Surgical site infections (SSIs) are infections of the tissues, organs, or spaces exposed by surgeons during performance of an invasive procedure. SSIs are classified into incisional and organ/space infections, and the former are further subclassified into superficial (limited to skin and subcutaneous tissue) and deep incisional categories.38,39 The development of SSIs is related to three factors: (a) the degree of microbial contamination of the wound during surgery, (b) the duration of the procedure, and (c) host factors such as diabetes, malnutrition, obesity, immune suppression, and a number of other underlying disease states. Table 6-6 lists risk factors for development of SSIs. By definition, an incisional SSI has occurred if a surgical wound drains purulent material or if the surgeon judges it to be infected and opens it.
Table 6-5Prophylactic use of antibiotics (adapted from ref 25) ||Download (.pdf) Table 6-5 Prophylactic use of antibiotics (adapted from ref 25)
ALTERNATIVE (E.G., PENICILLIN ALLERGIC)
Gastroduodenal area; small intestine, nonobstructed
Clindamycin or vancomycin + aminoglycoside or aztreonam or fluoroquinolone
Biliary tract: open procedure, laparoscopic high risk
Cefazolin, cefoxitin, cefotetan, ceftriaxone, ampicillin-sulbactam,
Clindamycin or vancomycin + aminoglycoside or aztreonam or fluoroquinolone
Metronidazole + aminoglycoside or fluoroquinolone
Biliary tract: laparoscopic low risk
Cefoxitin, cefotetan, cefazolin + metronidazole
Clindamycin + aminoglycoside or aztreonam or fluoroquinolone
Metronidazole + aminoglycoside or flouroquinolone
Colorectal surgery, obstructed small intestine
Cefazolin or ceftriaxone plus metronidazole, Ertapenem, cefoxitin, cefotetan, ampicillin-sulbactam
Clindamycin + aminoglycoside or aztreonam or fluoroquinolone, metronidazole + aminoglycoside or fluoroquinolone
Head and neck; clean contaminated
Cefazolin or cefuroxime + metronidazole, ampicillin-sulbactam
Surgical wounds are classified based on the presumed magnitude of the bacterial load at the time of surgery (Table 6-7).40 Clean wounds (class I) include those in which no infection is present; only skin microflora potentially contaminate the wound, and no hollow viscus that contains microbes is entered. Class I D wounds are similar except that a prosthetic device (e.g., mesh or valve) is inserted. Clean/contaminated wounds (class II) include those in which a hollow viscus such as the respiratory, alimentary, or genitourinary tracts with indigenous bacterial flora is opened under controlled circumstances without significant spillage of contents.
Table 6-6Risk factors for development of surgical site infections ||Download (.pdf) Table 6-6 Risk factors for development of surgical site infections
Chronic inflammatory process
Peripheral vascular disease
Chronic skin disease
Carrier state (e.g., chronic Staphylococcus carriage)
Open compared to laparoscopic surgery
Poor skin preparation
Contamination of instruments
Inadequate antibiotic prophylaxis
Local tissue necrosis
Prolonged hospitalization (leading to nosocomial organisms)
Resistance to clearance (e.g., capsule formation)
While elective colorectal cases have classically been included as class II cases, a number of studies in the last decade have documented higher SSI rates (9% to 25%).41,42,43 One study identified two-thirds of infections presenting after discharge from hospital, highlighting the need for careful follow-up of these patients.41 Infection is also more common in cases involving entry into the rectal space.42 In a recent single center quality improvement study using a multidisciplinary approach, one group of clinicians has demonstrated the ability to decrease SSI from 9.8% to 4.0%.43
Contaminated wounds (class III) include open accidental wounds encountered early after injury, those with extensive introduction of bacteria into a normally sterile area of the body due to major breaks in sterile technique (e.g., open cardiac massage), gross spillage of viscus contents such as from the intestine, or incision through inflamed, albeit nonpurulent tissue. Dirty wounds (class IV) include traumatic wounds in which a significant delay in treatment has occurred and in which necrotic tissue is present, those created in the presence of overt infection as evidenced by the presence of purulent material, and those created to access a perforated viscus accompanied by a high degree of contamination. The microbiology of SSIs is reflective of the initial host microflora such that SSIs following creation of a class I wound are invariable, due solely to skin microbes found on that portion of the body, while SSIs subsequent to a class II wound created for the purpose of elective colon resection may be caused by either skin microbes or colonic microflora, or both.
