Introduction

When the skin is incised, microorganisms colonising the surface may contaminate the exposed tissues and subsequently proliferate and lead to an SSI. Interventions that reduce the number of microorganisms on the skin surrounding the incision may therefore decrease the risk of SSI. The microbial flora on the skin comprises transient microorganisms that are acquired by touch and easily removed by washing with soap, and resident flora that normally live in the skin appendages such as hair follicles. The resident flora are generally not pathogenic but are not so readily removed by soap although their numbers can be reduced by antiseptics. The purpose of the review was to determine the clinical effectiveness of preoperative bathing or showering with antiseptics for the prevention of SSI.

Overview of evidence

One systematic review was identified.

One well-conducted systematic review³⁴ (six RCTs, n = 10 007 participants) examined the evidence for preoperative bathing or showering with antiseptics for the prevention of SSI. [EL = 1+] Patients were undergoing orthopaedic, vascular, biliary tract, inguinal hernia, breast, vasectomy and other general surgical operations. The incidence of SSI was the primary outcome measure in all studies although definitions varied among studies. Four studies had two treatment arms and two had three treatment arms. The only antiseptic used in the included studies was chlorhexidine.

Two RCTs (Figure 5.1) compared the effect on SSI of showering with 4% chlorhexidine against no showering. The smaller trial (n = 64 participants) found no difference in the SSI rate between the two groups (RR 1.33, 95% CI 0.65 to 2.72), while the larger trial (n = 978) found statistically significantly fewer SSIs in the group that used chlorhexidine (9/541) than in the group that did not shower (20/437) (RR 0.36, 95% CI 0.17 to 0.79).

Figure 5.1

Comparison of showering with 4% chlorhexidine versus no showering on SSI incidence in two trials.

Five studies in the systematic review examined the effect of preoperative showering or bathing with 4% chlorhexidine solution compared with a detergent or bar soap. Three RCTs (n = 7691 participants) used a detergent and three RCTs (n = 1443) used bar soap as a comparator. It should be noted that one of these studies³⁵ used a detergent that was subsequently discovered to have antimicrobial properties.

A meta-analysis (Figure 5.2) of these five RCTs (n = 8445 participants) demonstrated that the incidence of SSIs was not statistically significantly different between groups showering with chlorhexidine (375/3919) and with detergent or bar soap (487/4526) (RR 0.90, 95% CI 0.79 to 1.02, I² = 35.3%).

Figure 5.2

Meta-analysis of five trials comparing the effect of showering with chlorhexidine versus showering with detergent or bar soap on SSI incidence.

One included RCT (n = 1093) found that total body washing with chlorhexidine produced a statistically significant reduction in SSI incidence compared with partial body washing where only the skin area at the site of incision was washed (RR 0.40, 95% CI 0.19 to 0.85).

Overview of evidence

One systematic review³⁴ that examined the evidence for preoperative bathing or showering with antiseptics for the prevention of SSI made reference to an analysis comparing ‘one wash against more than one wash’ that had been published in a previous Cochrane Library issue, but which had been subsequently withdrawn. [EL = 1+]

This analysis was removed because no trial specifically randomised patients by number of washes and the methodology was deemed insufficiently rigorous for publication.

Similarly, no trials were identified that specifically randomised patients by timing of washes.

Health economics overview of evidence

One RCT was identified.

One RCT³⁶ compared a chlorhexidine detergent shower three times before elective surgery with three showers using detergent. This study did not find that chlorhexidine was associated with a statistically significant reduction in wound infection. The average cost of both non-infected and infected patients was found to be higher in the chlorhexidine than in the detergent group. The average cost of a non-infected chlorhexidine-treated patient was £847.95 compared with £804.60 for a non-infected patient treated with detergent alone, whereas the average cost of an infected patient was £1,459.70 (chlorhexidine) and £1,414.22 (detergent). The authors concluded that preoperative whole-body disinfection with a chlorhexidine detergent was not a cost-effective treatment for reducing wound infection, being more costly and showing no statistically significant clinical benefit.

Evidence statements

There is evidence from one RCT that showering or bathing using chlorhexidine significantly reduces the rate of SSI compared with no showering. [EL = 1+]

There is evidence of no difference in SSI incidence when chlorhexidine or detergent/bar soap is used for preoperative showering or bathing. [EL = 1+]

There is no (systematic review or RCT) evidence that examines the clinical effectiveness of the timing or number of preoperative showers to prevent SSI.

Health economics evidence statement

There is evidence to indicate that preoperative showering with a chlorhexidine detergent is not a cost-effective intervention to prevent SSIs when compared with preoperative showering with a detergent or bar soap.

GDG interpretation

One study demonstrated a significant reduction in SSI associated with a chlorhexidine preoperative shower compared with no showering or a partial body wash, and, in one study, whole-body showering with chlorhexidine was compared with a partial wash. In a separate meta-analysis, chlorhexidine was demonstrated to be no more effective than bar soap or detergent in the prevention of SSI and one RCT found it not to be a cost-effective intervention.

Therefore, while there is evidence to support the efficacy of preoperative showering as a measure to reduce the rate of SSI, there is evidence of no difference of effect on SSI rate between chlorhexidine as a cleansing agent and plain detergent or soap. In addition, chlorhexidine has been found not to be cost-effective.

None of the studies provided evidence to indicate whether the number and timing of preoperative showers affected the rate of SSI but the GDG view was that showering should take place as close to or on the day of surgery.

Introduction

The removal of hair may be necessary to adequately view or access the operative site and it is sometimes undertaken because of a perceived increased risk of microbial contamination of the operative site from the presence of hair. However, micro-abrasions of the skin caused by shaving using razors may support the multiplication of bacteria, within the skin and on the skin surface, particularly if undertaken several hours prior to surgery. An increase in the number of microorganisms colonising the skin surrounding the operative site may facilitate contamination of the wound and subsequent development of SSI. Therefore, when hair removal is indicated, the method used should minimise damage to the skin. The purpose of the review was to determine the clinical effectiveness of preoperative removal of hair from the operative site to prevent SSI.

Overview of evidence

One systematic review and one additional RCT were identified.

One well-conducted systematic review³⁷ (11 RCTs, n = 4627 participants) was identified that examined the evidence for preoperative hair removal for the prevention of SSI. [EL = 1+] RCTs were included where adult patients undergoing any surgery in a designated operating theatre were allocated to groups comparing any hair removal schedule. Methods of hair removal included were shaving using razors, clipping and depilatory cream.

