Non-compliance with recommendations for the practice of antibiotic
prophylaxis and risk of surgical site infection: results of a multilevel
analysis from the INCISO Surveillance Network
Katiuska Miliani1, Franc ¸ois L’He ´riteau1and Pascal Astagneau1,2* on behalf of the INCISO
Network Study Group†
1Regional Coordinating Centre for Nosocomial Infection Control (C-CLIN Paris Nord), Paris, France;
2Department of Public Health, Pierre et Marie Curie University School of Medicine, Paris, France
Received 15 July 2009; returned 10 September 2009; revised 14 September 2009; accepted 20 September 2009
Objectives: The aim of this study was to determine which surgical antibiotic prophylaxis (SAP) prac-
tices alter surgical site infection (SSI) risk.
Methods: Data were collected during a 7 year surveillance period (2001–07) from volunteer surgery
wards participating in the INCISO Surveillance Network in Northern France. Main SAP practices, i.e.
antibiotic choice, timing of first dose and total SAP duration, were evaluated and compliance checked
based on French recommendations. The study focused on selected procedures in digestive, orthopae-
dic, gynaecological and cardiovascular surgery, for which standard SAP is recommended. Multilevel
logistic regression analysis (a two-level random effect model) was carried out to identify SAP-, patient-
and procedure-specific factors associated with SSI.
Results: Of 8029 patients who underwent the selected surgeries, 91.3% received SAP and 2.5% devel-
oped SSI. Among those receiving SAP, 83.3% received appropriate antibiotic agents and 76.6% had an
optimal timing of administration. SAP duration was considered to be appropriate in 35.0%, too long
(SAP unnecessarily prolonged) in 45.2% and too short (lack of intra-operative redosing when
recommended) in 19.8%. In the multivariate analysis, a too-short SAP duration remained the only
inappropriate practice associated with higher SSI risk (odds ratio51.8, 95% confidence interval: 1.14–
2.81), after adjustment for surgery procedure group, the National Nosocomial Infections Surveillance
System risk index, age and infection risk variability among hospitals. No significant relationships were
observed between SSI and the other SAP parameters.
Conclusions: A too-short SAP duration was the most important SAP malpractice associated with an
increased risk of SSI. Information directed at practitioners should be reinforced based on standard
Keywords: surgical antibiotic prophylaxis, practice assessment, multilevel logistic regression analysis, France
Surgical site infections (SSIs) are the most common type of
nosocomial infection acquired by surgical patients.1–3SSIs are
the third cause of nosocomial infections in France, accounting
for up to 14% of all nosocomial infections in 2006.4These infec-
tions are a substantial burden in terms of healthcare cost and
post-operative morbidity and mortality.5,6Since 1960, substantial
research has demonstrated the effectiveness of antimicrobial pro-
phylaxis to reduce the risk of post-operative infections for many
procedures in parallel with hygiene control measures.7–9
Likewise, numerous guidelines for optimal use of prophylactic
antibiotics have been published in recent years.10–12Henceforth,
antibiotic prophylaxis is indicated clearly for most clean-
contaminated surgical wounds and some clean surgical wounds,
such as vascular prostheses and orthopaedic implants.
*Corresponding author. Institut Biome ´dical des Cordeliers, C-CLIN Paris Nord, 15–21 rue de l’Ecole de Me ´decine, 75006 Paris, France.
Tel: þ33-1-40-46-42-00; Fax: þ33-1-40-46-42-17; E-mail: email@example.com
†Members are listed in the Acknowledgements section.
Journal of Antimicrobial Chemotherapy (2009) 64, 1307–1315
Advance Access publication 16 October 2009
# The Author 2009. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved.
For Permissions, please e-mail: firstname.lastname@example.org
Consensus guidelines10–12state that adequate prophylaxis
requires: (i) antimicrobial agents with targeted spectra of activity
against organisms likely to be encountered in the particular sur-
gical field; (ii) timely pre-operative administration of the anti-
biotics; (iii) bactericidal concentrations of the drugs in serum
and tissues during the entire time that the incision is open; and
(iv) a duration of up to 24 h following surgery. However, despite
the existence of these guidelines, several studies have shown
that compliance with these practices is not optimal.13–15The
widespread abuse of prophylactic antibiotics is pervasive,
harmful for patients and increases financial costs.16Prolonging
antibiotic prophylaxis beyond 24 h not only fails to improve SSI
rates, but could also favour the emergence of multidrug-resistant
pathogens such as methicillin-resistant Staphylococcus aureus
and Clostridium difficile.17–19Furthermore, inappropriate timing
of administration leads to decreased efficacy.20
To address the issue of whether inappropriate surgical anti-
biotic prophylaxis (SAP) increases the risk of SSI, we conducted
a study based on data collected by the surveillance network of
SSI incidence in Northern France (INCISO Network). The aim
of this study was to determine which SAP practice parameters
(antibiotic choice, timing of first dose and total duration of pro-
phylaxis) alter SSI risk in patients for whom SAP was indicated
and compliance assessed, and to estimate the individual contri-
bution of each SAP parameter to the SSI risk.
