Updated Recommendations for Control of Surgical Site Infections

J. Wesley Alexander, MD, ScD; Joseph S. Solomkin, MD; Michael J. Edwards, MD


Annals of Surgery. 2011;253(6):1082-1093. 

In This Article

Systemic Prophylactic Antibiotics

A recent "meta-analysis of meta-analyses" involving 250 clinical trials and 4809 patients has provided an estimation of the relative benefit of systematic prophylactic antibiotics to reduce infection for 23 different types of surgery.[11] The type of antibiotic, timing, dosing, and type of procedure varied widely in this analysis, but the relative risk of developing infection for all types of operations with prophylactic systemic antibiotics versus no prophylactic antibiotics varied from 0.19 to 0.82, suggesting a generalized benefit regardless of the degree of contamination. Taken as a whole, the use of prophylactic systemic antibiotics decreased the incidence of wound infections by about one-half. This does not mean that prophylactic antibiotics should be used for every case, in as much as there are significant costs involved with their administration, they can have serious adverse effects and there is a risk of the development of antibiotic resistant pathogens or C. difficele colitis. Because of this, there has been reluctance to use prophylactic antibiotics in clean cases. However, prospective randomized studies have shown a clear benefit in clean elective operations such as hernia and breast procedures (SDC-137-141). Recent reports have also shown significant protection against infections in patients with a cesarean section (SDC-142-143). A review of the use of antimicrobial prophylaxis in colorectal surgery, including 182 trials with 3880 participants and 50 different antibiotics, showed a definite benefit of prophylactic antibiotics compared to a placebo or no treatment (RR = 0.30).[12] In that same study, combined therapy against both aerobic and anerobic organisms and combined oral and intravenous antibiotic prophylaxis compared to intravenous alone had significant benefits (RR, 0.41 and 0.74, respectively).

Timing of Administration

In earlier studies, the administration of antibiotics before or at the time of operation in experimental animals with contaminated wounds would effectively prevent wound infections, but this benefit was lost after 12 hours (SDC-144–145). In a further refinement, there was a progressive loss of effectiveness of antibiotics within the first 3 hours after inoculation in a model where organisms were injected intracutaneously in guinea pigs (SDC-146). Subsequent clinical studies confirmed that preoperative administration of antibiotics were superior to postoperative administration (SDC-147-149). These studies initially defined the "golden period" with recommendations to administer prophylactic antibiotics in the first 2 hours before surgery. In subsequent clinical studies, the rates of surgical wound infection have been more clearly related to the time of administration with respect to surgical incision with a U-shaped curve for the length of time either before or after the time of incision (SDC-150). One study showed that the incidence of an SSI was least when cefuroxime (and metronidazole in colorectal cases) was administered 30 to 60 minutes before incision (SDC-151). Another recent study showed that the least infections occurred when the antibiotics were given 10 to 20 minutes before incision (SDC-152). The most effective time for antibiotics to be given was 1 to 30 minutes before the incision in 1922 patients undergoing elective total hip arthroplasty in 11 hospitals (SDC-153). In a large multicenter report involving 4472 patients in 29 hospitals, the timing of the administration of perioperative antibiotics was studied involving cardiac, hip-knee arthroplasty and hysterectomy cases.[13] The best protection was seen when the antibiotic was given within the first 30 minutes before incision (SDC-Fig. 1 at https://links.lww.com/SLA/A121). Timing for vancomycin and other long-acting antibiotics is different. In a study of 2048 patients undergoing coronary artery bypass, the optimal time for administration of vancomycin was between 16 to 60 minutes (SDC-154). A further analysis of that study showed that when vancomycin was given at appropriate times, compared to inappropriate times, there was a savings of $4154/patient treated (SDC-155).

