Antimicrobial prophylaxis in the surgical patient☆

Postoperative infections 

Bacteria introduced into normally sterile body sites are the dominant cause of postoperative infection. Immunosuppression from perioperative stress, and from concomitant treatments such as blood transfusion, may also contribute to postoperative infection [1]. Although postoperative fungal infections remain much less common than bacterial infections, postoperative fungal infections are becoming more frequent, particularly in immunosuppressed patients [2], [3]. The most obvious and frequent location for a postoperative infection is the surgical site, but pneumonia is also a common postoperative infection in susceptible patients undergoing surgeries that entail endotracheal intubation or compromise the respiratory tract, thorax, or upper abdomen [4], [5]. Although we commonly talk about “wound” infections, these can be more explicitly referred to as surgical site infections and further characterized by the depth of infection and by the presence or absence of a foreign body or prosthetic material [3]. In addition to surgical site and pulmonary infections, bacteremia from these infections or from catheters can lead to sepsis and to endocarditis [6]. Finally, urinary tract infections occur in surgical patients, usually as a consequence of an indwelling urinary catheter [7].

Nonantimicrobial strategies for reducing postoperative infection 

In addition to antimicrobials and standard antiseptic surgical technique [3], some nonantimicrobial strategies have been demonstrated to reduce the incidence of postoperative infection, including maintaining normal body temperature [8], maintaining normal blood sugar levels [9], and hyperoxygenation [10].

Kurz et al [8] randomized 200 patients undergoing colorectal surgery to routine intraoperative thermal care or to supplemental warming. Blinded investigators evaluated the surgical sites for infection daily until discharge and at a 2-week follow-up clinic visit. Surgical site infection was defined as culture-positive purulent drainage. Final intraoperative core temperature was 34.7°C in patients randomized to routine care and 36.6°C in patients randomized to supplemental warming. Surgical site infection occurred in 18 (19%) of 96 patients randomized to routine care, but in only 6 (6%) of 104 randomized to supplemental warming (P=0.009).

Van den Berghe et al [9] randomly assigned adults who were admitted to the surgical intensive care unit (SICU) on a mechanical ventilator, to receive either conventional insulin treatment to maintain blood glucose below 210 mg/dL, or intensive insulin treatment to maintain blood glucose between 80 and 110 mg/dL. The study was terminated early, after the enrollment of 1548 patients, because 8.0% of patients receiving conventional treatment had expired in the SICU, compared with only 4.6% of patients receiving intensive treatment (P=0.04). The reduction in SICU mortality was principally caused by a reduction in multiple-organ failure with a proven septic focus in patients who were in the SICU for more than 5 days (20.2% conventional treatment, 10.6% intensive treatment, P=0.005). Intensive insulin treatment also reduced overall in-hospital mortality by 34%, and bloodstream infections by 46%.

Greif et al [10] randomly assigned 500 patients undergoing colorectal resection to receive either 30% or 80% inspired oxygen during the operation and for 2 hours afterward. Blinded investigators evaluated the surgical sites for infection daily until discharge and at a 2-week follow-up clinic visit. Surgical site infection was defined as culture-positive purulent drainage. Arterial oxygen saturation was normal in both groups; however, the arterial and subcutaneous partial pressure of oxygen was significantly higher in the patients randomized to 80% oxygen. Surgical site infection occurred in 28 (11.2%) of 250 patients randomized to 30% inspired oxygen, but in only 12 (5.2%) of 250 patients randomized to 80% inspired oxygen (P=0.01).

Although the three clinical trials described above can be criticized (for example, the control patients in Kurz's study had an extraordinarily high infection rate, and investigators in Van den Bergh's study were not blinded), the interventions have a solid physiologic basis and are supported by the findings of other studies in humans and other animals. Allogenic blood transfusion is another risk factor for postoperative infection [11], [12]; however, blood transfusion is also a general measure of severity of illness, and randomized trials restricting allogenic blood transfusion have failed to show a significant benefit [13]. In sum, assiduous maintenance of homeostasis, including body temperature, blood glucose, and tissue oxygenation in the perioperative period can significantly reduce postoperative infection.

