fraser VJ, Bailey MS, Olsen MA,
comparison of two dosage regimens of teicoplanin with a
Damiano RJ Jr. Multivariate analysis of risk factors for deep
combination of flucloxacillin and tobramycin. J Antimicrob
and superficial sternal infection after coronary artery bypass
grafting at a tertiary care medical center. Semin Thorac
11. Finkelstein R, Rabino G, Mashiah T, et al. Vancomycin
Cardiovasc Surg 2004;16:53– 61.
versus cefazolin prophylaxis for cardiac surgery in the set-
32. Muto C, Jernigan JA, Ostrowsky BE, et al. SHEA guideline
ting of a high prevalence of methicillin-resistant staphylo-
for preventing nosocomial transmission of multidrug-
coccal infections. J Thorac Cardiovasc Surg 2002;123:326 –32.
resistant strains of Staphylococcus aureus and enterococcus.
12. Maki DG, Bohn MJ, Stolz SM, Kroncke GM, Acher CW,
Infect Control Hosp Epidemiol 2003;24:362– 86.
Myerowitz DP. Comparative study of cefazolin, cefaman-
33. Spelman D, Harrington G, Russo P, Wesselingh S. Clinical,
dole, and vancomycin for surgical prophylaxis in cardiac
microbiological, and economic benefit of a change in antibi-
and vascular operations. A double-blind randomized trial.
otic prophylaxis for cardiac surgery. Infect Control Hosp
J Thorac Cardiovasc Surg 1992;104:1423–34.
Epidemiol 2002;23:402– 4.
13. Kaiser AB, Petracek MR, Lea JW IV, et al. Efficacy of cefazolin,
34. Martone WJ. Spread of vancomycin-resistant enterococci:
cefamandole, and gentamicin as prophylactic agents in cardiac
why did it happen in the United States? Infect Control Hosp
surgery. Results of a prospective, randomized, double-blind
Epidemiol 1998;19:539 – 45.
trial in 1030 patients. Ann Surg 1987;206:791–7.
35. Smith TL, Pearson ML, Wilcox KR, Cruz C, et al. Emergence
14. Townsend TR, Reitz BA, Bilker WB, Bartlett JG. Clinical trial
of vancomycin resistance in Staphylococcus aureus. N Engl
of cefamandole, cefazolin, and cefuroxime for antibiotic
J Med 1999;340:493–501.
ENGELMAN ET AL
Ann Thorac Surg
ANTIBIOTIC PROPHYLAXIS IN CARDIAC SURGERY
36. Centers for Diseases Control and Prevention. Recommenda-
An analysis of peri- and postoperative serum cefuroxime
tions for preventing the spread of vancomycin resistance.
and vancomycin levels. J Hosp Infect 1997;37:237– 47.
Recommendations of the Hospital Infection Control Prac-
59. Stanbridge TN, Greenall DJ. Netilmicin prophylaxis in open-
tices Advisory Committee. MMWR Morb Mortal Wkly Rep
heart surgery. J Antimicrob Chemother 1984;13:59 – 66.
60. Zanetti G, Giardina R, Platt R. Intraoperative redosing of
37. McNamara DR, Steckelberg JM. Vancomycin. J Am Acad
cefazolin and risk for surgical site infection in cardiac sur-
Orthop Surg 2005;13:89 –92.
gery. Emerg Infect Dis 2001;7:828 –31.
38. Ryan T, Mc Carthy JF, Rady MY, et al. Early bloodstream
61. Craig WA. Basic pharmacodynamics of antibacterials with
infection after cardiopulmonary bypass: frequency rate, risk
clinical applications to the use of beta-lactams, glycopeptides,
factors, and implications. Crit Care Med 1997;25:2009 –14.
and linezolid. Infect Dis Clin North Am 2003;17:479–501.
39. Parenti MA, Hatfield SM, Leyden JJ. Mupirocin: a topical
62. Phillips E, Louie M, Knowles SR, et al. Cost-effectiveness
antibiotic with a unique structure and mechanism of action.
analysis of six strategies for cardiovascular surgery prophy-
Clin Pharm 1987;6:761–70.
laxis in patients labeled penicillin allergic. Am J Health Syst
40. Ward A, Campoli-Richards DM. Mupirocin. A review of its
Pharm 2000;57:339 – 45.
antibacterial activity, pharmacokinetic properties and ther-
63. Kishiyahma J, Adelman D. The cross-reactivity and immu-
apeutic use. Drugs 1986;32:425– 44.
␤-lactam antibiotics. Drug Safety 1994;10:318 –27.
41. Wenzel RP, Perl TM. The significance of nasal carriage of
64. Shepherd G. Clinical experience using only PrePen and
penicillin G to detect penicillin allergy in hospitalized adults.
Staphylococcus aureus and the incidence of postoperative
J Allergy Clin Immunol 1997;99(Suppl):134.
wound infection. J Hosp Infect 1995;31:13–24.
65. Lin R. A perspective on penicillin allergy. Arch Intern Med
42. Tulloch LG. Nasal carriage in staphylococcal skin infections.
1992;152:930 – 4.
Br Med J 1954;2:912–3.
66. Gadde J, Spence M, Wheeler B, Adkinson NF Jr. Clinical
43. Kallen AJ, Wilson CT, Larson RJ. Perioperative intranasal
experience with penicillin skin testing in a large inner-city
mupirocin for the prevention of surgical-site infections:
STD clinic. JAMA 1993;270:2456 – 63.
systematic review of the literature and meta-analysis. Infect
67. Lewis DR, Longman RJ, Wisheart JD, Spencer RC, Brown
Control Hosp Epidemiol 2005;26:916 –22.