In the United States, hospitals are required to conduct surveillance for the development of SSIs for a period of 30 days after the operative procedure.44 Such surveillance has been associated with greater awareness and a reduction in SSI rates, probably in large part based upon the impact of observation and promotion of adherence to appropriate care standards. Beginning in 2012, all hospitals receiving reimbursement from the Center for Medicare and Medicaid Services are required to report SSIs.
A recent refinement of risk indexes has been implemented through the National Healthcare Safety Network, a secure, web-based system of surveillance utilized by the Centers for Disease Control and Prevention for surveillance of health care associated infections. This refinement utilized data reported from 847 hospitals in nearly one million patients over a two- year period to develop procedure-specific risk indices for SSIs.45
SSIs are associated with considerable morbidity and occasional lethality, as well as substantial health care costs and patient inconvenience and dissatisfaction.46 For that reason, surgeons strive to avoid SSIs by using the maneuvers described in the previous section. Also, the use of prophylactic antibiotics may serve to reduce the incidence of SSI rates during certain types of procedures. For example, it is well accepted that a single dose of an antimicrobial agent should be administered immediately prior to commencing surgery for class I D, II, III, and IV types of wounds. It seems reasonable that this practice should be extended to patients in any category with high National Nosocomial Infection Surveillance (NNIS) scores, although this remains to be proven. Thus, the utility of prophylactic antibiotics in reducing the rate of wound infection subsequent to clean surgery remains controversial, and these agents should not be employed under routine circumstances (e.g., in healthy young patients). However, because of the potential dire consequences of a wound infection after clean surgery in which prosthetic material is implanted into tissue, patients who undergo such procedures should receive a single preoperative dose of an antibiotic.
A number of health care organizations within the United States have become interested in evaluating performance of hospitals and physicians with respect to implementing processes that support delivery of standard of care. One major process of interest is reduction in SSIs, since the morbidity (and subsequent cost) of this complication is high. Several of these organizations are noted in Table 6-8. Appropriate guidelines in this area incorporating the principles discussed previously have been developed and disseminated.47 However, observers have noted that adherence to these guidelines has been poor.48 Most experts believe that better adherence to evidence-based practice recommendations and implementing systems of care with redundant safeguards will result in reduction of surgical complications and better patient outcomes. More important, the Center for Medicare and Medicaid Services, the largest third party insurance payer in the United States, has required reporting by hospitals of many processes related to reduction of surgical infections, including appropriate use of perioperative antibiotics. This information, which is currently reported publicly by hospitals, has led to significant improvement in reported rates of these process measures. However, the effect of this approach on the incidence of SSIs is not known at this time.
Table 6-7Wound class, representative procedures, and expected infection rates ||Download (.pdf) Table 6-7 Wound class, representative procedures, and expected infection rates
EXAMPLES OF CASES
EXPECTED INFECTION RATES
Clean (class I)
Hernia repair, breast biopsy specimen
Clean/contaminated (class II)
Cholecystectomy, elective GI surgery (not colon)
Clean/contaminated (class II)
Contaminated (class III)
Penetrating abdominal trauma, large tissue injury, enterotomy during bowel obstruction
Dirty (class IV)
Perforated diverticulitis, necrotizing soft tissue infections
Surgical management of the wound also is a critical determinant of the propensity to develop a SSI. In healthy individuals, class I and II wounds may be closed primarily, while skin closure of class III and IV wounds is associated with high rates of incisional SSIs (~25% to 50%). The superficial aspects of these latter types of wounds should be packed open and allowed to heal by secondary intention, although selective use of delayed primary closure has been associated with a reduction in incisional SSI rates.49 It remains to be determined whether NNIS-type stratification schemes can be employed prospectively in order to target specific subgroups of patients which will benefit from the use of prophylactic antibiotic and/or specific wound management techniques. One clear example based on CoGeNT data from clinical trials is that class III wounds in healthy patients undergoing appendectomy for perforated or gangrenous appendicitis can be primarily closed as long as antibiotic therapy directed against aerobes and anaerobes is administered. This practice leads to SSI rates of approximately 3% to 4%.50
Recent investigations have studied the effect of additional maneuvers in an attempt to further reduce the rate of SSIs. The adverse effects of hyperglycemia on WBC function have been well described.51 A number of recent studies in patients undergoing several different types of surgery describe increased risk of SSI in patients with hyperglycemia.52,53 Although randomized trials have not been performed, it is recommended that clinicians maintain appropriate blood sugar control in patients in the perioperative period to minimize the occurrence of SSI.