Two RCTs reported in the systematic review compared the effect of shaving using razors with no hair removal (total n = 358 adults). No SSIs were found in either group in the smaller study (n = 80) whereas, in the larger study (n = 278), 9.6% of people who were shaved developed an SSI compared with 6% people who were not shaved (RR 1.59, 95% CI 0.77 to 3.27).

A recent RCT³⁸ compared the effect of shaving using razors with no hair removal in spinal surgery patients in Turkey. [EL = 1+] There was no statistically significant difference between the two groups (RR 4.51, 95% CI 0.51 to 40.14).

Adding this latest study to the Cochrane meta-analysis and using a fixed effects model shows that there was no statistically significant difference in SSI incidence between shaving using razors and no hair removal (RR 1.82, 95% CI 0.93 to 3.59) (Figure 5.3).

Figure 5.3

Meta-analysis of three trials comparing the effect of shaving and no hair removal on SSI incidence.

One trial (n = 267 adults) reported in the systematic review compared SSI incidence in two groups randomised to either hair removal with depilatory cream (10/126) or to no hair removal (11/141). There was no statistically significant difference between the two groups (RR 1.02, 95% CI 0.45 to 2.31).

There were no studies comparing clipping of hair with no hair removal.

Three RCTs (n = 3193 participants) compared the relative effects of shaving using razors with those of clipping on the incidence of SSI: 2.8% (46/1627) of people who were shaved developed an SSI compared with 1.3% (21/1566) who had hair clipped preoperatively. This was a statistically significant difference (RR 2.02, 95% CI 1.21 to 3.36).

Seven trials (n = 1213 participants) reported in the systematic review compared the relative effects of shaving using razors with the use of depilatory cream for hair removal. Meta-analysis undertaken using a fixed effects model showed statistically significantly more SSIs in patients who were shaved (65/670) compared with those who had hair removed with depilatory cream (38/543) (RR 1.54, 95% CI 1.05 to 2.24).

There were no studies that compared clipping with depilatory cream.

Introduction

The timing of hair removal may be important since deep skin organisms may be encouraged to the skin surface following skin damage and may, therefore, contaminate the operative field.

Overview of evidence

One systematic review was identified.

The same Cochrane systematic review³⁷ (11 RCTs, n = 4627 participants) examined the evidence for the timing of preoperative hair removal for the prevention of SSI. [EL = 1+]

One RCT reported in the review compared timings of hair removal. Participants were adults undergoing general clean surgery in a designated operating theatre. Observations of SSI at 15 and 30 days postoperatively were made for hair removal performed the night before and the morning of the patient’s surgery. Both shaving using razors and clipping were investigated.

Health economics overview of evidence

Five studies were included.^39–43

The studies examined and compared different techniques of preoperative hair removal (shaving using razors, use of depilatory cream and clipping, as well as no hair removal).

It was difficult to ascertain the most cost-effective form of hair removal from these studies, most of which were more than 20 years old. Therefore, an economic model was developed to evaluate the cost-effectiveness of the various hair removal techniques in a UK context (see Appendix D). It showed that electric clippers were the most cost-effective method for preoperative hair removal.

Health economics evidence statements

There is evidence from the literature that the use of razors to remove patients’ hair prior to surgery is not cost-effective.

Evidence from a decision-analytic model showed that the use of electric clippers for preoperative hair removal was cost-effective when compared with no hair removal, shaving using razors, or depilatory cream. The use of electric clippers was not only found to generate more quality-adjusted life years (QALYs) but was also found to be less expensive than these two interventions.

Evidence statements

There is evidence that there is no difference in SSI incidence following preoperative hair removal (using depilatory cream or by shaving using razors) or no hair removal. [EL = 1+]

There is evidence that fewer SSIs occur following hair removal with clippers or depilatory creams compared with shaving using razors. [EL = 1+]

There is insufficient evidence to determine whether the timing of the preoperative shaving using razors or clipping of hair at the operative site affects the incidence of SSI. [EL = 1+]

There is a risk of skin reactions with the use of depilatory creams.

There is evidence that shaving using razors is associated with more SSIs than any other method of hair removal. [EL = 1+]

GDG interpretation

There is no evidence that hair removal in general influences the incidence of SSI, but it might be appropriate in some clinical circumstances. However, if hair has to be removed, there is evidence that shaving using razors increases the risk of SSI.

There is insufficient evidence on whether the timing of hair removal affects the risk of SSI but the GDG consensus was that where hair removal is required it should be undertaken as close to the time of surgery as possible but clipping on the day of surgery may be preferable. Electric clippers with single-use disposable heads are the most cost-effective method of hair removal.

Introduction

It has been traditional for patients to put on clean clothing (and in some units to remove underwear) on the ward before being taken to the operating theatre. Any risk of infection from airborne spread from socially clean clothing is unlikely to be large because, in comparison with the operating team, little patient movement occurs during operations thus reducing the dispersal of microorganisms from skin and clothing. The purpose of the review was to determine whether patient theatre attire can affect the incidence of SSI.

Overview of evidence

No studies were identified that examined patient theatre attire and postoperative SSI rates.

Evidence statement

There was no evidence identified to determine whether patient theatre attire can affect the incidence of SSI.

GDG interpretation

There is no evidence concerning patient theatre attire but operating department clothing should maintain the dignity and comfort of the patient and allow easy access to the operative site as well as other areas for placement of intravenous cannulas, catheters and epidurals, etc. Operative wear may also be preferred when the patient’s own clothes may be at risk of contamination from blood, body and washout fluids.

Introduction

It is traditional for the operating team to put on freshly laundered, but non-sterile, theatre wear prior to a surgical procedure in an operating theatre environment, and to change this scrub suit for a fresh set should any of it become soiled by blood or other body fluids. Scrub suits are usually re-laundered but other components are usually disposable. The purpose of the review was to determine the clinical effectiveness of theatre staff wearing non-sterile theatre wear (scrub suits, masks, hats and overshoes) for the prevention of SSI.

Overview of evidence

Evidence statements

There is no evidence available that examines whether the wearing of scrub suits or head attire or overshoes by scrubbed or circulating theatre staff can prevent SSI.