The INCISO surveillance system is a regional network of participat-
ing volunteer surgery wards in Northern France that was set up in
1997. This network, coordinated by the regional centre for nosoco-
mial infection control in Northern France (C-CLIN Paris Nord), has
been widely described elsewhere.5,21Since 1998, a yearly 3 month
survey has been proposed to all volunteer surgical wards between
January and March according to a protocol established by the coor-
dinating centre. Each participating ward had to enrol the first 200
consecutive patients who underwent a surgical procedure for a
30 day post-operative follow-up. For each patient, the surgical team,
helped by infection control practitioners, completed a standardized
form with two sections for pre-operative and post-operative
follow-up data. The form was stapled to the medical record at
inclusion and a copy was sent to the infection control practitioner
for follow-up. The overall items collected in the form were SSI risk
factors such as age, pre-operative hospital stay, time of surgical
incision and skin closure, type of surgery, American Society of
Altemeier classification,23emergency conditions, whether video
surgery or multiple procedures were performed, length of post-
operative stay and SSI occurrence during in- and out-hospital stay.
All patients should be followed up to 30 days after operation for
SSI identification. If they were discharged before day 30, they were
given an appointment by the surgeon for a post-discharge visit. The
surgeon was responsible for diagnosing each SSI according to stan-
dard criteria,24whether at in- or out-patient hospital visit. Infection
control practitioners had to collect clinical and bacteriological infor-
mation for SSI identification and to check whether the surgeon took
into account available information for infection diagnosis. All data
were transmitted to the regional coordinating centre at the end of
eachsurveillance period for
Nosocomial Infections Surveillance (NNIS) System risk index was
calculated based on the 75th percentile of operation length described
previously.25A management report was edited yearly displaying the
overall results of the data network to all participant wards.
Since 2000, an annual evaluation of SAP practices has been pro-
posed to all volunteer surgical wards participating in the INCISO
Network. Thus, a standardized form for SAP data was completed
for patients aged ?15 years. Data collected in the SAP form pro-
vided information on whether antibiotics were used for treatment or
SAP, as well as a detailed description of each SAP course related to
route, dosage, administration time of all doses, time of discontinu-
ation, names of antibiotics administered and total duration of
Our study was limited to groups of surgical procedures for which
standard SAP is recommended. Thus, data focused on selected
digestive, orthopaedic, gynaecological and cardiovascular surgery,
collected during a 7 year surveillance period (2001–07). The diges-
tive group included appendectomy (excluding perforated appendix),
small bowel surgery and colorectal surgery, the orthopaedic group
included knee and hip prosthesis surgery, the gynaecology group
included hysterectomy (vaginal or abdominal) and breast surgery
(excluding lumpectomy) and the cardiovascular group included vas-
cular surgery (lower limb arteries, thoracic and abdominal aorta)
and cardiac surgery (open/closed heart surgery, heart valve surgery
and coronary artery bypass surgery). Since wound contamination
class 3 (contaminated) and 4 (dirty) interventions often required
antibiotic therapy, only surgical procedures that belonged to wound
contamination class 1 (clean) or 2 (clean-contaminated) were
Compliance with the main SAP practices, i.e. antibiotic choice,
timing of first dose and total duration of prophylaxis, was assessed
based on the standard recommendations of the French Society of
Anesthesia and Intensive Care (SFAR) 1999 update,11with minor
modifications limiting the duration of SAP in orthopaedic and cardi-
ovascular surgery up to 24 h.12SAP compliance was defined for
each SAP practice parameter and they were considered indepen-
dently of each other. Thus, for antibiotic choice, compliance was
stated if the antimicrobial agents used were those recommended by
the SFAR. For the timing of first dose, compliance was stated if
injection occurred within 60 min before incision (90 min for vanco-
mycin). Regarding total duration of prophylaxis, compliance was
stated in the following circumstances: (i) if a single dose was given
for patients undergoing digestive and gynaecological surgery or
when vancomycin was used; (ii) if additional intra-operative doses
were re-injected when intervention lasted more than twice the half-
life of the antibiotic and until the wound was closed, even if non-
compliant antibiotics were used (Table 1); and (iii) if post-operative
doses, not exceeding 24 h, were re-injected only for the cardiovascu-
lar and orthopaedic surgery groups.
