INTRODUCTION

Healthcare-associated infections (HAIs) are frequent on surgical wards and represent a high burden on patients and
hospitals in terms of morbidity, mortality, prolonged length of hospital stay and additional costs. Surgical site infections
(SSIs) are an important source and may even be the most frequent HAI after excluding asymptomatic bacteriuria. Apart
from endogenous risk factors, such as immune suppression, obesity or advanced age, the role of external risk factors in
SSI pathogenesis is now clearly established. Multimodal, multicenter or supranational preventive intervention programs
based on guidelines, 'bundles or safety checklists are gaining momentum on a global scale. In parallel, randomized
studies provide insight into poorly explored risk factors and practical intervention measures. The National Institute for
Health and Clinical Excellence (NICE) in England, Wales and Northern Ireland issued guidance for the prevention and
treatment of SSI in October 2008, and the 1999 SSI guidelines of the CDC are currently under revision.
We summarize the state-of-the-art regarding SSI prevention among adult inpatients, highlight important epidemiological
features and discuss pitfalls of surveillance and the possible role of benchmarking SSI rates. The practical questions
regarding the most effective measures to reduce SSI and the SSI rates achievable today are also addressed, as well as
the theoretical possibility of achieving a zero SSI policy on a surgical ward, at least for clean orthopedic surgery.
METHODS
The aim of the research was to provide an overview of the current state-of-the-art of SSI prevention with an emphasis on
literature published during the last 5 years, particularly the most recent. Landmark studies and important publications are
incorporated for an overriding purpose. The first author performed a PubMed search of the literature to identify English,
French and German language publications prior to 10 January 2010 using the following MeSH terms in various
combinations: 'surgical site infection', 'nosocomial', 'surgery', 'orthopaedic', 'infection', 'prosthesis', 'arthroplasty', 'zero',
'prophylaxis', 'prevention', 'bundles' and 'guidelines'. The search was verified by the second and last author for pertinence
to the topic. Reference lists of identified articles were searched manually to retrieve additional literature published after
January 2004. Animal studies and studies with an outcome other than SSI, for example, colonization studies, in
vitro studies and pediatric reports were excluded. We concentrated on articles with data on the post-discharge
surveillance of SSI. A total of 205 articles were retained and form the basis of this review. Sterilization techniques of
surgical instruments and problems related to prion diseases are not included as we considered these to be outside the
scope of this article.

DIFFICULTIES IN ASSESSING INFECTIONS

An active surveillance program is the cornerstone for the detection of HAI and accurate calculation of SSI rates within an
institution. Unfortunately, there is no consensus on a universally accepted surveillance strategy that would allow
benchmarking of assessable infections. However, the National Nosocomial Infections Surveillance (NNIS) index for
stratification could be useful, for example, and would represent at least an attempt to provide comparisons between
hospitals. Surgical patients do not only acquire SSIs but they may be at higher risk to experience a HAI than nonsurgical
patients, thus resulting in an additional burden on both the patient and the institution. If other institutional HAI rates remain
unchanged or are increasing, a reduction of SSI may have only a minor impact on patient safety. To reduce infections in
surgical wards, an infection control strategy must target all HAIs, including less severe HAIs, such as urinary tract
infections.
Surveillance can be active, for example, by scheduled medical follow-up visits, or passive, but should be conducted within
a minimum delay of 30 days following surgery (1 year in the case of implant surgery. Of note, programs without active
post-discharge surveillance readily miss between 48 and 79% of all SSIs. An additional problem is wound depth. While
most surveillance strategies easily detect post-surgical deep or organ space infections, they often fail to record superficial
infections. This is an important issue as these infections may represent up to 65% of the overall SSI burden in absolute
numbers. Hematogenous infections are witnessed in implant-related surgery. Although these are 'surgical site' in terms of
localization, they originate from other sites and might be interpreted as primary SSIs if data are not validated by an expert.
Finally, in surveillance programs based on microbiological evidence alone, the use of sonication of implants may enhance
and alter the proportion of microbiologically documented SSIs. In summary, the surveillance strategy has a large influence
on infection rates: the more sophisticated the strategy, the more likely the chance of detecting SSI and not to claim zero
rates of HAI or SSI.