Benefits and risks of antimicrobials 

The benefits of perioperative antimicrobial prophylaxis include a reduction in surgical site infection, pneumonia, sepsis, endocarditis, and urinary tract infection. The risks include allergic reactions to antimicrobials, toxic effects of antimicrobials, adverse interactions of antimicrobials with other medications, selection pressure for the emergence of antimicrobial-resistant organisms, and the cost of the antimicrobials. Therefore, the use of antimicrobial prophylaxis should be limited to those operations with high infection rates or serious consequences of infection [14], [15].

Principles of perioperative antimicrobial use 

Perioperative antimicrobial prophylaxis is directed against the most likely infecting organisms and does not have to cover every potential pathogen [16]. In surgeries not entering a chronically colonized body cavity, surgical site infections are most likely to be caused by skin organisms such as staphylococci and streptococci. Cefazolin is effective against these organisms and is therefore usually appropriate for these kinds of surgeries. Although prophylactic vancomycin might be appropriate for patients at high risk for infection with methicillin-resistant staphylococci, a randomized trial in a high-risk setting failed to show benefit [17], and vancomycin use promotes the emergence of resistant organisms, especially enterococci [18].

Antimicrobial prophylaxis for surgeries involving the lower gastrointestinal tract should cover gram-negative enteric bacteria and bowel anaerobes, especially Bacteroides fragilis. Cefoxitin and cefotetan are appropriate for such surgeries.

Third-generation cephalosporins cefotaxime, ceftriaxone, cefoperazone, ceftizoxime, ceftizoxime and fourth-generation cephalosporins such as cefepime are contraindicated for antimicrobial prophylaxis because: (1) most of them are less active than cefazolin against organisms likely to cause postoperative infection such as staphylococci, (2) they are active against organisms that rarely cause postoperative infection, (3) their use promotes the emergence of resistance organisms, especially enterococci, and (4) they are more expensive than more effective alternatives [19].

Penicillin allergy 

Patient report of penicillin allergy is notoriously unreliable. Approximately 85% of patients who report penicillin allergy do not have an allergy when assessed by skin testing [20]. Patients who are not penicillin-allergic by skin testing can safely receive penicillin [21]. Use of alternate antimicrobials, especially vancomycin, for surgical prophylaxis in patients reporting penicillin allergy increases cost and increases the prevalence of antimicrobial-resistant bacteria, such as vancomycin-resistant enterococci. A cost-effectiveness analysis that did not account for the increased prevalence of antimicrobial-resistant bacteria found that routine preoperative skin testing of cardiovascular surgery patients reporting penicillin allergy was more cost-effective than routine use of vancomycin [22]. A 6-month clinical trial of routine preoperative skin testing in elective orthopedic surgery patients reporting allergy to penicillins or to cephalosporins found a significant reduction in vancomycin use and no instances of immediate antimicrobial reaction [23]. Therefore, to minimize the cost of surgical antimicrobial prophylaxis, and to reduce the prevalence of antimicrobial-resistant bacteria, it is probably worth using skin testing when feasible to guide the choice of a prophylactic antimicrobial in patients reporting penicillin allergy prior to surgery.

Duration of antibiotic prophylaxis after surgery 

Usually, a single dose of antimicrobial within 1/2 hour prior to skin incision is effective infection prophylaxis [24]. An antimicrobial is measurably less effective if given more than 2 hours prior to skin incision [25], [26]. If the surgery lasts longer than 4 hours, or involves major blood loss, or the antimicrobial has a very short half-life (eg, cefoxitin) then additional doses of antimicrobial may be of benefit. Many surgeons continue antimicrobials for 2–3 days after surgery with the rationale that surgical wound drains and intravenous catheters might lead to bacterial seeding of the surgical site; however, there is evidence that this practice does not further decrease the risk of infection [27], [28], [29].