Analysis focused only on patients for whom SAP was documen-
ted. Patients who did not receive SAP were excluded. All variables
were considered as categorical: age was divided into 10 year inter-
vals from 40 to 60 years; duration of surgery was divided into two
classes according to the 75th percentile of surgical duration (.p75
yes or no); and pre-operative hospital stay was divided into two
classes consistent with pre-operative stay .48 h yes or no. For each
SAP practice parameter variable, the reference class was that defined
as compliant. Non-compliant SAP was categorized for the timing of
first dose (too-early administration: .60 min before incision or
.90 min if vancomycin was used; too-late administration: after
incision) and for the total duration of prophylaxis (too-short
duration: lack of intra-operative redosing despite the fact that it
is recommended; too-long duration: unnecessary re-injections or
post-operative doses .24 h). Univariate analysis was performed
Miliani et al.
using the Pearson x2test to screen potential risk factors for SSI.
Multivariate analysis was achieved by multilevel logistic regression
analyses, with surgical procedures clustered within hospitals, using
the GLLAMM software that runs in the statistical package STATA
(Stata Corporation, College Station, TX, USA). A two-level random
effect model was estimated.26Odds ratios (ORs) and their 95% con-
fidence intervals (95% CIs) are reported, as well as hospital-level
(level 2) variances and standard errors, and goodness-of-fit statistics
based on 22 log likelihood ratio test (22LL). The random effect
model applied in this analysis allowed adjustment of the risk esti-
mates for random variation between hospitals. This model is basi-
cally a logistic regression model, supplemented with an extra term
in the equation for the random effects associated with differences in
infection risk between hospitals. Ordinary logistic regression models
do not take into account inter-hospital variability and thus they
might overestimate the contribution of patient- and prophylaxis-
related factors. Variables for which P?0.10 in univariate analysis
were included in the multivariate model. The final model was com-
puted with a manual stepwise backwards elimination. Variables
remained in the multivariate model if the likelihood ratio test was
significant (P,0.05). All computations were performed using
STATA statistical software, release 10.1 (Stata Corporation).
A total of 8029 surgical patients were chosen from 302 wards
belonging to 149 hospitals according to the criteria defined
above. Overall, 7330 (91.3%) patients were finally analysed, and
699 (8.7%) were excluded from the study as they did not receive
SAP but had an indication. As shown in Table 2, the distribution
by procedure groups was as follows: digestive, 23.6%; gynaecol-
ogy, 13.0%; orthopaedic, 48.5%; and cardiovascular, 14.9%. The
study population was predominantly female (60.8%), with a
median age of 64.2 years, ASA score ,3 (74.1%), operated
within the standard operative time (80.0%), under elective surgery
(91.9%) and pre-operative hospital stay ?48 h (91.1%). A total of
184 SSIs were diagnosed, 105 (57.1%) during the hospital stay.
The median time between surgery and SSI diagnosis was 11 days
(interquartile range: 7–17 days). SSI was deep or organ/space in
34.8% and required re-operation in 25.6%. The overall SSI inci-
dence rate was 2.5% and varied significantly according to surgical
procedure, age, NNIS System risk index, ASA score, wound con-
tamination class, duration of surgery, multiple procedures per-
formed and a long pre-operative hospital stay (Table 2).
The compliance with SAP practices varied according to pro-
cedure group (Table 3). Overall, compliance rates were rather
high for antibiotic choice (83.3%) and timing of first dose
(76.6%), but poorer for total duration of SAP (35%), since 45%
had a too-long SAP duration. The choice and timing of SAP
administration were highly compliant (.80%) in the orthopae-
dic group as compared with the other procedure groups.