Despite the overwhelming evidence during the last 40 years, an analysis of 34,133 US Medicare patients undergoing cardiac, colorectal, or orthopedic surgery and hysterectomy from January 1 to November 30, 2001, showed only 55.7% of the patients received their prophylactic antibiotic within 1 hour before incision (SDC-156). In a large National Surveillance Program in the United States, compliance with the administration of antibiotics within 60 minutes of surgery rose from 47.6% from the beginning of 2002 to 73.1% in the last quarter of 2005 (SDC-157). With better surveillance, the administration of antibiotics within the optimal time, at one hospital increased from 40% in 1985 to 99% in 1998 (SDC-158).

Numerous studies have now shown that a single dose of antibiotics for chemoprophylaxis is as effective as multiple doses in a variety of surgical procedures, including gynecologic surgery (SDC-159-160), colorectal surgery (SDC-161), surgery for closed long bone fractures (SCD-162) and a variety of other operations (SDC-163-166). Other studies have shown that prolonged administration of antibiotics for longer than 24 hours adds no benefit in neurosurgical procedures (SDC-167), thoracic surgery (SDC-168), colorectal cancer surgery (SDC-169), gastric and colorectal surgery (SDC-170), cardiovascular surgery (SDC-171) and penetrating abdominal trauma (SDC-172-173). Prolonged use of antibiotics not only increases the costs directly but also may be associated with an increased risk of acquired antibiotic resistance (SDC-171).

Not all studies, however, have shown that single-dose antibiotics are as effective as multiple dose antibiotics. One study compared one dose with three doses of cefmetazole in elective colorectal surgery (SDC-174). The single dose group had an incisional SSI of 14.2% compared to 4.3% in the 3-dose group (P = 0.009). A prospective randomized study of 838 adult patients undergoing elective coronary artery bypass grafting showed that 8.3% of patients receiving a single dose and 3.6% of patients with antibiotic administered more than 24 hours developed an infection (P = 0.004) (SDC-175). The reasons for these infections occurring at a higher incidence with single dose administration may well be related to ineffective concentrations in the blood occurring before the end of closure for a variety of reasons (SDC-176-178).

Modifying Factors

It is generally acknowledged that systemic antibiotics are effective only if they maintain therapeutic concentrations in the wound throughout the period of potential contamination. There are several factors, which can modify the concentrations of antibiotics in the surgical wounds in addition to the ones already discussed. These include renal function, body weight, half-life of the antibiotic, use of cardiopulmonary bypass, use of transfusions for blood loss, aggressive fluid therapy, patient age, and rate of diffusion into the wound.

Renal Function

Most antibiotics are excreted by the kidney, some more rapidly than others. Therefore, renal function determines the half-life of serum concentrations to a major degree. Only a few antibiotics commonly used for prevention of infection are not affected by renal function including but not limited to clindamycin and metronidazole. Renal function is also decreased by the presence of hypotension, which will in turn increase half-life of renal-excreted drugs.


Obese patients require higher doses of antibiotics to achieve effective tissue levels. Morbidly obese patients given 2 rather than 1 gram of cefazolin preoperatively had higher tissue and serum concentration, which resulted in a reduction of wound infections from 16.5% to 5.6% (SDC-179). In another study, 230 patients with different degrees of obesity (BMI 40–49, BMI 50–59, BMI ≥ 60) were examined for blood concentrations of cefazolin after administration of 2 grams preoperatively (SDC-180). Blood and tissue levels were measured at the time of closure. Serum concentrations of the "resistance breakpoint" of 32 μg/ml or higher were obtained in 73%, 68%, and 52%, respectively. Therapeutic tissue levels at the time of closure were achieved in only 48.1%, 28.6%, and 10.0%, respectively from the lowest BMI to the highest BMI.


Drugs with a short half-life such as cefazolin do not have the ability to sustain therapeutic concentrations during long operations. In patients undergoing cardiopulmonary bypass, the prevalence of wound infections was lowest with prophylaxis with vancomycin compared to either cefazolin or cefamandol (SDC-181). Fluoroquinolones also have a prolonged half-life and could be useful for lengthy operations (SDC-182).