The respective effects of body temperature and the level of inhaled oxygen during surgery on SSI rates also have been studied, and both hypothermia and hypoxia during surgery are associated with a higher rater of SSIs. Although an initial study provided evidence that patients who received high levels of inhaled oxygen during colorectal surgery developed fewer SSIs,54 a recent meta-analysis suggest that the overall benefit is small and may not warrant use.55 Further evaluation via multicenter studies is needed prior to implementation of hyperoxia as standard therapy, but it is clear that intraoperative hypothermia and hypoxia should be prevented.
Effective therapy for incisional SSIs consists solely of incision and drainage without the additional use of antibiotics. Antibiotic therapy is reserved for patients in whom evidence of significant cellulitis is present, or who concurrently manifest a systemic inflammatory response syndrome. The open wound often is allowed to heal by secondary intention, with dressings being changed twice a day. The use of topical antibiotics and antiseptics to further wound healing remains unproven, although anecdotal studies indicate their potential utility in complex wounds that do not heal with routine measures.56 Despite a paucity of prospective studies,57 vacuum-assisted closure is increasingly used in management of large, complex open wounds and can be applied to wounds in locations that are difficult to manage with dressings (Fig. 6-2). One also should consider obtaining wound cultures in patients who develop SSIs and whom have been hospitalized or reside in long-term care facilities due to the increasing incidence of infection caused by multidrug resistant organisms. The treatment of organ/space infections is discussed in the following section.
Negative pressure wound therapy in a patient after amputation for wet gangrene (A), and in a patient with enterocutaneous fistula (B). It is possible to adapt these dressings to fit difficult anatomy and provide appropriate wound care while reducing frequency of dressing change. It is important to evaluate the wound under these dressings if patient demonstrates signs of sepsis with an unidentified source, since typical clues of wound sepsis such as odor and drainage are hidden by the suction apparatus.
Microbial contamination of the peritoneal cavity is termed peritonitis or intra-abdominal infection, and is classified according to etiology. Primary microbial peritonitis occurs when microbes invade the normally sterile confines of the peritoneal cavity via hematogenous dissemination from a distant source of infection or direct inoculation. This process is more common among patients who retain large amounts of peritoneal fluid due to ascites, and among those individuals who are being treated for renal failure via peritoneal dialysis. These infections invariably are monomicrobial and rarely require surgical intervention. The diagnosis is established based on identification of risk factors as noted previously, physical examination that reveals diffuse tenderness and guarding without localized findings, absence of pneumoperitoneum on an imaging study, the presence of more than 100 WBCs/mL, and microbes with a single morphology on Gram’s stain performed on fluid obtained via paracentesis. Subsequent cultures typically will demonstrate the presence of gram positive organisms in patients undergoing peritoneal dialysis. In patients without this risk factor organisms can include E. coli, K. pneumoniae, pneumococci, and others, although many different pathogens can be causative. Treatment consists of administration of an antibiotic to which the organism is sensitive; often 14 to 21 days of therapy are required. Removal of indwelling devices (e.g., a peritoneal dialysis catheter or a peritoneovenous shunt) may be required for effective therapy of recurrent infections.