CURRENT BENCHMARKS
The preventable proportion of HAI ranges from 10 to 70% in the literature. So far, a complete absence of nosocomial
infection over several months has only been reported for catheter-related bloodstream infections. To the best of our
knowledge, only one report has claimed almost zero SSI over an extended period of time (one of 176 neurosurgical shunt
operations) and it would appear that this is not easy to achieve outside minimally invasive, ambulatory surgery. In general,
emergency surgery, dirty contaminated surgery and surgery in resource-poor settings yield higher SSI rates than general
or clean elective surgery in high-income countries (Table 1), presumably because clean elective surgery is performed in
soft tissues or bone structures that are not routinely colonized by bacteria.
Among all clean elective surgery, cataract and orthopedic surgery have the fewest SSIs, with an overall crude incidence of
0.05 to 1.1%, respectively. According to a large prevalence study in northern France, the relative SSI risk of genitourinary,
cardiovascular, gynecologic and gastrointestinal surgery compared with orthopedic surgery was 2.1, 2.4, 2.6, 3.4 and 4.8,
respectively. Within the group of clean elective orthopedic surgery, SSI risks may vary between different procedures.
Arthroscopies harbor the absolute lowest SSI rates (0.10.4%) (Table 2) and primary arthroplasties carry the lowest
infection risk (0.50.9%) for implant-related procedures. Differences in arthroplasty infection rates may depend on the
surveillance methods employed and the proportion of knee arthroplasties, which reveal a slightly higher infection risk than
hip arthroplasties. The surveillance system used can also explain outliers, such as in the case of the United States
Medicare system, with a documented lower incidence of primary arthroplasties (0.23%). However, only deep SSIs
identified at 90 days post-implantation are reported in this system.

PATHOGENESIS
Most SSIs are believed to be acquired during surgery. This is supported by the success of SSI prevention measures
directed towards activities in the operating theater and a few reports demonstrating matching strains of pathogens from
the surgeon's fingers and postoperative infection. However, despite much research on SSI, there are currently no data on
the actual proportion acquired in the operating theater versus post-operative care onwards. Similarly, within the subgroup
of SSIs acquired during surgery, the proportion originating from the patient versus that transmitted by the surgical staff,
operating theater procedure or the environment remains unknown.

INDEPENDENT RISK FACTORS FOR SURGICAL SITE INFECTIONS

Over the past decades, identified risk factors have been assessed through epidemiological studies due to a lack of
knowledge of the detailed pathogenesis of SSIs. The literature reports dozens of independent risk factors, mostly
assessed by retrospective casecontrol (Table 3) or beforeafter quasi-experimental studies. In this article, we only
discuss the risk factors that have been shown to be independent determinants by multivariate analysis. Among the most
frequently cited are diabetes mellitus, age, obesity and incorrect or lack of antibiotic prophylaxis; other mentioned factors,
such as low socioeconomic status or postoperative pain, have not been thoroughly investigated. It is likely that with the
construction of large databases for nationwide surveillance programs, new and as yet unidentified risk factors will emerge
from research. Data on the clinical impact related to both endogenous and exogenous independent risk factors reveal risk
indices oscillating between 1.3 and 4.5 (Table 3). Approximately half of all identified risk factors are endogenous and
difficult to modify in the immediate preoperative and perioperative phase. For exogenous risk factors, past experience has
identified those with the best and least costly chance to decrease SSIs

EVIDENCE-BASED PREVENTIVE MEASURES

At present, four preventive measures are considered as having a high level of evidence (grade IA) according to major
evidence-based guidelines: surgical hand preparation; appropriate antibiotic prophylaxis; and postponing of an elective
operation in the case of active remote infection. Although hair clipping before surgery was considered grade IA evidence
in the 1999 CDC guidelines, this high grading is now a matter of debate.

Surgical Hand Preparation Surgical hand preparation is probably the most important SSI prevention strategy, although
there is no strict randomized study comparing surgery with and without previous hand antisepsis preparation sensu
strictu. Its importance is supported by expert opinion, experimental studies and success stories of SSI reduction via mere
hand hygiene promotion campaigns. However, owing to their multimodal design, most hand hygiene campaigns cannot
distinguish between SSI reduction due to improved antisepsis in the operating theater versus better patient and wound
care on the ward. Few interventions have targeted the operating theater as in the seminal intervention of Ignaz
Semmelweis (18181865) who recognized that the incidence of puerperal fever was high in his obstetric clinic. After the
compulsory introduction of hand antisepsis for obstetricians using chlorinated lime, he succeeded in lowering the
incidence of this life-threatening, postpartum maternal infection. A cluster-randomized, crossover trial recently reported the
equivalence of surgical hand preparation with unmedicated soap and water versus alcohol-based handrub on
postoperative SSI rates. Parienti et al. reported the equivalence of a handrubbing protocol with a 75% alcohol-based
solution versus a handscrubbing protocol containing 4% povidone-iodine or 45% chlorhexidine gluconate in terms of SSI
prevention. A Cochrane review also addressed the issue of preoperative surgical hand preparation. Hand rubbing with an
alcohol-based formulation was considered as effective as scrubbing, for which the ideal duration remains
unknown. Although it is probable that the minimum duration is 23 min for both techniques. Either alcohol-based hand
rubs or aqueous antiseptic scrubs can then be subsequently used between patients, provided hands are not visibly
soiled. However, the rapid antimicrobial action, wider spectrum of activity, lower side effects and the absence of the risk of
hand contamination by rinsing water in resource-poor areas might favor alcohol-based solutions. Brushes are not
recommended for surgical hand preparation.