Selection of Drugs for Prophylaxis

In general, the antibiotic(s) used for prophylaxis must be effective against the expected pathogens. Typical drugs based on recommendations from a variety of sources (SDC-183-189) are shown in Table 1. Initial doses are shown in Table 2.

An important problem in selecting alternative agents for prophylaxis where there is possible allergy is the need, in all settings, for Gram-positive coverage for streptococci and methicillin-susceptible staphylococci. Untreated, these organisms have a high likelihood of causing infection due to a variety of virulence factors. If cephalosporins or penicillin derivatives cannot be used, clindamycin remains the best alternative. In addition, so called community acquired MRSA (CA-MRSA) are commonly susceptible to clindamycin, especially the USA300 strain that now predominates in the United States. This organism is replacing other strains of methicillin resistant staphylococci as the predominant pathogen in surgical site infections. Systemic aminoglycosides do not provide effective coverage for these organisms and quinolones have little activity against CA-MRSA. For this reason, where prophylaxis is needed for Gram-negative facultative and aerobic organisms, clindamycin should be added. In penicillin allergic patients undergoing procedures where Bacteroides fragilis and other anaerobes must be covered (colorectal surgery, appendectomy), gentamicin or a quinolone is recommended with the addition of clindamycin. There has been a continuing increase in Bacteroides fragilis resistance to clindamycin but this agent seems to retain sufficient activity to continue on as an effective prophylactic agent. The difficulty with replacing it with metronidazole, a more active anaerobic agent, is the loss of activity against Gram-positive organisms.

The use of prophylactic vancomycin is becoming increasingly common, and we now recommend this for operative procedures in which prosthetic materials are placed. It is also recommended for patients undergoing median sternotomy or craniotomy. This is, however, a recommendation without a large body of data on effective dosing. Vancomycin has been used extensively as a therapeutic agent for MRSA. This drug has an unusual distribution and does not behave as single compartment beta-lactams do. Steady state is not reached until the fourth dose on a q12 hourly dosing regimen. Therapeutic efficacy correlates most closely with a complex variable termed area under the curve. This is a variable determined by the height of the peak level and the trough level. In therapeutic settings, doses of 15 to 20 mg/kg (as actual body weight) given every 8 to 12 hour are recommended for most patients with normal renal function to achieve the suggested serum concentrations. In seriously ill patients, a loading dose of 25 to 30 mg/kg (based on actual body weight) can be used to facilitate rapid attainment of target trough serum vancomycin concentration.

Prophylactic therapy is based on having effective levels of the antimicrobial present throughout the operative period from skin incision to closure, and in some settings, one or two postoperative doses. It is not known, however, if dosing vancomycin at levels below 20 mg/kg will achieve these levels quickly. The primary problem is that dosing on an actual body weight basis will mean, for a 100 kg patient, 2 grams of vancomycin, infused at a maximal rate of 1 gm/hour. This will take 2 hours preoperatively, an excessive time that consumes substantial hospital resources. We, therefore, recommend beginning the infusion 1 hour before scheduled incision and continue the infusion intraoperatively until completed.


Many operations extend past the half life of the commonly used antibiotics. Current recommendations for redosing interval are given in Table 3, based upon renal function.


The administration of systemic prophylactic perioperative antibiotics is among the most important of the currently available methods to prevent wound infection. Except for vancomycin and the fluorquinolones, the most effective time for administration is within the first 30 minutes before the incision is made. The cephalosporins provide good early penetration into the wounds. Longer acting antibiotics like vancomycin and fluorquinolones should be given between 1 and 2 hours before the incision. Redosing of antibiotics is important for short-acting antibiotics and should be given approximately 3 hours after the incision is made. Dosage should be adjusted for large body size. The combined evidence shows no benefit of administration of an antibiotic after the wound is closed in the vast majority of cases where there is not massive contamination. The practice of routine administration of 3 doses of antibiotics should be abandoned. Adjustments for redosing should be made for renal function and rate of drug elimination.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.