Secondary microbial peritonitis occurs subsequent to contamination of the peritoneal cavity due to perforation or severe inflammation and infection of an intra-abdominal organ. Examples include appendicitis, perforation of any portion of the gastrointestinal tract, or diverticulitis. As noted previously, effective therapy requires source control to resect or repair the diseased organ; débridement of necrotic, infected tissue and debris; and administration of antimicrobial agents directed against aerobes and anaerobes.58 This type of antibiotic regimen should be chosen because in most patients the precise diagnosis cannot be established until exploratory laparotomy is performed, and the most morbid form of this disease process is colonic perforation, due to the large number of microbes present. A combination of agents or single agents with a broad spectrum of activity can be used for this purpose; conversion of a parenteral to an oral regimen when the patient’s ileus resolves provides results similar to those achieved with intravenous antibiotics. Effective source control and antibiotic therapy is associated with low failure rates and a mortality rate of approximately 5% to 6%; inability to control the source of infection is associated with mortality greater than 40%.59
The response rate to effective source control and use of appropriate antibiotics has remained approximately 70% to 90% over the past several decades.60 Patients in whom standard therapy fails typically develop one or more of the following: an intra-abdominal abscess, leakage from a gastrointestinal anastomosis leading to postoperative peritonitis, or tertiary (persistent) peritonitis. The latter is a poorly understood entity that is more common in immunosuppressed patients in whom peritoneal host defenses do not effectively clear or sequester the initial secondary microbial peritoneal infection. Microbes such as Enterococcus faecalis and faecium, Staphylococcus epidermidis, Candida albicans, and Pseudomonas aeruginosa commonly are identified, typically in combination, and their presence may be due to their lack of responsiveness to the initial antibiotic regimen, coupled with diminished activity of host defenses. Unfortunately, even with effective antimicrobial agent therapy, this disease process is associated with mortality rates in excess of 50%.61
Formerly, the presence of an intra-abdominal abscess mandated surgical reexploration and drainage. Today, the vast majority of such abscesses can be effectively diagnosed via abdominal computed tomographic (CT) imaging techniques and drained percutaneously. Surgical intervention is reserved for those individuals who harbor multiple abscesses, those with abscesses in proximity to vital structures such that percutaneous drainage would be hazardous, and those in whom an ongoing source of contamination (e.g., enteric leak) is identified. The necessity of antimicrobial agent therapy and precise guidelines that dictate duration of catheter drainage have not been established. A short course (3 to 7 days) of antibiotics that possess aerobic and anaerobic activity seems reasonable, and most practitioners leave the drainage catheter in situ until it is clear that cavity collapse has occurred, output is less than 10 to 20 mL/d, no evidence of an ongoing source of contamination is present, and the patient’s clinical condition has improved.
Hepatic abscesses are rare, currently accounting for approximately 15 per 100,000 hospital admissions in the United States. Pyogenic abscesses account for approximately 80% of cases, the remaining 20% being equally divided among parasitic and fungal forms.62 Formerly, pyogenic liver abscesses mainly were caused by pylephlebitis due to neglected appendicitis or diverticulitis. Today, manipulation of the biliary tract to treat a variety of diseases has become a more common cause, although in nearly 50% of patients no cause is identified. The most common aerobic bacteria identified in recent series include E coli, K pneumoniae, and other enteric bacilli, enterococci, and Pseudomonas spp., while the most common anaerobic bacteria are Bacteroides spp., anaerobic streptococci, and Fusobacterium spp. Candida albicans and other related yeast cause the majority of fungal hepatic abscesses. Small (<1 cm), multiple abscesses should be sampled and treated with a 4 to 6 week course of antibiotics. Larger abscesses invariably are amenable to percutaneous drainage, with parameters for antibiotic therapy and drain removal similar to those mentioned previously. Splenic abscesses are extremely rare and are treated in a similar fashion. Recurrent hepatic or splenic abscesses may require operative intervention—unroofing and marsupialization or splenectomy, respectively.
Secondary pancreatic infections (e.g., infected pancreatic necrosis or pancreatic abscess) occur in approximately 10% to 15% of patients who develop severe pancreatitis with necrosis. The surgical treatment of this disorder was pioneered by Bradley and Allen, who noted significant improvements in outcome for patients undergoing repeated pancreatic débridement of infected pancreatic necrosis.63 Current care of patients with severe acute pancreatitis includes staging with dynamic, contrast material-enhanced helical CT scan to evaluate the extent of pancreatitis (unless significant renal dysfunction exists in which case one should forego the use of contrast material) coupled with the use of one of several prognostic scoring systems. Patients who exhibit clinical signs of instability (e.g., oliguria, hypoxemia, large-volume fluid resuscitation) should be carefully monitored in the ICU and undergo follow-up contrast enhanced CT examination when renal function has stabilized to evaluate for development of local pancreatic complications (Fig. 6-3). A recent change in practice has been the elimination of the routine use of prophylactic antibiotics for prevention of infected pancreatic necrosis. Enteral feedings initiated early, using nasojejunal feeding tubes placed past the ligament of Treitz, have been associated with decreased development of infected pancreatic necrosis, possibly due to a decrease in gut translocation of bacteria. These topics have been recently reviewed.64,65
Contrast-enhanced CT scan of pancreas 1½ weeks after presentation showing large central peripancreatic fluid collection.
The presence of secondary pancreatic infection should be suspected in patients whose systemic inflammatory response (fever, elevated WBC count, or organ dysfunction) fails to resolve, or in those individuals who initially recuperate, only to develop sepsis syndrome 2 to 3 weeks later. CT-guided aspiration of fluid from the pancreatic bed for performance of Gram’s stain and culture analysis can be useful. A positive Gram’s stain or culture from CT-guided aspiration, or identification of gas within the pancreas on CT scan, mandate surgical intervention.