Evidence from two quasi-RCTs show that there is no difference in the rate of SSI when face masks are worn during clean or dirty surgery. [EL = 1+]

GDG interpretation

In theatre, there is a need to minimise the risk of microbial contamination of the operating site from the theatre environment. Traditionally, this has been referred to as theatre discipline. Although there is limited evidence concerning the use of specific non-sterile theatre wear (scrub suits, masks, hats and overshoes), the GDG consensus was that wearing non-sterile theatre wear is important in maintaining theatre discipline and may therefore contribute to minimising the risk of SSI.

A separate issue of the protection of operating staff from exposure to patients’ body fluids was beyond the scope of the GDG and is covered by health and safety regulations.

Introduction

It is traditional to change non-sterile theatre wear into conventional clothing when leaving the operating environment and to put on fresh theatre wear when re-entering. The purpose of the review was to determine whether staff exiting and re-entering the operating room can affect the incidence of SSI.

Overview of evidence

No studies were identified that examined the effect of staff movement in and out of the operating room on SSI rates.

Evidence statement

There is no evidence to determine whether staff exiting and re-entering the operating area has an effect on the incidence of SSI.

GDG interpretation

It is good practice to discard all used theatre wear prior to leaving the operating area to prevent healthcare workers, patients and visitors being exposed to the risk of contamination. However, there is no evidence that this practice has any effect on the incidence of SSI.

The GDG consensus was that staff should not leave the operating theatre suite wearing non-sterile theatre wear as this is important in the maintenance of theatre discipline and may therefore contribute to minimising the risk of SSI.

Introduction

The anterior nares (front of the nose, within the nostril) are the main reservoir for the multiplication of Staphylococcus aureus in the body, and S. aureus spreads from this site to other places on the skin surface. Up to one-third of people carry S. aureus persistently in their nares and about a further one-third do so intermittently. S. aureus is the most common cause of SSI in all types of surgery, with the microorganism frequently being derived from the patients themselves. Measures to clear carriage of S. aureus from the anterior nares around the time of surgery have thus been investigated to assess whether they reduce SSI. Such measures usually involve applying topical antiseptics or antibiotics active against S. aureus. Theoretically, it may take several days of treatment to clear S. aureus from the anterior nares and also from other carriage sites and prolonged treatment may be difficult to achieve in practice for all patients.

It is important for studies in this area to determine whether the measures used have actually reduced S. aureus carriage and whether both S. aureus and total SSI rates have been influenced. This is because eliminating S. aureus carriage from a patient might, for example, leave them prone to acquiring carriage (and hence infection) with other bacteria.

The purpose of the review was to determine the clinical effectiveness of nasal decontamination using topical antimicrobial agents for the prevention of SSI.

Overview of evidence

Five RCTs were identified.

Five RCTs^45–49 examined the effects of nasal decontamination for prevention of SSI. Participants were undergoing orthopaedic, digestive, cardiothoracic, gynaecological, neurological, oncological and general surgery.

Three studies compared the effects of intranasal mupirocin with placebo,^45–47 although participants in one trial⁴⁷ were all S. aureus carriers. A further trial⁴⁸ compared mupirocin with no intervention and another⁴⁹ compared the effect of chlorhexidine mouthwash and nasal gel with that of placebo on SSI incidence.

Two RCTs^45,46 (n = 4478 participants) examined whether there was any difference in SSI incidence following nasal decontamination with mupirocin and with placebo. [EL = 1+] Data were pooled in a meta-analysis (Figure 5.4). There was no heterogeneity and no statistically significant difference in SSI incidence between the two groups (OR 0.97, 95% CI 0.77 to 1.21).

Figure 5.4

Meta-analysis of two trials comparing the effect of nasal decontamination with mupirocin versus placebo on SSI incidence.

Two RCTs^46,47 examined mupirocin compared with a placebo in patients carrying S. aureus only. [EL = 1+] Heterogeneity between studies prevented pooling (I² = 66%) and individual findings for SSI incidence were not statistically significant for either study (respectively OR 0.84,⁴⁶ 95% CI 0.55 to 1.28, n = 891, and OR 1.88,⁴⁷ 95% CI 0.83 to 4.25, n = 157).

These two studies^46,47 also presented findings for a comparison of mupirocin with placebo for S. aureus infections in S. aureus carriers (n = 1128). [EL = 1+] There was no statistically significant difference in S. aureus infection incidence between the two groups of S. aureus carriers (OR 0.69, 95% CI 0.39 to 1.22).

One trial⁴⁸ (n = 395 participants) compared the SSI incidence following nasal decontamination with mupirocin with that following no nasal decontamination in patients undergoing abdominal digestive surgery. [EL = 1+] There was no statistically significant difference in SSI rate between treatment arms (OR 1.39, 95% CI 0.76 to 2.52).

One trial⁴⁹ (n = 954 participants) comparing the effect of chlorhexidine with that of placebo found no statistically significant difference in SSI rates between groups (OR 0.88, 95% CI 0.58 to 1.33). [EL = 1+]

One trial⁴⁹ reported one adverse event. [EL = 1+] One participant receiving chlorhexidine oral rinse and nasal gel complained of tooth staining. No other adverse events were detailed in this or any other included study.

Overview of evidence

No single RCT compared timing of nasal decontamination for prevention of SSI.

Health economics overview of evidence

Two full economic analysis papers^50,51 were identified.

A cost-effectiveness analysis⁵⁰ compared mupirocin ointment treatment with no preventative treatment in cardiothoracic surgery patients. The outcome used was cost per SSI prevented. It was found that treating 1000 surgical patients with mupirocin would lead to a cost saving of $747,969, which is $16,633 saved per SSI prevented. However no staff costs were considered for the application of mupirocin, which would make using mupirocin ointment more expensive.

A cost-effectiveness analysis⁵¹ compared the following strategies:

• screening patients for S. aureus colonisation with nasal culture and treating carriers with mupirocin
• no screening but treating all patients with mupirocin
• no screening with no preventative treatment.

The outcomes of the analysis were cost per infection avoided and cost per life year saved. The study concluded that both strategies that used mupirocin were cost-saving.

As neither published analysis was conducted in the UK, a new model was developed to assess the cost-effectiveness of mupirocin nasal ointment to prevent SSI caused by S. aureus (see Appendix E). Three strategies were compared:

• no treatment
• screen for S. aureus and treat identified carriers with mupirocin
• treat all patients with mupirocin.