Considerations for specific surgeries (Table 1) 

Cardiac procedures 

Table 1. Antimicrobial prophylaxis for surgery

	Procedure	Likely pathogens	Antimicrobiala
	Cardiac: pacemaker or defibrillator insertion, and open heart, eg, coronary artery bypass and prosthetic valve	Staphylococci, corynebacteria, enteric gram-negative bacilli	Cefazolinb 1–2 gm IV, or cefuroximeb 1.5 gm IV, or vancomycinc 1 gm IV
	Gastrointestinal: appendectomy without perforation	Enteric gram-negative bacilli, anaerobes, enterococci	Cefoxitin 1–2 gm IV or cefotetan 1–2 gm IV
	Gastrointestinal: biliary tract, in a high-riskd patient only	Enteric gram-negative bacilli, enterococci, clostridia	Cefazolin 1–2 gm IV, or cefoxitin 1–2 gm IV, or cefotetan 1–2 gm IV
	Gastrointestinal: colorectal	Enteric gram-negative bacilli, anaerobes, enterococci	Oral: neomycin plus erythromycin basee Intravenous: cefoxitin 1–2 gm IV, or cefotetan 1–2 gm IV or cefazolin 1–2 gm IV plus metronidazole 0.5 gm IV
	Gastrointestinal: esophageal, gastroduodenal, in a high-riskf patient only	Enteric gram-negative bacilli, gram-positive cocci	cefazolin 1–2 gm IV
	Genitourinary: in a high-riskg patient only	Enteric gram-negative bacilli, enterococci	Oral: ciprofloxaxin 0.5 gm PO or trimethoprim-sulfamethoxazole 160–800 mg PO
			Intravenous: ciprofloxacin 0.4 gm IV trimethoprim-sulfamethoxazole 160–800 mg IV
	Gynecologic/obstetric: abortion, first trimester, in a high-riskh patient only	Enteric gram-negative bacilli, anaerobes, enterococci, group B strep	Oral: doxycycline 300 mg POi Intravenous: aqueous penicillin G 2 million units IV
	Gynecologic/obstetric: abortion, second trimester	Enteric gram-negative bacilli, anaerobes, enterococci, group B strep	Cefazolin 1 gm IV
	Gynecologic/obstetric: cesarean section, in a high-riskj patient only	Enteric gram-negative bacilli, anaerobes, enterococci, group B strep	Cefazolin 1 gm IV after cord clamping
	Gynecologic/obstetric: hysterectomy: vaginal or abdominal	Enteric gram-negative bacilli, anaerobes, enterococci, group B strep	Cefoxitin 1–2 gm IV, or cefotetan 1–2 gm IV, or cefazolin 1–2 gm IV
	Head and neck: with incision through oral or pharyngeal mucosa	Oral anaerobes, enteric gram-negative bacilli, staphylococci	Ampicillin-sulbactam 1.5–3 gm IV or clindamycin 600–900 mg IV, plus gentamicin 1.5 mg/kg IV or cefazolin 1–2 gm IV
	Neurologic: craniotomy	Staphylococci	Cefazolin 1–2 gm IV or vancomycinc 1 gm IV
	Ophthalmic	Staphylococci, streptococci, enteric gram-negative bacilli, Pseudomonas aeruginosa	Topical drops over 2–24 hours: gentamicin, or tobramycin, or ciprofloxacin, or ofloxacin, or neomycin-gramicidin-polymyxin B
			Subconjunctival: cefazolin 100 mg
	Orthopedic	Staphylococci	Cefazolin 1–2 gm IV or vancomycinc 1 gm IV
	Thoracic: noncardiac	Staphylococci, streptococci, enteric gram-negative bacilli	Cefazolin 1–2 gm IV, or cefuroxime 1.5 gm IV, or vancomycinc 1 gm IV
	Vascular: arterial repair, prosthetic material, abdominal aorta	Staphylococci, streptococci	Cefazolin 1–2 gm IV or vancomycinc 1 gm IV
	Vascular: groin incision, leg amputation for arterial insufficiency	Staphylococci, streptococci, enteric gram-negative bacilli, clostridia	Cefoxitin 1–2 gm IV or vancomycinc 1 gm IV

(Adapted from Antimicrobial prophylaxis in surgery. Med Lett Drugs Ther 2001;43:92–97; with permission.)