The approach of open necrosectomy with repeated debridements, although life saving, is associated with significant morbidity and prolonged hospitalization. Efforts to reduce the amount of surgical injury, while still preserving the improved outcomes associated with debridement of the infected sequestrum have led to a variety of less invasive approaches.66 These include endoscopic approaches, laparoscopic approaches and other minimally invasive approaches. There are a limited number of randomized trials reporting the use of these new techniques currently. An important concept common to all of these approaches, however, is the attempt to delay surgical intervention, since a number of trials have identified increased mortality when intervention occurs during the first two weeks of illness.
Data supporting the use of endoscopic approaches to this problem include nearly a dozen case series and a randomized trial.67,68 The reported mortality rate was 5%, with a 30% complication rate. Most authors noted the common requirement for multiple endoscopic debridements (similar to the open approach), with a median of 4 endoscopic sessions required. Fewer series report experience with the laparoscopic approach, either transgastric or transperitoneal, entering the necrosis through the transverse mesocolon or gastrocolic ligament. The laparoscopic technique is carefully described in a recent publication.69 Laparoscopic intervention is limited by the difficulty in achieving multiple debridements and the technical expertise required to achieve an adequate debridement. Mortality in 65 patients in 9 case series reported was 6% overall.
Debridement of necrosis through a lumbar approach has been advocated by a number of authors. This approach, developed with experience in a large number of patients,70 has been recently subjected to a single center randomized prospective trial.71 This approach includes delay of intervention when possible until 4 weeks after the onset of disease. Patients receive transgastric or preferably retroperitoneal drainage of the sequestrum. If patients do not improve over 72 hours, they are treated with video-assisted retroperitoneal drainage (VARD), consisting of dilation of the retroperitoneal drain tract, placement of and irrigation, and debridement of the pancreatic bed (Fig. 6-4). Repeat debridements are performed as clinically indicated, with most patients requiring multiple debridements. In the trial reported, patients randomized to VARD (n=43) compared to those randomized to the standard open necrosectomy (n=45) had a decreased incidence of the composite endpoint of complications and death (40% vs. 69%), with comparable mortality rate, hospital, and ICU lengths of stay. Patients randomized to VARD had fewer incisional hernias, new-onset diabetes, and need for pancreatic enzyme supplementation.
Infected pancreatic necrosis. (A) Open necrosectomy specimen with pancreatic stent in situ. It is important to gently debride only necrotic pancreatic tissue, relying on repeated operation to ensure complete removal. (B) For video-assisted retroperitoneal debridement (VARD), retroperitoneal access is gained through radiologic placement of a drain, followed by dilation 2-3 days later. (C) Retroperitoneal cavity seen through endoscope during VARD.
It is apparent that patients with infected pancreatic necrosis can safely undergo procedures that are more minimal than the gold-standard open necrosectomy with good outcomes. However, to obtain good outcomes these approaches require an experienced multidisciplinary team consisting of interventional radiologists, gastroenterologists, surgeons, and others. Important concepts for successful management include careful preoperative planning, delay (if possible) to allow maturation of the fluid collection, and the willingness to repeat procedures as necessary till the majority if not all nonviable tissue has been removed.
Infections of the Skin and Soft Tissue
These infections can be classified according to whether or not surgical intervention is required. For example, superficial skin and skin structure infections such as cellulitis, erysipelas, and lymphangitis invariably are effectively treated with antibiotics alone, although a search for a local underlying source of infection should be undertaken. Generally, drugs that possess activity against the causative gram-positive skin microflora are selected. Furuncles or boils may drain spontaneously or require surgical incision and drainage. Antibiotics are prescribed if significant cellulitis is present or if cellulitis does not rapidly resolve after surgical drainage. Community-acquired methicillin resistant Staphylococcus aureus (MRSA) infection should be suspected if infection persists after treatment with adequate drainage and administration of first line antibiotics. These infections may require more aggressive drainage and altered antimicrobial therapy.72
Aggressive soft tissue infections are rare, difficult to diagnose, and require immediate surgical intervention plus administration of antimicrobial agents. Failure to do so results in an extremely high mortality rate (~80%–100%), and even with rapid recognition and intervention, current mortality rates are high (16%–24%).73 Eponyms and classification in the past have been a hodgepodge of terminology, such as Meleney’s synergist gangrene, rapidly spreading cellulitis, gas gangrene, and necrotizing fasciitis, among others. Today it seems best to delineate these serious infections based on the soft tissue layer(s) of involvement (e.g., skin and superficial soft tissue, deep soft tissue, and muscle) and the pathogen(s) that cause them.