The results with baseline values showed that treating all patients with mupirocin was the dominant strategy, resulting in the least number of SSIs and the lowest cost.

A deterministic threshold sensitivity analysis suggested that the cost of treating an SSI would have to be below £600 before the strategy of treating all patients with mupirocin exceeded £20,000 per QALY (the willingness-to-pay threshold used by NICE to determine cost-effectiveness). The point estimates on which baseline values were based were not statistically significant at the 5% level and a probabilistic sensitivity analysis was carried out to reflect the uncertainty in the effect size parameters. This suggested that there was approximately a 50% chance that treating all patients with mupirocin would be cost-effective.

However, this analysis did not model the potential harm of increased antibiotic resistance from treating all patients with mupirocin. Full details of the models are provided in Appendix E.

Health economics evidence statement

An economic evaluation with clinical effectiveness based on a single trial suggested that there was a 50% chance that treating all patients with mupirocin nasal ointment to prevent SSI caused by S. aureus is a cost-effective strategy.

Evidence statements

There is evidence that nasal decontamination with mupirocin or chlorhexidine administered to all patients undergoing surgery does not affect the overall rate of SSI.

There is evidence that nasal decontamination with mupirocin given to S. aureus carriers undergoing surgery does not reduce either the incidence of S. aureus SSI or the incidence of all-cause SSI. [EL = 1+]

There is insufficient evidence from RCTs to determine incidence of adverse effects with nasal decontamination treatment. [EL = 1+]

There is no evidence available that examined the clinical effectiveness of the timing of nasal decontamination strategies.

GDG interpretation

Mupirocin or chlorhexidine nasal decontamination does not reduce the overall rate of SSI. Nevertheless, in S. aureus carriers, there was a non-statistically significant reduction in SSIs caused by S. aureus when mupirocin was used.

An economic model suggested that there was considerable uncertainty about the cost-effectiveness of treating all patients with mupirocin nasal ointment to prevent SSI caused by S. aureus, and the GDG consensus was that it should not be recommended, especially as the potential harm of increased antibiotic resistance had not been factored into the model.

Introduction

Most SSIs are acquired intraoperatively from the bacterial flora colonising the patient’s skin, gastrointestinal tract and mucous membranes. At present, the best method to prevent SSI after colorectal surgery is a matter of debate. Traditional surgical practice has suggested that removal of faecal matter from the colon and rectum prior to elective colorectal surgery confers an advantage, and mechanical bowel preparation has become a fundamental component of intestinal surgery in many units. Mechanical bowel preparation has been considered to be advantageous for many reasons, including operative time, ease of handling of the bowel, rate of stoma formation and the ability to palpate lesions in the bowel wall. The purpose of the review was to determine the clinical effectiveness of preoperative mechanical bowel preparation for the prevention of SSI.

Overview of evidence

12 RCTs were identified.

A systematic review (nine RCTs, n = 1592 participants) published in 2005 was identified that investigated SSI incidence (as a secondary outcome) following mechanical bowel preparation in patients undergoing colorectal surgery. All nine trials were included here, although two trial reports published after the Cochrane review was prepared were used as they contained fuller detail. A further three trials^52–54 published within the last 2 years were also identified. [EL = 1+]

This gives a total of 12 included trials with patients who were all undergoing colorectal surgery. Different mechanical bowel preparation solutions were administered in the studies: polyethylene glycol, mannitol, sodium picosulfate, laxative/enema/mannitol and in two studies the solution was not reported.

Data from all trials were pooled in a meta-analysis (12 RCTs, n = 5383). All of these studies examined the clinical effectiveness of preoperative mechanical bowel preparation for the prevention of SSI.

There was no heterogeneity and no statistically significant difference in SSI incidence between the treatment and control groups (I² = 0% and OR 1.08, 95% CI 0.88 to 1.32 – fixed effects model) (Figure 5.5).

Figure 5.5

Meta-analysis of 12 trials comparing the effect of mechanical bowel preparation versus no mechanical bowel preparation on SSI incidence.

Evidence statement

There is evidence from a meta-analysis that there is no difference in the incidence rate of SSI for patients receiving bowel preparation when compared with no preparation in colorectal surgery. [EL = 1+]

GDG interpretation

The GDG recognised that there are different types of surgery (left- or right-sided colonic resections), different bowel preparations and different diseases (cancer or diverticular disease) that may have an impact on rates of SSI. The GDG recognised that there may be other indications where bowel preparation may be used, in particular to minimise the risk of an anastomotic leak and the formation of a stoma.

However, there is no evidence that bowel preparation influences the incidence of SSI in patients undergoing colorectal surgery.

Introduction

It is conventional for the operating team not to wear hand jewellery during surgical procedures, although some of the team may feel strongly about not removing wedding rings, and equally strongly that nail polish or nail extensions should be avoided. The purpose of the review was to evaluate the effects of the removal of nail polish, nail extensions and hand jewellery by the surgical scrub team on the prevention of postoperative SSI.

Overview of evidence

One systematic review⁵⁵ was identified that examined the effect of the surgical scrub team removing finger rings and nail polish on postoperative SSI rates.

No trials were identified that compared the wearing of finger rings with their removal. No trials were identified that compared the removal or wearing of nail polish with respect to SSI.

One well-conducted systematic review⁵⁵ (one RCT, n = 102 participants) looked at the effects of removing finger rings and nail polish on the incidence of SSI [EL = 1+]. Only one small trial was included. Participants were scrub team members. The study outcome was the bacterial load on fingernails before and after surgical scrubbing expressed as the number of colony-forming units (CFUs). The trial found no statistically significant difference in the number of CFUs between the two groups in the pre-scrubbing as in the post-scrubbing. Since there is insufficient evidence to establish a direct association between CFUs and SSI, the systematic review could not determine whether the removal or not of nail polish, hand jewellery or nail extensions has an effect on SSI rate.

Evidence statement

There is insufficient evidence to determine whether the removal or not of nail polish, hand jewellery or nail extensions has an effect on SSI rate. [EL = 1+]

GDG interpretation

There is no RCT evidence available to relate SSI to jewellery, nail polish or artificial nails. However, there was GDG concern that in certain circumstances artificial nails, nail polish and jewellery may conceal underlying soiling and impair hand decontamination. The GDG recognised that some rings may be difficult to remove.