a Give as a single intravenous dose completed 1-half hr prior to the first skin incision. For prolonged procedures additional doses may be required at usual dosing interacts. b Some consultants recommend an additional dose when patients are removed from bypass during open-heart surgery. c For settings in which methicillin-resistant S aureus and S epidermidis frequently cause surgical site infection, or for patients allergic to cephalosporins. Rapid IV administration may cause hypotension, which can be exacerbated by induction of anesthesia. Infuse slowly and treat with diphenhydramine (Benadryl®, Parke-Davis, and others). For procedures in which enteric gram-negative bacilli are likely pathogens, such as vascular surgery in the groin, include cefazolin or cefuroxime for patients not allergic to cephalosporins. d Age >70 years, acute cholecystitis, nonfunctioning gall bladder, obstructive jaundice, or common duct stones. e After appropriate diet and catharsis, one gram each at 1 PM, 2 PM, and 11 PM the day before an 8 AM operation. f Morbid obesity, esophageal obstruction, decreased gastric acidity, or decreased gastrointestinal motility. g Urine culture positive or unavailable, preoperative bladder catheter, transrectal prostatic biopsy. h Patients with previous pelvic inflammatory disease, previous gonorrhea, or multiple sex partners. i Divided into 100 mg 1 hr before the abortion, and 200 mg 1 hr after. j Active labor or premature rupture of membranes.

Antimicrobial prophylaxis with cefazolin reduces the risk of infection after cardiac procedures, including the transvenous pacemaker placement [30]. In institutions with a high risk of infection with methicillin-resistant staphylococci, vancomycin may be an appropriate alternative, though a randomized trial in a high-risk setting failed to show benefit [17]. An inception cohort study demonstrated a reduction in sternal wound infection after cardiac surgery in patients treated with intranasal mupirocin before and after surgery [31].

The rationale for antimicrobial prophylaxis 

Endocarditis is an uncommon yet life-threatening infection. It usually occurs in people with abnormal or prosthetic heart valves and requires bacteremia with organisms that can reside on the valves. The source of the bacteremia can be inapparent or can be caused by a focal infection such as cellulitis, an abscess, or pneumonia. Some surgical and dental procedures can produce transient bacteremia, and, though the great majority of endocarditis is not attributable to an invasive procedure [66], periprocedure antimicrobials are administered to patients at risk with the goal of reducing the risk for this serious complication. Under the aegis of The American Heart Association, a panel of experts has devised recommendations for the use of antimicrobial prophylaxis to reduce the risk of bacterial endocarditis after invasive procedures [6]. Despite these guidelines, antibiotic prophylaxis against endocarditis is frequently both overused and underused [67].

At risk patients 

Patients at high risk for endocarditis include those with prosthetic valves, prior endocarditis, or complex cyanotic heart disease (Table 2). Patients at moderate risk include those with congenital cardiac malformations other than complex cyanotic heart disease; with rheumatic and other acquired structurally abnormal valves; hypertrophic cardiomyopathy; and with mitral valve prolapse including an abnormal, regurgitant mitral valve (Table 2). Patients at no greater risk than the general population include those with isolated secundum atrial septal defects, surgically repaired atrial and ventricular defects (more than 6 months after successful repair), surgically repaired patent ductus arteriosus (more than 6 months after successful repair), prior coronary artery bypass, implanted cardiac pacemakers and defibrillators, and benign murmurs (Table 2).

Table 2. Patient risk categories for endocarditis

	High risk
	Prosthetic valves, including bioprosthetic and homograft valves
	Prior endocarditis
	Complex cyanotic heart disease
	Surgically constructed systemic-pulmonary shunts
	Moderate risk
	Congenital cardiac malformations other than complex cyanotic heart disease
	Rheumatic and other acquired, structurally abnormal valves
	Hypertrophic cardiomyopathy
	Mitral valve prolapse with a thickened or continuously regurgitant valve
	Low risk (no greater risk than the general population)
	Isolated secundum atrial septal defects
	Surgically repaired: atrial septal defects, ventricular septal defects, or patent ductus arteriosus (more than 6 months after successful repair)
	Prior coronary artery bypass
	Implanted cardiac pacemakers and defibrillators
	Prior Kawasaki's disease or rheumatic fever without valve dysfunction
	Mitral valve prolapse without a thickened or continuously regurgitant valve
	Benign murmurs

(Adapted from Dajani AS, Taubert KA, Wilson W, et al. Prevention of bacterial endocarditis. Recommendations by The American Heart Association. JAMA 1997;277:1794–1801; with permission.)