Conversely, the compliance rate with total SAP duration was
poorer in this group (11.2%). A total of 1417 (19.4%) pro-
cedures received SAP in complete compliance, 4074 (55.8%)
had two SAP practice parameters compliant, 1678 (23.0%) had
one SAP parameter compliant and 131 (1.8%) procedures
received SAP in complete non-compliance. Regarding the anti-
biotic choice, the types of antibiotic that were non-compliant
varied among the four surgical groups. For digestive procedures,
the most frequent non-compliant antibiotics were first-generation
Table 1. Indications of intra-operative redosing according to the antibiotic used for SAP by procedure
Surgical procedure group
Recommended intra-operative redosinga
Digestive: appendectomy (excluding perforated appendix), small bowel surgery and
if procedure takes .3 h
if procedure takes .2 h
if procedure takes .2 h
Orthopaedic: knee and hip prosthesis surgery
if procedure takes .4 h
cefamandole or cefuroxime
if procedure takes .2 h
Gynaecology: vaginal or abdominal hysterectomy and breast surgery (excluding
if procedure takes .4 h
yes, for clindamycin if procedure takes .4 h
Cardiovascular: vascular surgery (lower limb arteries, thoracic and abdominal aorta)
and cardiac surgery (open/closed heart surgery, heart valve surgery andcoronary artery bypass surgery)
repeated doses at 4 h intervals until the wound is closed
cefamandole or cefuroxime
repeated doses every 2 h until the wound is closed
aBased on recommendations of the SFAR.11In general, SFAR recommendations stated that additional antimicrobial doses should be provided intra-operatively if intervention lasted more than twice the
half-life of the antibiotic.
bWhen allergic to b-lactam antibiotics.
Surgical prophylaxis and infection risk
Table 2. Incidence of SSI, risk factors and SAP, N¼7330 surgical procedures
Variable No. of proceduresNo. of SSIs (%)Pa
Surgical procedure (N¼7330)
Female gender (N¼7329)
Age by class in years (N¼7330)
NNIS System risk indexb(N¼7164)
Wound contamination class (N¼7330)
Duration of surgery .p75 (N¼7176)
Emergency circumstance (N¼7316)
Multiple procedures performed (N¼7304)
Pre-operative hospital stay .48 h (N¼7326)
Antibiotic choice for SAPd(N¼7330)
Timing of first dosee(N¼7287)
Miliani et al.
cephalosporins (8.4%), third-generation cephalosporins (5.2%),
aminopenicillins without b-lactam inhibitors (4.0%), nitroimida-
zoles without gentamicin (3.6%) and other non-specified anti-
biotics (4.5%). For gynaecology procedures, the most frequent
non-compliant antibiotics were cephamycins (12.0%), amoxicil-
(3.8%), aminopenicillins without b-lactam inhibitors (2.4%) and
other non-specified antibiotics (3.0%). For cardiovascular pro-
cedures, thenon-compliant antibiotics
resistant penicillins (4.1%) and cefoxitin (1.0%) and for ortho-
paedic procedures the non-compliant antibiotics were mainly
represented bycefalotin (2.1%),
penicillinase-resistant penicillins (0.8%). Table 4 shows the anti-
biotics most commonly prescribed for SAP (compliant and non-
compliant) by surgical procedure.
In univariate analysis, antibiotic choice for SAP and total
SAP duration were significantly associated with higher SSI
cefoxitin (1.1%) and
incidence as well as the usual SSI risk factors, such as NNIS
System risk index, pre-operative hospital stay, age and multiple
procedures performed. Conversely, the timing of the first dose
was not associated with SSI incidence (Table 2). In the multi-
variate analysis (Table 5), a too-short SAP duration remained
the only inappropriate practice associated with higher SSI risk
(OR¼1.8, 95% CI: 1.14–2.81), after adjustment for surgery
procedure group, NNIS System risk index and age. Table S1
[available as Supplementary data at JAC Online (http://jac.ox-
fordjournals.org/)] shows the output estimates of the final model.
Based on a large survey including most important surgical pro-
cedures requiring SAP, we found that a too-short SAP duration,
i.e. lack of intra-operative redosing despite the fact that it is
Table 3. Compliance of SAP according to standard French recommendations by surgical group, N¼7330 surgery procedures
(N¼1730) n (%)
(N¼956) n (%)
(N¼3555) n (%)
(N¼1089) n (%)
(N¼7330) n (%)
Timing of first doseb
Total duration of prophylaxisc
aNon-compliance was stated if the antibiotics used for prophylaxis were those not recommended in the SFAR recommendations.11
bNon-compliance was categorized as: too-early administration, .60 min before incision or .90 min if vancomycin was used; and too-late administration,
cNon-compliance was categorized as: too-short duration, lack of intra-operative redosing despite the fact that it is recommended; and too-long duration,
unnecessary re-injections or post-operative doses .24 h.