Introduction

Antibiotic prophylaxis has been used effectively to prevent SSIs after appropriate operative procedures since 1969. Prophylaxis usually involves a single dose of antibiotic often given intravenously, close to the time of surgery (at induction of anaesthesia) and must be seen as different to treatment that entails a course of antibiotics over a period of time. In common with therapeutic use, the use of antibiotics for prophylaxis carries a risk of adverse drug reactions (including Clostridium difficile-associated diarrhoea) and increased prevalence of antibiotic-resistant bacteria. The choice of antibiotic prophylaxis should be influenced by the strength of the association between the antibiotic used and C. difficile diarrhoea. In this review the clinical effectiveness of antibiotic prophylaxis for various types of surgical procedures in the prevention of SSI was examined.

Searches were run for intravenous (IV) and oral antibiotic use, limited by study design (RCT and systematic reviews) but not by year.

Overview of evidence

The evidence is ordered by location of surgery and by surgery type. Evidence statements are grouped by the wound classification – clean, clean-contaminated, contaminated or dirty.

Head and neck surgery

Open reduction and internal fixation of compound mandibular fractures

One systematic review was included.

A systematic review was identified⁵⁸ (four RCTs, n = 461 participants) that examined the use of prophylactic antibiotics in the surgical treatment of maxillofacial fractures. [EL = 1+] Patients were undergoing surgery for mandibular or facial fractures and were randomised to receive either antibiotic or placebo/no prophylaxis. The antibiotics used were not reported in three studies and in the fourth IV cefazolin was administered. All studies included wound infection as an outcome. There was a mixture of open and closed reductions in one trial.

A meta-analysis (Figure 5.6) of the four studies found statistically significantly fewer wound infections in participants given antibiotic prophylaxis compared with those given placebo or no treatment (OR 0.18, 95% CI 0.10 to 0.32). Removal of the trial that mixed open and closed reduction of fractures did not remove statistical significance (OR 0.25, 95% CI 0.08 to 0.30).

Figure 5.6

Meta-analysis of four trials comparing the effect of antibiotic prophylaxis versus placebo or no treatment on SSI incidence in maxillofacial fracture surgery.

Contaminated/clean-contaminated head and neck surgery

One systematic review was included.

A systematic review⁶⁰ (12 RCTs) investigating antibiotic prophylaxis compared with placebo or with different antibiotic types or schedules in head and neck surgery was identified. [EL = 1+]. Three trials (237 participants) investigated the effect on wound infection of antibiotic prophylaxis compared with placebo.

All three trials included participants undergoing surgery for head and neck cancer. The antibiotics used were ampicillin/cloxacillin, cefazolin and cefoperazone/cefotaxime. One trial stopped placebo administration after examination of the results of the first 16 patients. All participants subsequently recruited instead received cefotaxime.

A meta-analysis (Figure 5.7) of these three trials found that there were statistically significantly fewer wound infections in patients who received antibiotics (19/155) than in those who received placebo (35/82) (OR 0.06, 95% CI 0.02 to 0.18).

Figure 5.7

Meta-analysis of three trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in clean-contaminated/contaminated head and neck surgery.

Breast cancer surgery

One systematic review and an RCT were identified.

One Cochrane systematic review⁶¹ (six RCTs, n = 1302 participants) was identified that included people with breast cancer undergoing breast surgery with or without immediate reconstruction as part of their treatment. [EL = 1+] The antibiotics used were azithromycin, cefonicid (two trials), clarithromycin, co-amoxiclav and cefazolin.

Five RCTs compared antibiotic with placebo and found statistically significantly fewer infections in the group receiving prophylaxis (RR 0.66, 95% CI 0.48 to 0.89) (Figure 5.8).

Figure 5.8

Meta-analysis of five trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in breast cancer surgery with or without immediate reconstruction.

One RCT included in the review compared clarithromycin with no intervention (RR not estimable, no events in either group).

A further, subsequently published trial (n = 618 participants) was identified.⁶² [EL = 1+] This study included patients scheduled for non-reconstructive breast surgery and compared the administration of a single dose of flucloxacillin immediately after anaesthesia induction with no treatment. The incidence of wound infection was similar in the two groups (OR 0.71, 95% CI 0.32 to 1.56) (Figure 5.9).

Figure 5.9

Comparison of single-dose flucloxacillin prophylaxis versus placebo on SSI incidence in two trials.

Cardiac pacemaker insertion

One systematic review was identified.

A systematic review⁶³ (seven RCTs, n = 2023 participants) of antibiotic prophylaxis for permanent pacemaker insertion was identified. [EL = 1+] All trials compared antibiotics with ‘control’, which was presumed to be a placebo or no treatment – this was implied although not specifically stated. The antibiotics used were flucloxacillin/benzylpenicillin, cloxacillin, cloxacillin/amoxicillin, ampicillin/flucloxacillin, cefazolin, cefazedone and flucloxacillin. The definition of infection was not given but included pocket infection and lead infection and may also have included septicaemia.

Meta-analysis of these studies demonstrated an overall statistically significant protective effect of antibiotic prophylaxis (5/1011) compared with no antibiotic treatment (37/1012) (OR 0.26, 95% CI 0.10 to 0.66) for infection.

Open-heart surgery

Three RCTs were identified.

Two trials were identified that examined the effect of antibiotic prophylaxis compared with placebo in coronary artery bypass graft (CABG)^64,65 and one trial in aorto-coronary bypass operations.⁶⁶ [EL = 1−, EL = 1+, EL = 1+, respectively] The antibiotics used were meticillin, cefradine and cefalotin. All studies were halted to examine infection rates in both groups. One study⁶⁵ was restarted with placebo still given, while the other two had protocols modified.

A meta-analysis (Figure 5.10) of these three RCTs showed that antibiotic prophylaxis reduced the rate of wound infections compared with placebo (OR 0.08, 95% CI 0.03 to 0.27).

Figure 5.10

Meta-analysis of three trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in open-heart surgery.

General thoracic surgery

Two RCTs of patients undergoing operations in general thoracic surgery units were identified.

One RCT⁶⁷ (n = 211 participants) randomised patients to receive either cefalotin or placebo at induction of anaesthesia. [EL = 1+] Patients were undergoing lung, hernia, gastroplasty and oesophageal surgery. Seven wound infections were found in the antibiotic group (n = 118 participants) and 22 in the placebo group (n = 93 participants) (OR 0.20, 95% CI 0.08 to 0.50).