At risk procedures 

Any procedure involving infected tissue at the surgical site is associated with a significant risk of bacteremia. Upper aerodigestive tract procedures associated with some risk of bacteremia include dental extractions and implants, tonsillectomy, rigid bronchoscopy, esophageal sclerotherapy and dilation, biliary tract surgery, and other procedures violating the oral or intestinal mucosa (Table 3). Procedures with negligible risk of bacteremia include restorative dentistry, local anesthetic injection, intracanal endodontistry, suture removal, adjustment of orthodontic appliances, endotracheal intubation, flexible bronchoscopy, tympanostomy, transesophageal echocardiography, and endoscopy without biopsy.

Table 3. Procedure risk categories for endocarditis in uninfected patients

	Upper aerodigestive tract procedures with some risk of bacteremia
	Procedures violating the oral, esophageal or intestinal mucosa, including:
	Prophylactic cleaning with anticipated bleeding
	Dental extractions and implants
	Periodontal surgery, scaling, planing, and probing
	Subgingival or intraligamentary periodontic manipulation or injection
	Endodontic surgery beyond the apex
	Initial placement of orthodontic bands, but not brackets
	Tonsillectomy—adenoidectomy
	Rigid bronchoscopy
	Esophageal sclerotherapy and dilation
	Biliary tract surgery including ERCP with biliary obstruction
	Upper aerodigestive tract procedures with negligible risk of bacteremiaa
	Restorative dentistry
	Local anesthetic injection not into dental ligaments
	Intracanal endodontistry
	Suture removal
	Adjustment of orthodontic appliances
	Endotracheal intubation
	Flexible bronchoscopy
	Tympanostomy tube insertion
	Transesophageal echocardiography
	Gastrointestinal endoscopy with or without biopsy
	Lower gastrointestinal and genitourinary tract procedures with some risk of bacteremia
	Prostate surgery
	Cystoscopy
	Urethral dilation
	Lower gastrointestinal and genitourinary tract procedures with negligible risk of bacteremiaa
	Vaginal hysterectomy
	Vaginal delivery
	Cesarean section
	Uterine dilation and curettage
	Therapeutic abortion
	Tubal ligation
	Insertion and removal of intrauterine devices
	Urethral catheterization

ERCP=endoscopic retrograde cholangiopancreatography. (Adapted from Dajani AS, Taubert KA, Wilson W, et al. Prevention of bacterial endocarditis. Recommendations by The American Heart Association. JAMA 1997; 277:1794–1801; with permission.)

a Any procedure involving infected tissue at the surgical site is asociated with a significant risk of bacteremia.

Lower gastrointestinal and genitourinary tract procedures associated with some risk of bacteremia include prostate surgery, cystoscopy, and urethral dilation. Procedures with negligible risk include vaginal hysterectomy, normal vaginal delivery, cesarean section, uterine dilation and curettage, therapeutic abortion, tubal ligation, insertion and removal of intrauterine devices, and urethral catheterization (Table 3).

Antimicrobials 

Antimicrobials used in endocarditis prophylaxis are aimed at the most likely causative organisms (Table 4). In upper aerodigestive tract procedures, viridians (alpha-hemolytic) streptococci are the most likely causative organisms, and in lower gastrointestinal and genitourinary tract procedures enterococci (Enterococcus faecalis) are the most likely causative organisms. Oral amoxicillin or intravenous ampicillin is usually the antimicrobials of choice. In upper aerodigestive tract procedures, alternative antimicrobials for penicillin-allergic patients include clindamycin, cephalexin, cephadroxil, azithromycin, and clarithromycin. Erythromycin is no longer listed as an alternative because of the availability of better-tolerated alternatives. In lower gastrointestinal and genitourinary tract procedures, vancomycin is the primary alternative to ampicillin. In the highest-risk patients undergoing lower gastrointestinal and genitourinary tract procedures, combination antimicrobial prophylaxis including gentamicin is used against enterococci because enterococci are frequently resistant to antimicrobials.