NM. The pharmacokinetics of a single dose of gentamicin (4
44. Kluytmans JA, Wertheim HF. Nasal carriage for Staphylo-
mg/kg) as prophylaxis in cardiac surgery requiring cardio-
coccus aureus and prevention of nosocomial infections.
pulmonary bypass. Cardiovasc Surg 1999;7:398 – 401.
Infection 2005;33:3– 8.
68. Polk R. Anaphylactoid reaction to glycopeptide antibiotics.
45. Nicholson MR, Huesman LA. Controlling the usage of
J Antimicrob Chemother 1991;27:17–29.
intranasal mupirocin does impact the rate of Staphylococcus
69. Valero R, Gomar C, Fita G, et al. Adverse reactions to
aureus deep sternal wound infections in cardiac surgery
vancomycin prophylaxis in cardiac surgery. J Cardiothorac
patients. Am J Infect Control 2006;34:44 – 8.
Vasc Anesth 1991;5:574 – 6.
46. Fawley W, Parnell P, Hall J, Wilcox MH. Surveillance for
70. Vander Salm TJ, Okike ON, Pasque MK, et al. Reduction of
mupirocin resistance following introduction of routine peri-
sternal infection by application of topical vancomycin. J Tho-
operative prophylaxis with nasal mupirocin. J Hosp Infect
rac Cardiovasc Surg 1989;98:618 –22.
71. Oakley RE, Nimer KA, Bukhari E. Is the use of topical
47. Shrestha NK, Banbury MK, Weber M, et al. Safety of targeted
vancomycin to prevent mediastinitis after cardiac surgery
perioperative mupirocin treatment for preventing infections
justified? J Thorac Cardiovasc Surg 2000;119:190 –1.
after cardiac surgery. Ann Thorac Surg 2006;81:2183–8.
72. Friberg O, Svedjeholm R, Soderquist B, et al. Local gentamicin
48. Bratzler D, Houck P. Antimicrobial prophylaxis for surgery:
reduces sternal wound infections after cardiac surgery: a ran-
an advisory statement from the National Surgical Infection
domized controlled trial. Ann Thorac Surg 2005;79:153–61.
Prevention Project. Clin Infect Dis 2004;38:1706 –15.
73. Eklund AM, Valtonen M, Werkkala KA. Prophylaxis of
49. Classen DC, Evans RS, Pestotnik SL. The timing of prophy-
sternal wound infections with gentamicin-collagen implant:
lactic administration of antibiotics and the risk of surgical-
randomized controlled study in cardiac surgery. J Hosp
wound infection. N Engl J Med 1992;326:281– 6.
Infect 2005;59:108 –12.
50. DiPiro JT, Vallner JJ, Bowden TA Jr, Clark BA, Sisley JF.
Intraoperative serum and tissue activity of cefazolin and
cefoxitin. Arch Surg 1985;120:829 –32.
51. Galandiuk S, Polk HC Jr, Jagelman DG, Fazio VW. Re-
emphasis of priorities in surgical antibiotic prophylaxis.
Classification of Recommendations
Surg Gynecol Obstet 1989;169:219 –22.
52. Buylaert WA, Herregods LI, Mortier EP. Cardiopulmonary
Class I. Conditions for which there is evidence or general
bypass and the pharmacokinetics of drugs. An update. Clin
agreement, or both, that a given procedure is useful and
Pharmacokinet 1989;17:10 –26.
53. Fellinger EK, Leavitt BJ, Hebert JC. Serum levels of prophy-
Class II. Conditions for which there is conflicting evidence or a
lactic cefazolin during cardiopulmonary bypass surgery.
divergence of opinion, or both, about the usefulness/efficacy of a
Ann Thorac Surg 2002;74:1187–90.
54. Klamerus KJ, Rodvold KA, Silverman NA, Levitsky S. Effect
of cardiopulmonary bypass on vancomycin and netilmicin
Class IIa. Weight of evidence favors usefulness/efficacy.
disposition. Antimicrob Agents Chemother 1988;32:631–5.
Class IIb. Usefulness/efficacy is less well established by evi-
55. Krivoy N, Yanovsky B, Kophit A, et al. Vancomycin seques-
tration during cardiopulmonary bypass surgery. J Infect
Class III. Conditions for which there is evidence or general
agreement, or both, that the procedure is not useful/effective.
56. Miglioli PA, Merlo F, Grabocka E, Padrini R. Effects of
cardio-pulmonary bypass on vancomycin plasma concentra-
tion decay. Pharm Res 1998;38:275– 8.
Level of Evidence
57. Edmiston CE, Krepel C, Kelly H, et al. Perioperative antibi-
otic prophylaxis in the gastric bypass patient: do we achieve
Level A. Data derived from multiple randomized clinical trials.
therapeutic levels? Surgery 2004;136:738 – 47.
Level B. Data derived from a single randomized trial or from
58. Vuarisalo S, Pokela R, Syrjala H. Is single-dose antibiotic
prophylaxis sufficient for coronary artery bypass surgery?
Level C. Consensus expert opinion.
Excellence in Prevention – descriptions of the prevention programs and strategies with the greatest evidence of success Name of Program/Strategy: Peacemakers Report Contents Implementation considerations (if available) Cost effectiveness report (Washington State Institute of Public Policy – if available) Washington State results (from Performance Based Prevention System (P
nausea and vomitingRudolph M. Navari, MD, PhDIndiana University School of Medicine–South Bend, South Bend, INNausea and vomiting associated with cancer chemotherapy are experienced by 70%–80% of patients receiving chemotherapy and can result in signiﬁcant morbidity. Chemotherapy-induced nausea and vomiting (CINV) adversely affects patient quality of life, often leading to poor compliance