The other RCT⁶⁸ (n = 127 participants) randomised participants to receive either cefazolin or placebo half an hour before surgery. [EL = 1+] Patients were undergoing pulmonary resection, atypical pulmonary resection, bullectomy, chest wall resection, oesophageal surgery and surgery for mediastinal tumours. There were statistically significantly fewer wound infections in the antibiotic group (2/70) than in the placebo group (8/57) (OR 0.18, 95% CI 0.04 to 0.89).

A meta-analysis (Figure 5.11) of these two studies that included a total of 238 participants also found that there were statistically significantly fewer wound infections with antibiotic prophylaxis compared with placebo (OR 0.20, 95% CI 0.09 to 0.43).

Figure 5.11

Meta-analysis of two trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in general thoracic surgery.

Upper gastrointestinal tract surgery

Stomach and duodenal surgery

Four RCTs were identified.

Four trials^69–72 were identified that compared the use of antibiotic prophylaxis with placebo or no antibiotic in stomach and duodenal surgery. Three reported wound infection outcomes for patients and one reported wound infections as a proportion of the overall number of wounds.⁶⁹

One RCT⁶⁹ included patients undergoing general surgery and randomised them to either cefaloridine (376 wounds) or no antibiotic (386 wounds). [EL = 1+] There was one wound infection in those undergoing gastric surgery with antibiotic prophylaxis (33 wounds) and six infections in the gastric surgery patients who did not receive antibiotics (30 wounds). This difference was not statistically significant (OR 0.13, 95% CI 0.01 to 1.11).

One RCT⁷⁰ included 83 patients undergoing surgery for high-risk gastroduodenal disease who were divided into two treatment arms, one of which received two doses of cefaloridine, the other no antibiotic. [EL = 1+] A further low-risk treatment arm was not considered here. No wound infections were found in the cefaloridine group (n = 41 patients) compared with 11 in the no antibiotic group (n = 42 patients). This difference was statistically significant (OR 0.03, 95% CI 0.00 to 0.58).

One RCT⁷¹ included 39 patients undergoing gastroduodenal surgery with a high postoperative risk. [EL = 1+] One infection was found in the cefamandole group (n = 19 patients) and seven were reported in the placebo group (n = 20 patients). This difference was statistically significant (OR 0.10, 95% CI 0.01 to 0.94).

One RCT⁷² included 68 consecutive patients undergoing elective surgery of the gastrointestinal tract. [EL = 1+] There were no infections in the antibiotic group (n = 32 patients), but 11 in the placebo group (n = 36 patients). This difference was statistically significant (OR 0.03, 95% CI 0.00 to 0.61).

A meta-analysis (Figure 5.12) of the three trials that reported wound infections in patients rather than as a proportion of all wounds found an overall statistically significant protective effect of antibiotics compared with placebo or no antibiotics (OR 0.05, 95% CI 0.01 to 0.22).

Figure 5.12

Meta-analysis of three trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in upper gastrointestinal tract surgery.

Hepatobiliary surgery

Laparoscopic gallbladder surgery

One systematic review and two RCTs were identified.

One relevant systematic review⁷⁴ and two more recently published RCTs from India⁷⁵ and Taiwan⁷⁶ were included.

The systematic review (six RCTs, n = 974 patients) compared the effect of antibiotic prophylaxis with that of placebo on wound infection in patients undergoing low-risk laparoscopic cholecystectomy. [EL = 1+] The pooled OR was 0.71, 95% CI 0.32 to 1.60, suggesting that there was no difference in wound infection incidence following antibiotic prophylaxis (12/567) or placebo administration (12/407) in laparoscopic cholecystectomy.

One trial⁷⁵ included 93 patients of ASA score 1 and 2 diagnosed as having gall stone disease undergoing laparoscopic cholecystectomy. [EL = 1+] Forty patients were randomised to receive 1.5 g cefuroxime in 100 ml saline at anaesthesia induction while 53 patients received normal saline similarly administered. There were three postoperative wound infections – one in the antibiotic group and two in the placebo group. This finding was not statistically significant (OR 0.65, 95% CI 0.06 to 7.47).

One trial⁷⁶ included 277 patients with symptomatic gallbladder stones or polyps disease with or without acute cholestasis who were candidates for laparoscopic cholecystectomy. [EL = 1+] One hundred and forty-one patients were randomised to receive 1 g cefazolin given at anaesthetic induction and 136 received 10 ml isotonic sodium chloride solution similarly. There were two infections, both of which occurred in the placebo group. This finding was not statistically significant (OR 0.19, 95% CI 0.01 to 4.00).

A meta-analysis (Figure 5.13) of all participants’ wound infection outcomes was performed that yielded a similar non-statistically significant result (OR 0.63, 95% CI 0.30 to 1.32).

Figure 5.13

Meta-analysis of three trials comparing the effect of antibiotic prophylaxis versus placebo on SSI incidence in hepatobiliary surgery.

Lower gastrointestinal tract surgery

Hernia repair

One systematic review and one RCT were identified.

A recently updated Cochrane systematic review⁷⁹ (12 RCTs, n = 6705 participants) was identified that evaluated antibiotic prophylaxis compared with placebo for prevention of wound infection in hernia repair. [EL = 1+]

Six trials (n = 2436 participants) used prosthetic material for hernia repair (hernioplasty) whereas the remaining studies (n = 4269 participants) did not (herniorrhaphy).

For hernioplasty, there were 17 wound infections among the patients who received prophylaxis (n = 1196 participants) compared with 37 in those receiving placebo (n = 1240 participants). This difference in wound infection incidence was statistically significant (OR 0.48, 95% CI 0.27 to 0.85).

For herniorrhaphy, there were 103 wound infections among the patients who received prophylaxis (n = 2932 participants) compared with 66 in those receiving placebo (n = 1337 participants). This difference in wound infection incidence did not quite reach statistical significance.

Overall, for both hernia repair methods, there were fewer wound infections among the participants who received prophylaxis (120/4128 participants) compared with those receiving placebo (103/2577 participants). This was a statistically significant finding (OR 0.64, 95% CI 0.48 to 0.85).