Antibiotic Prophylaxis The effectiveness of the administration of preoperative antibiotic agents can be assumed, at least
for most surgical interventions. Exceptions are clean elective surgery without foreign material, for example, hernia repair,
removal of implant material, dermatologic surgery, and some foot and ankle surgery, where the number-to-treat needed to
prevent one SSI episode might be too small to justify routine administration for a few superficial easy-to-treat SSIs. In
orthopedic implant surgery, prophylaxis helps in reducing SSI rates from 48% without antibiotics to 13%, according to
trials performed in the 1970s and 1980s.

Successful and adequate prophylactic antibiotic use depends on a number of key principles: the patient's individual history
of allergy must be considered; first- or second-generation cephalosporins are sufficient for most types of surgery. The
proportional distribution of SSI pathogens has remained stable over the last two decades.even if the recent challenge of
community-acquired methicillin-resistant Staphylococcus aureus (MRSA) and other resistant pathogens might become a
threat in the future. In the case of skin colonization with methicillin-resistant staphylococci, a glycopeptide antibiotic is
recommended. Some experts advocate that general vancomycin prophylaxis may be appropriate in settings with more
than 20% MRSA prevalence, but from an epidemiologic standpoint, there is no threshold for routine glycopeptide
prophylaxis in settings with endemicity for methicillin-resistant staphylococci. Furthermore, there is no evidence that a
glycopeptide would be superior to cephalosporins for patients without MRSA carriage. With a few exceptions, such as
dermatologic or cataract surgery, antibiotics have to be administered parenterally without adaptation of the recommended
dose for renal insufficiency. Suboptimal dosing is a risk factor for SSI. An additional benefit of antibiotics in irrigation fluid
or of antibiotic-loaded cements in primary arthroplasty surgery has not yet been proven. Timing is of the utmost
importance and prophylaxis should be within 1 h before incision. Recent reports attempted to further define the optimal
time frame for administration of antibiotic prophylaxis, but showed contrasting results. While two studies identified the best
time window as between 0 and 30 min before incision, another from Switzerland identified a delay of 3060 min to be
superior in terms of SSI reduction. Undoubtedly, this debate will continue until definitive research is conducted, ideally
using a cluster-randomized crossover trial. One dose is sufficient. For operating procedures longer than 4 h or in the case
of significant blood loss, redosing may be justified, but this recommendation depends on the half-life of the antimicrobial
agent. At present, it remains unknown if a repeated administration should use the same or a reduced dosage, but there is
no advantage of a continuous versus intermittent infusion in the case of redosing. In any case, a prolongation of
prophylaxis beyond 24 h postoperative is of little benefit (except for surgery of large burn wounds where pre-emptive
administration of antibiotics may be beneficial beyond 24 h) and favors the acquisition of antibiotic resistance, particularly
among Gram-negative pathogens. When a tourniquet is used, the entire dose should be administered prior to its
inflation. At present, it remains unknown whether standard doses should be routinely enhanced for obese patients, for
instance, patients with a BMI of more than 30 kg/m 2 or higher. Different experts recommend a higher prophylactic dose for
selected pathogens or infections, but this is an empirical opinion and no clinical trials have demonstrated the advantage of
this approach so far.
Postponing Elective Surgery in the Case of Symptomatic Remote Infection This issue is regarded as high evidence
in the CDC guidelines, although there are no randomized trials on the topic of postponing to the best of our knowledge.
However, a number of case series and retrospective studies exist reporting a hematogenous origin for total joint
arthroplasty infections and secondary infections of other orthopedic implants. Although the incidence of true
hematogenous arthroplasty infections in the case of active remote infection is probably lower than formerly believed, most
experts agree that elective surgery should be postponed until the remote infection is cured. The necessity to postpone in
the case of asymptomatic bacteriuria or even urinary tract infection in elective orthopedic surgery or nonimplant-related
surgery is less certain. Common sources of hematogenous implant infections are often the skin, the GI tract or the lungs.

EXPERT COMMENTARY
Zero HAI (or zero SSI) is theoretically possible in clean elective surgery, but there is no published report so far of zero
infection in a surgical unit over a prolonged period. All HAIs have to be targeted, thus requiring different prevention
programs conducted at the same time. This goal requires multidisciplinary, multifaceted commitment, dedicated infection
control teams and efforts, and institutional and behavioral elements to ensure sustainability. National surveillance
networks may help benchmarking. Current benchmarks for SSI are 0.1% risk in arthroscopy and cataract surgery. A
multimodal approach is recommended, including optimized antibiotic prophylaxis, active post-discharge surveillance and
continuous performance feedback.