Table 2. Continued
VariableNo. of proceduresNo. of SSIs (%)Pa
Total duration of prophylaxisf(N¼6821)
bIncludes the following elements: ASA score, wound contamination class and duration of surgery.
c1, healthy; 2, mild systemic disorder; ?3, severe systemic disorder.
dNon-compliance was stated if the antibiotics used for prophylaxis were those not recommended in the SFAR recommendations.11
eNon-compliance was categorized as: too-early administration, .60 min before incision or .90 min if vancomycin was used; and too-late administration,
fNon-compliance was categorized as: too-short duration, lack of intra-operative redosing despite the fact that it is recommended; and too-long duration,
unnecessary re-injections or post-operative doses .24 h.
Surgical prophylaxis and infection risk
recommended, was the most important SAP malpractice associ-
ated with an increased SSI risk. Other studies had shown the sig-
nificance of intra-operative SAP redosing for preventing SSI
risk.27–29A retrospective study, restricted to cardiac surgery,
showed the benefit of intra-operative cefazolin redosing on SSI
risk reduction.27Two randomized trials demonstrated the neces-
sity for intra-operative cefazolin or cefoxitin redosing in various
surgerical procedures that lasted .3 h.28,29These studies,
however, were designed to prove an effect of SAP redosing on
SSI risk. In contrast, we conducted a comprehensive study
aiming to investigate which SAP practice parameters had a
major role in SSI risk. SAP duration, more precisely a lack of
intra-operative redosing, appeared to be the most important SAP
malpractice parameter associated with an increased risk of SSI.
Furthermore, our results were obtained using a multivariate
regression analysis permitting adjustment for various risk
factors, such as surgical procedures and patient conditions. Also,
inter-hospital variability was measured using a two-level random
intercept model allowing adjustment of the risk estimates for
random variation among hospitals. van Kasteren et al.,30in a
large survey on hip prosthesis with a study design similar to
ours, did not find a significant relationship between SAP dur-
ation and SSI risk after comparable adjustment for confounders.
However, this study did not assess compliance for each SAP
parameter, in particular whether intra-operative redosing was
recommended or not. To our knowledge, no other comparable
published study has reported that poor compliance with this
specific SAP recommendation was the most important malprac-
tice associated with increased SSI rate.
Strikingly, lack of intra-operative redosing, despite the fact
that it is recommended, was one of the most frequent malprac-
tices, observed in almost 20%. This rate varied according to
surgery; the highest non-compliance rate was for orthopaedic
surgery procedures and the lowest for gynaecological pro-
cedures. According to recommendations, when the duration of a
procedure is expected to exceed the time for which the active
concentration of an antibiotic should be maintained, repeat
intra-operative doses are indicated.11,12The recommended inter-
val time for redosing (twice the half-life of the antibiotic) varies
from 2 h to .4 h according to the antibiotic used. Although
evidence-based guidelines exist to support the appropriateness of
SAP practices, there is a substantial gap between these
guidelines and their implementation in daily practice.13–15The
awareness and knowledge of surgeons and anaesthesiologists
could be improved using information and training based on stan-
dard recommendations as well as practice assessment study and
morbidity–mortality reviews with root cause analysis of patient
cases presenting with severe SSI.
Almost half of the SAP durations were too long, mainly in
orthopaedic surgery where this proportion attained nearly 60%.
This poor compliance could be partly explained by a discre-
pancy between French recommendations, which allow SAP dur-
ation to be extended up to 48 h for these particular procedures,11
and the international standard used as the reference in our study,
which is limited to 24 h.12Indeed, most studies reported that a
24 h regimen is often standard for orthopaedic or cardiovascular
surgery.7,14,31Nonetheless, we did not find any significant
relationship between too-long SAP duration and SSI risk.