Table 4. American Heart Association Recommendations for Endocarditis Prophylaxis

	Upper Aerodigestive Tract Procedure with Some Risk of Bacteremia (Table 3)
	High- or moderate-risk patient (Table 2)
	No contraindication to penicillins:
	Oral: amoxicillin 2 gm PO 1 hr prior to the procedure
	Intravenous: ampicillin 2 gm IV 1/2 hr prior to the procedure
	Penicillins contraindicated:
	Oral: clindamycin 600 mg PO 1 hr prior to the procedure
	or
	cephalexin 2 gm PO 1 hr prior to the procedure
	or
	cephadroxil 2 gm PO 1 hr prior to the procedure
	or
	azithromycin 500 mg PO 1 hr prior to the procedure
	or
	clindamycin 500 mg PO 1 hr prior to the procedure
	Intravenous: clindamycin 600 mg IV 1/2 hr prior to the procedure
	or
	cefazolin 1 gm IV 1/2 hr prior to the procedure
	Lower gastrointestinal or genitourinary tract procedure with some risk of bacteremia (Table 3)
	High Risk Patient (Table 2)
	No Contraindication to Penicillins:
	ampicillin 2 gm IV 1/2 hr prior to the procedure
	plus
	Gentamicin 1.5 mg/kg IV 1/2 hr prior to the procedure
	and, 6 hr later
	Amoxicillin 1 gm PO, or ampicillin 1 gm IV
	Penicillins contraindicated:
	vancomycin 1 gm IV 1.5 hr prior to the procedure
	plus
	Gentamicin 1.5 mg/kg IV 1/2 hr prior to the procedure
	Moderate-risk patient (Table 2)
	No contraindication to penicillins:
	Oral: amoxicillin 2 gm PO 1 hr prior to the procedure
	Intravenous: ampicillin IV 2 gm 1/2 hr prior to the procedure
	Penicillins contraindicated:
	Vancomycin 1 gm IV 1.5 hr prior to the procedure

(Adapted from Dajani AS, Taubert KA, Wilson W, et al. Prevention of bacterial endocarditis. Recommendations by The American Heart Association. JAMA; 1997, 277:1794–1801, with permission.)

Special considerations 

As with antimicrobial prophylaxis against postoperative infection, prophylaxis against endocarditis would be expected to be most effective if the antimicrobial is given within an hour prior to the procedure. Patients who chronically take antimicrobials, such as those who take penicillin for secondary prevention of rheumatic fever, may be colonized with bacteria resistant to penicillins. For these patients, it is appropriate to use an antimicrobial with a different mechanism of action than the one taken chronically. For example, for the patient who is taking penicillin to prevent rheumatic fever, either clindamycin or azithromycin would be an appropriate alternative.

Summary 

The primary prophylactic measure against postoperative infection is antiseptic technique in patient preparation, during surgery, and in postoperative patient care. Antimicrobial prophylaxis against postoperative infection is not indicated for procedures with a low infection rate because the expected benefit of antimicrobial treatment is less than the risk of an adverse medication reaction. Antimicrobial prophylaxis has been demonstrated to be of greater benefit than risk in some procedures with higher infection rates; however, because the problem is complex and the data are limited, extrapolating these findings to the practitioner's setting and the individual patient remains a challenge (Table 1).

Although antimicrobial prophylaxis for bacterial endocarditis is not effective for most patients, the seriousness of the potential infection has driven the creation of guidelines recommending prophylaxis for at-risk patients undergoing at-risk procedures. Applying these guidelines appropriately could help to reduce unwarranted use of antimicrobials.

In the prophylactic use of antimicrobials, as in many medical interventions, the difficulty is balancing the risks of the intervention with the potential benefits. Although we do not have either the randomized, controlled trials or the detailed, patient-specific information to estimate this balance precisely, there are general guidelines to help the clinician choose treatment for most patients.