A further RCT⁸⁰ that was not referred to in the Cochrane review was also identified that compared the effect on wound infection of a single dose of amoxicillin and clavulanic acid with that of normal saline in elective open repair of inguinal hernia using mesh. [EL = 1+] There were five reports of wound infection in the antibiotic group (n = 190 participants) compared with nine in the placebo group (n = 189 participants). This was not a statistically significant difference (OR 0.54, 95% CI 0.18 to 1.64).

Pooling the results of this RCT with the review of hernioplasty demonstrated a statistically significant difference between the two groups (OR 0.49, 95% CI 0.30 to 0.81).

Pelvis surgery

Limb surgery

Open surgery for closed long bone fracture

One systematic review was identified.

One Cochrane systematic review⁸⁴ was identified that investigated the effect of antibiotic prophylaxis for surgery for proximal femoral and other closed long bone fractures. [EL = 1+] This review included trials examining wound infection for hip fracture as well as trials for long bone and other unspecified closed fractures. Only long bone and other unspecified closed fracture trials that examined the effect of prophylactic antibiotics versus placebo were included.

This left five trials available for inclusion in this review.

A meta-analysis (Figure 5.14) of three trials that considered the deep and superficial infection rates following multiple doses of a single antibiotic compared with placebo found that statistically significantly fewer wound infections occurred in the antibiotic group (RR 0.49, 95% CI 0.25 to 0.96, I² = 28.8%) overall. No statistically significant difference in either deep or superficial wound infection rates individually was observed.

Figure 5.14

Meta-analyses of three trials comparing the effect of prophylaxis with multiple doses of a single antibiotic versus placebo for deep, superficial and all SSI prevention in open surgery for closed long bone fractures.

A meta-analysis (Figure 5.15) of two trials that considered the deep and superficial infection rates following a single dose of one antibiotic as prophylaxis compared with placebo found that statistically significantly fewer wound infections occurred in the antibiotic group (RR 0.44, 95% CI 0.30 to 0.64) overall. Statistically significant differences in both deep and superficial wound infection rates individually were also observed.

Figure 5.15

Meta-analyses of two trials comparing the effect of single dose antibiotic prophylaxis versus placebo for deep, superficial and all SSI prevention in open surgery for closed long bone fractures.

Evidence statements – clean surgery

There is evidence that administration of antibiotics in craniotomy results in fewer wound infections compared with placebo. [EL = 1+]

There is evidence that administration of antibiotics in spinal surgery results in fewer wound infections compared with placebo. [EL = 1+]

There is evidence that pre- or perioperative antibiotics used as prophylaxis for breast cancer surgery results in fewer wound infections compared with placebo, although there is insufficient evidence to determine whether this effect is also true when antibiotics are compared with no prophylaxis. [EL = 1+]

There is insufficient evidence to determine the effect of antibiotic prophylaxis on wound infection in immediate breast reconstruction surgery with or without implants. [EL = 1+]

There is evidence that antibiotic prophylaxis during cardiac pacemaker surgery results in fewer infections compared with no antibiotic prophylaxis. [EL = 1+]

There is evidence that antibiotic prophylaxis reduces wound infection incidence in open-heart surgery compared with placebo. [EL = 1+]

There is evidence that antibiotic prophylaxis reduces wound infection incidence in thoracic surgery compared with placebo. [EL = 1+]

There is evidence that antibiotic prophylaxis reduces the incidence of wound infection compared with placebo in hernia repair in general and when hernioplasty is used. [EL = 1+]

However, currently there is evidence of no difference in wound infection rates when antibiotic prophylaxis or placebo is used in herniorrhaphy. [EL = 1+]

There is insufficient evidence available (owing to poor reporting) to determine the effect on wound infection in abdominal hysterectomy of antibiotic prophylaxis compared with placebo or no prophylaxis. [EL = 1−]

There is evidence that prophylactic antibiotics result in fewer wound infections in non-elective caesarean sections and for all patients undergoing a caesarean delivery. [EL = 1+]

There is currently evidence of fewer wound infections when antibiotic prophylaxis is given in elective caesarean delivery compared with placebo or no prophylaxis. [EL = 1+]

There is evidence from two meta-analyses that single and multidose antibiotic prophylaxis results in fewer wound infections than use of placebo or no prophylaxis in surgery for long bone and other unspecified closed fractures. [EL = 1+]

There is evidence that antibiotic prophylaxis results in fewer wound infections than placebo in surgery for hip fracture. [EL = 1+]

There is currently evidence of no difference in superficial wound infection rates when antibiotic prophylaxis or placebo is given in hip fracture surgery. However, there is some evidence that deep infection rate is reduced with antibiotic prophylaxis compared with placebo. [EL = 1+]

There is evidence from one trial that the use of antibiotics results in fewer wound infections than placebo in patients undergoing leg amputation for arteriosclerosis. [EL = 1+]

There is evidence that the use of systemic antibiotics results in fewer wound infections in patients undergoing peripheral arterial reconstruction. [EL = 1+]

Evidence statements – clean-contaminated surgery

There is evidence that there are fewer infections when patients are given antibiotic prophylaxis for contaminated/clean-contaminated head and neck cancer surgery compared with placebo. [EL = 1+]

There is evidence that antibiotic prophylaxis reduces wound infection incidence in gastroduodenal surgery compared with placebo or no antibiotic. [EL = 1+]

There is evidence that antibiotic prophylaxis reduces wound infection incidence in biliary tract surgery compared with placebo or no antibiotic. [EL = 1+]

There is evidence of no difference of effect of antibiotic prophylaxis compared with placebo for the prevention of wound infection in laparoscopic cholecystectomy. [EL = 1+]

There is evidence that systemic antibiotics result in fewer wound infections in surgery for appendicitis when compared with placebo. [EL = 1+]

There is evidence that there are fewer wound infections in surgery for appendicitis when single or multiple antibiotics given as a single dose preoperatively or perioperatively compared with placebo. [EL = 1+]

There is evidence that there are fewer wound infections in surgery for appendicitis when a single antibiotic is given at operation and subsequently given postoperatively as a single or multiple dose compared with placebo. [EL = 1+]

There is evidence that there are fewer wound infections in surgery for appendicitis when multiple antibiotics are given at operation and subsequently given postoperatively in multiple doses compared with placebo. [EL = 1+]