Indeed, as expected, too-long SAP duration fails to improve SSI
risk. Besides, a too-long SAP duration carries a risk of
multidrug-resistant bacteria selection. This result corroborates
the issue that duration of SAP should be shortened to a
maximum of 24 h after surgery.17–19More recently, studies
comparing single-dose prophylaxis to multiple-dose prophylaxis
did not show benefits of post-operative doses.32–35
Conversely to other studies,20,36,37we did not find that com-
pliance with antibiotic choice or timing of first dose adminis-
tration influenced SSI risk. This result could be explained partly
by the fact that a small proportion of practices dismissed rec-
ommendations according to these two criteria. Regarding anti-
biotic choice for SAP, potential confounders that could influence
SSI risk, such as NNIS System risk index or other peri-operative
conditions, were taken into account in the final analysis. For
example, the type of surgery was an important confounder
factor. Indeed, antibiotic choice was significantly more often
compliant in orthopaedic surgery (93.8%) than in non-
orthopaedic surgery (73.4%), while the SSI risk in the latter was
significantly higher (OR¼7.6, 95% CI: 4.8–12.0). Thus, the
crude OR of antibiotic choice (non-compliant versus compliant)
was significantly associated with SSI risk (OR¼2.1, 95% CI:
1.5–3.0; P,0.0001), whereas OR adjusted only for the type of
surgery (non-orthopaedic versus orthopaedic) was not (adjusted
OR 1.3, 95% CI: 0.95–1.8; P¼0.10). In contrast, timing of the
first dose was not associated with SSI risk, neither in univariate
Table 4. Antibiotics most commonly prescribed for SAP, N¼7330 surgery procedures
(N¼1730) n (%)
(N¼956) n (%)
(N¼3555) n (%)
(N¼1089) n (%)
In bold, antibiotics recommended by the SFAR.11
aVancomycin is an antibiotic recommended in cases of allergy to b-lactam antibiotics for orthopaedic and cardiovascular procedures.
Miliani et al.
Table 5. Risk factors for SSI in the multivariate logistic regression analysis, N¼6751
Random intercept modela
full modelfinal model
OR (95% CI)P OR (95% CI)P
Non-compliant antibiotic choice1.2 (0.83–1.82) 0.30——
Total duration of prophylaxisb
NNIS System risk indexc
Age .40 years
Multiples procedures performed
Pre-operative hospital stay .48 h
Model validation results
full model final model
Number of surgical procedures
Number of hospitals
Number of SSIs
aOutput model obtained by retaining the significant variables (P,0.05).
bNon-compliance was categorized as: too-short duration, lack of intra-operative redosing despite the fact that it is recommended; and too-long duration, unnecessary re-injections or post-operative doses
cIncludes the following elements: ASA score, wound contamination class and duration of surgery.
Surgical prophylaxis and infection risk
nor in multivariate analysis. One could argue that a more
detailed time stratification would have provided different results.
However, when we stratified timing in intervals of ?30 min,
31–60 min, 61–90 min and .90 min before incision and two
classes for administration after incision (?30 or .30 min), no
significant association was observed with SSI risk (P¼0.82,
Pearson x2test). Whatever the methodological issues, we
support the idea that timeliness of antibiotic administration is
crucial. Several prior studies demonstrated that administering
SAP .2 h before surgery resulted in lower tissue levels at the
time of incision,36and administering SAP after the incision is
closed fails to prevent contamination of the surgical wound.20,37
However, in contrast with Classen et al.,36more recent evidence
recommends a first administration of SAP within 60 min before
Our study has methodological strengths that should be
emphasized. First, we analysed the correlation between the
appropriateness of SAP practices and SSI risk using data from a
large multicentre cohort of patients with all data collected pro-
spectively by a standardized protocol that included post-
discharge surveillance (up to 30 days after surgery). In addition,
we used standard definitions for surveillance of SSI and standard
evaluations of SAP practices. Furthermore, we analysed the
effect of various parameters of SAP and several potential
patient/procedure-specific risk factors on the risk of SSI by
using a multilevel multivariate analysis that takes into account
Even though data were prospectively recorded from a large
sample of participating surgery wards, our study had some limit-
ations inherent to the design and the study population. First of
all, there was a potential risk of selection bias, given that partici-
pation in the INCISO Surveillance Network is voluntary.
However, since participation was anonymous, the motivation to
deliver false or invalid data to the surveillance system is unli-
kely. In addition, we could analyse only the covariates that were
reported within the network and some confounding variables
that could have an impact on the SSI risk could have been
missed (glucose control, hypothermia, blood loss, experience of
the surgeon, pre-operative skin disinfection etc.). Also, because
follow-up was limited to 30 days, some SSIs were probably
missed. Although there was no modification in the SFAR rec-
ommendations11during the study period, SAP practices might
have changed over time. Indeed, SSI rates decreased during the
surveillance period,21probably owing to the feedback provided
to the wards.
In conclusion, this study emphasizes the importance of
intra-operative redosing according to antibiotic half-life and
procedure duration, without undermining the role of timely
administration described previously.20,36,37Information aimed at
practitioners should be reinforced in order to improve compli-
ance with SAP practices and eventually decrease SSI rates.
We thank the members of the INCISO Network Study Group
[H. Blanchard (Paris), A. Chalfine (Paris), G. Cherbonnel
(Evreux), Y. Costa (Lagny/Marne-la-Valle ´e), J. M. Germain
(Rouen), B. Grandbastien (Lille), J. L. Jost (Paris), Z. Kadi
(Amiens), M. Kitzis (Clichy), V. Merle (Rouen), L. Richard
(Paris) and J. C. Se ´guier (Saint-Germain-en-Laye)] and the
surgery wards for their participation in the INCISO Surveillance
These data have been generated as part of the routine work of
our INCISO Surveillance Network supported by the regional
None to declare.
Table S1 is available as Supplementary data at JAC Online
1. Gravel D, Taylor G, Ofner M et al. Point prevalence survey for
healthcare-associated infections within Canadian adult acute-care
hospitals. J Hosp Infect 2007; 66: 243–8.
2. Pellizzer G, Mantoan P, Timillero L et al. Prevalence and risk
factors for nosocomial infections in hospitals of the Veneto region,
north-eastern Italy. Infection 2008; 36: 112–9.
3. Reilly J, Stewart S, Allardice GA et al. Results from the Scottish
National HAI Prevalence Survey. J Hosp Infect 2008; 69: 62–8.
4. Thiolet J-M, Lacave ´ L, Jarno P et al. Pre ´valence des infections
nosocomiales, France, 2006. [Prevalence of nosocomial infections,
France, 2006.] Bulletin e ´pide ´miologique hebdomadaire (BEH) 2007;
51–52: 429–32. Available at: http://www.invs.sante.fr/beh/2007/51_52/
beh_51_52_2007.pdf (9 September 2008, date last accessed).
5. Astagneau P, Rioux C, Golliot F et al. Morbidity and mortality
associated with surgical site infections: results from the 1997–1999
INCISO surveillance. J Hosp Infect 2001; 48: 267–74.
6. Coello R, Charlett A, Wilson J et al. Adverse impact of surgical
site infections in English hospitals. J Hosp Infect 2005; 60: 93–103.
7. Dellinger EP, Gross PA, Barrett TL et al. Quality standard for
antimicrobial prophylaxis in surgical procedures. Infectious Diseases
Society of America. Clin Infect Dis 1994; 18: 422–7.
8. Polk HC Jr, Christmas AB. Prophylactic antibiotics in surgery
and surgical wound infections. Am Surg 2000; 66: 105–11.
9. James M, Martinez EA. Antibiotics and perioperative infections.
Best Pract Res Clin Anaesthesiol 2008; 22: 571–84.
10. Mangram AJ, Horan TC, Pearson ML et al. Guideline for
Prevention of Surgical Site Infection, 1999. Centers for Disease
Control and Prevention (CDC) Hospital Infection Control Practices
Advisory Committee. Am J Infect Control 1999; 27: 97–132; quiz
133–4; discussion 96.
11. Socie ´te ´ Franc ¸aise d’Anesthe ´sie et de Re ´anmation (SFAR).
Recommandations pour la pratique de l’antibioprophylaxie en chirurgie.
Actualisation 1999. [The French Society of Anesthesia and Intensive
Care. Recommendations for the practice of antibiotic prophylaxis in
surgery. Current status 1999.] 1999. Available at http://www.sfar.org/
consensusantibio.html (9 April 2009, date last accessed).
Miliani et al.
12. Stratchounski LS, Taylor EW, Dellinger EP et al. Antibiotic pol- Download full-text
icies in surgery: a consensus paper. Int J Antimicrob Agents 2005; 26:
13. Lallemand S, Thouverez M, Bailly P et al. Non-observance of
guidelines for surgical antimicrobial prophylaxis and surgical-site infec-
tions. Pharm World Sci 2002; 24: 95–9.
14. Bratzler DW, Houck PM, Richards C et al. Use of antimicrobial
prophylaxis for major surgery: baseline results from the National
Surgical Infection Prevention Project. Arch Surg 2005; 140: 174–82.
15. Tourmousoglou CE, Yiannakopoulou E, Kalapothaki V et al.
Adherence to guidelines for antibiotic prophylaxis in general surgery: a
critical appraisal. J Antimicrob Chemother 2008; 61: 214–8.