There is evidence that in children there is no difference of effect of antibiotic prophylaxis for non-complicated appendicitis. [EL = 1+] In children presenting with complicated appendicitis, there is evidence that antibiotics confer a protective effect against SSI. [EL = 1+]

There is evidence that antibiotic prophylaxis results in fewer wound infections than no antibiotic in colorectal surgery. [EL = 1+]

Evidence statement – contaminated surgery

There is evidence that antibiotic prophylaxis reduces the incidence of SSI in open reduction of mandibular fracture. [EL = 1+]

Evidence statement – dirty surgery

There is evidence that antibiotic prophylaxis results in fewer wound infections than placebo or no antibiotic in open limb fractures [EL = 1+]

Summary of evidence

There is evidence that prophylactic administration of antibiotics results in fewer SSIs compared with no other antibiotic treatment or with placebo in:

• craniotomy [EL = 1+]
• spinal surgery [EL = 1+]
• breast cancer surgery [EL = 1+]
• pacemaker insertion [EL = 1+]
• open-heart surgery [EL = 1+]
• thoracic surgery [EL = 1+]
• hernioplasty [EL = 1+]
• emergency and elective caesarean section [EL = 1+]
• long bone and other unspecified closed fractures [EL = 1+]
• hip fractures [EL = 1+]
• open limb fractures [EL = 1+]
• amputation [EL = 1+]
• peripheral arterial reconstruction [EL = 1+]
• head and neck surgery [EL = 1+]
• open reduction of mandibular fracture [EL = 1+]
• gastroduodenal surgery [EL = 1+]
• open biliary surgery [EL = 1+]
• appendicectomy [EL = 1+]
• colorectal surgery. [EL = 1+]

There is evidence to show that prophylactic antibiotics are not effective in:

• herniorrhaphy [EL = 1+]
• laparoscopic cholecystectomy. [EL = 1+]

There is insufficient evidence that prophylactic administration of antibiotics results in fewer SSIs compared with no other antibiotic treatment or with placebo in:

• breast reconstruction with or without implants [EL = 1+]
• abdominal hysterectomy (clean-contaminated) [EL = 1+]
• uncomplicated appendicectomy in children. [EL = 1+]

Health economics overview of evidence

Nineteen papers were identified for further review. Only three compared antibiotic prophylaxis with no antibiotic prophylaxis. One study was identified that compared a 24 hour prophylactic antibiotic regimen with a one-dose regimen.

One study⁸⁸ compared no antibiotic prophylaxis with antibiotic therapy. One study found no statistically significant difference in SSI rate in patients undergoing neck dissections⁸⁹ although this was based on retrospective data from 1977 to 1989. One study found a statistically significant difference in SSI rate in patients undergoing appendicectomies and colorectal operations.⁸⁸ One study was underpowered.⁹⁰ As none of these studies were carried out in the UK the costs are not generalisable to this setting.

The most recent study⁹¹ was a Brazilian study that used historical controls. A 24 hour prophylactic antibiotic regimen was compared with one-dose antibiotic prophylaxis given at anaesthesia induction. No statistically significant difference was found between SSI rate (2% and 2.1%, respectively, P = 0.67). A cost-minimisation analysis was thus carried out and using one dose of antibiotics was the lowest cost intervention. If similar SSI rates can be applied to a UK setting with reduced antibiotic prophylaxis, then a one-dose antibiotic prophylaxis protocol will be cost-saving compared with a 24 hour antibiotic regimen.

Timing of antibiotic prophylaxis

A prospective observational study using logistic regression to analyse data collected from patients undergoing elective clean or clean-contaminated surgery at a teaching hospital examined the timing of antibiotic prophylaxis administration as a risk factor for SSI.⁹² [EL = 2+] Patients were assigned to groups according to the time between their first dose of antibiotic prophylaxis and the initial surgical incision. The early group received prophylaxis 2–24 hours pre-incision, the preoperative group 0–2 hours pre-incision, the perioperative group up to 3 hours post-incision and the postoperative group received antibiotic prophylaxis 3–24 hours post-incision.

Forty-four of 2847 included patients (1.5%) developed SSI. Logistic regression demonstrated that there were statistically significantly more infections in the early and postoperative groups compared with the perioperative group (OR 4.3, 95% CI 1.8 to 10.4 and OR 5.8, 95% CI 2.4 to 13.8, respectively).

Results were further stratified according to the hour that prophylaxis was administered in relation to the time of surgery – the early group were excluded from this analysis. The lowest SSI rate occurred in patients receiving antibiotic prophylaxis 0–2 hours prior to surgery. A statistically significant trend was observed toward higher rates of infection with each successive hour that antibiotic administration was delayed after the surgical incision (z score = 2.00, P < 0.05 Wilcoxon test).

Evidence statement – timing of antibiotic prophylaxis

There is evidence that administration of antibiotic prophylaxis up to 2 hours preoperatively is associated with the lowest rates of infection in clean and clean-contaminated surgery. [EL = 2+]

GDG interpretation

Many of these studies used antibiotics that are not in current use and some were used for prolonged periods but comparable studies using modern antibiotics could not now be conducted ethically with the use of a placebo. In certain types of surgery (for example, orthopaedic prosthetic surgery) the GDG felt that even in the absence of adequate studies antibiotic prophylaxis would be appropriate.

Antibiotics are inexpensive and are likely to be cost-effective compared with no antibiotic prophylaxis if they prevent SSI as the cost of treating an SSI is approximately £3,500.

There is evidence that a single dose at the time of operation is effective.

Where antibiotic prophylaxis is administered, a repeat dose is only indicated when there is excessive blood loss or if surgery is unexpectedly prolonged.

If there is significant unexpected contamination encountered during an operation or existing infection then prophylaxis should be converted into a treatment regime.

The risk of adverse events, C. difficile diarrhoea, resistance and drug hypersensitivity must be considered. The GDG felt that the lack of evidence on the effectiveness of prophylaxis in the following procedures is insufficient reason to withhold antibiotic prophylaxis:

• breast reconstruction with/without implants
• abdominal hysterectomy (clean-contaminated)
• elective caesarean section
• uncomplicated appendicectomy in children.

In some of these groups (abdominal hysterectomy, elective caesarean section and appendicitis in children), unforeseen infection or contamination may be encountered that would make antibiotic prophylaxis appropriate. In breast reconstruction the presence of an implant may increase the risk of infection.