16. Yalcin AN, Serin S, Erbay H et al. Increased costs due to inap-
propriate surgical antibiotic prophylaxis in a university hospital. J Hosp
Infect 2002; 52: 228–9.
17. Privitera G, Scarpellini P, Ortisi G et al. Prospective study of
Clostridium difficile intestinal colonization and disease following single-
dose antibiotic prophylaxis in surgery. Antimicrob Agents Chemother
1991; 35: 208–10.
18. Fukatsu K, Saito H, Matsuda T et al. Influences of type and dur-
ation of antimicrobial prophylaxis on an outbreak of methicillin-resistant
Staphylococcus aureus and on the incidence of wound infection. Arch
Surg 1997; 132: 1320–5.
19. Harbarth S, Samore MH, Lichtenberg D et al. Prolonged anti-
biotic prophylaxis after cardiovascular surgery and its effect on surgical
site infections and antimicrobial resistance. Circulation 2000; 101:
20. Weber WP, Marti WR, Zwahlen M et al. The timing of surgical
antimicrobial prophylaxis. Ann Surg 2008; 247: 918–26.
21. Rioux C, Grandbastien B, Astagneau P. Impact of a six-year
control programme on surgical site infections in France: results of the
INCISO surveillance. J Hosp Infect 2007; 66: 217–23.
22. Owens WD, Felts JA, Spitznagel EL Jr. ASA physical status
classifications: a study of consistency of ratings. Anesthesiology 1978;
23. Altemeier WA, Burke JF, Pruitt BA et al. Definitions and classifi-
cations of surgical infections. In: Manual on Control of Infection in
Surgical Patients. Philadelphia: JB Lippincott Co, 1984; 19–30.
24. Horan TC, Gaynes RP, Martone WJ et al. CDC definitions of
nosocomial surgical site infections, 1992: a modification of CDC defi-
nitions of surgical wound infections. Am J Infect Control 1992; 20:
25. Culver DH, Horan TC, Gaynes RP et al. Surgical wound infec-
tion rates by wound class, operative procedure, and patient risk index.
National Nosocomial Infections Surveillance System. Am J Med 1991;
26. Rabe-Hesketh S, Skrondal A. Multilevel and Longitudinal
Modeling Using Stata (Second Edition). College Station, TX: Stata
27. Zanetti G, Giardina R, Platt R. Intraoperative redosing of
cefazolin and risk for surgical site infection in cardiac surgery. Emerg
Infect Dis 2001; 7: 828–31.
28. Scher KS. Studies on the duration of antibiotic administration for
surgical prophylaxis. Am Surg 1997; 63: 59–62.
29. DiPiro JT, Vallner JJ, Bowden TA Jr et al. Intraoperative serum
and tissue activity of cefazolin and cefoxitin. Arch Surg 1985; 120:
30. van Kasteren ME, Mannien J, Ott A et al. Antibiotic prophylaxis
and the risk of surgical site infections following total hip arthroplasty:
timely administration is the most important factor. Clin Infect Dis 2007;
31. Nelson CL, Green TG, Porter RA et al. One day versus seven
days of preventive antibiotic therapy in orthopedic surgery. Clin Orthop
Relat Res 1983: 258–63.
32. Bucknell SJ, Mohajeri M, Low J et al. Single- versus multiple-
dose antibiotics prophylaxis for cardiac surgery. Aust N Z J Surg 2000;
33. Fonseca SN, Kunzle SR, Junqueira MJ et al. Implementing
1-dose antibiotic prophylaxis for prevention of surgical site infection.
Arch Surg 2006; 141: 1109–13; discussion 1114.
34. Tang WM, Chiu KY, Ng TP et al. Efficacy of a single dose of
cefazolin as a prophylactic antibiotic in primary arthroplasty. J
Arthroplasty 2003; 18: 714–8.
35. Kriaras I, Michalopoulos A, Turina M et al. Evolution of antimi-
crobial prophylaxis in cardiovascular surgery. Eur J Cardiothorac Surg
2000; 18: 440–6.
36. Classen DC, Evans RS, Pestotnik SL et al. The timing of pro-
phylactic administration of antibiotics and the risk of surgical-wound
infection. N Engl J Med 1992; 326: 281–6.
37. Steinberg JP, Braun BI, Hellinger WC et al. Timing of
antimicrobial prophylaxis and the risk of surgical site infections: results
from the trial to reduce antimicrobial prophylaxis errors. Ann Surg
2009; 24: 24.
Surgical prophylaxis and infection risk