Patients at risk for these types of infections include those who are elderly, immunosuppressed, or diabetic; those who suffer from peripheral vascular disease; or those with a combination of these factors. The common thread among these host factors appears to be compromise of the fascial blood supply to some degree, and if this is coupled with the introduction of exogenous microbes, the result can be devastating. However, it is of note that over the last decade, extremely aggressive necrotizing soft tissue infections among healthy individuals due to streptococci have been described as well.
Initially, the diagnosis is established solely upon a constellation of clinical findings, not all of which are present in every patient. Not surprisingly, patients often develop sepsis syndrome or septic shock without an obvious cause. The extremities, perineum, trunk, and torso are most commonly affected, in that order. Careful examination should be undertaken for an entry site such as a small break or sinus in the skin from which grayish, turbid semipurulent material (“dishwater pus”) can be expressed, as well as for the presence of skin changes (bronze hue or brawny induration), blebs, or crepitus. The patient often develops pain at the site of infection that appears to be out of proportion to any of the physical manifestations. Any of these findings mandates immediate surgical intervention, which should consist of exposure and direct visualization of potentially infected tissue (including deep soft tissue, fascia, and underlying muscle) and radical resection of affected areas. Radiologic studies should not be undertaken in patients in whom the diagnosis seriously is considered, as they delay surgical intervention and frequently provide confusing information. Unfortunately, surgical extirpation of infected tissue frequently entails amputation and/or disfiguring procedures; however, incomplete procedures are associated with higher rates of morbidity and mortality (Fig. 6-5).
Necrotizing soft tissue infection. (A) This patient presented with hypotension due to severe late necrotizing fasciitis and myositis due to beta-hemolytic streptococcal infection. The patient succumbed to his disease after 16 hours despite aggressive debridement. (B) This patient presented with spreading cellulites and pain on motion of his right hip 2 weeks after total colectomy. Cellulitis on right anterior thigh is outlined. (C) Classic dishwater edema of tissues with necrotic fascia. (D) Right lower extremity after debridement of fascia to viable muscle.
During the procedure a Gram’s stain should be performed on tissue fluid. Antimicrobial agents directed against Gram-positive and Gram-negative aerobes and anaerobes (e.g., vancomycin plus a carbapenem), as well as high-dose aqueous penicillin G (16,000,000 to 20,000,000 U/d), the latter to treat clostridial pathogens, should be administered. Approximately 50% of such infections are polymicrobial, the remainder being caused by a single organism such as Streptococcus pyogenes, Pseudomonas aeruginosa, or Clostridium perfringens. The microbiology of these polymicrobial infections is similar to that of secondary microbial peritonitis, with the exception that Gram-positive cocci are more commonly encountered. Most patients should be returned to the operating room on a scheduled basis to determine if disease progression has occurred. If so, additional resection of infected tissue and debridement should take place. Antibiotic therapy can be refined based on culture and sensitivity results, particularly in the case of monomicrobial soft tissue infections. Hyperbaric oxygen therapy may be of use in patients with infection caused by gas-forming organisms (e.g., Clostridium perfringens), although the evidence to support efficacy is limited to underpowered studies and case reports.In the absence of such infection, hyperbaric oxygen therapy has not shown to be effective.74
Postoperative Nosocomial Infections
Surgical patients are prone to develop a wide variety of nosocomial infections during the postoperative period, which include SSIs, UTIs, pneumonia, and bacteremia. SSIs are discussed earlier, and the latter types of nosocomial infections are related to prolonged use of indwelling tubes and catheters for the purpose of urinary drainage, ventilation, and venous and arterial access, respectively.
The presence of a postoperative UTI should be considered based on urinalysis demonstrating WBCs or bacteria, a positive test for leukocyte esterase, or a combination of these elements. The diagnosis is established after >104 CFU/mL of microbes are identified by culture techniques in symptomatic patients, or >105 CFU/mL in asymptomatic individuals. Treatment for 3 to 5 days with a single antibiotic directed against the most common organisms (e.g., E. Coli, K. pneumonia) that achieves high levels in the urine is appropriate. Initial therapy is directed by Gram’s stain results and is refined as culture results become available. Postoperative surgical patients should have indwelling urinary catheters removed as quickly as possible, typically within 1 to 2 days, as long as they are mobile, to avoid the development of a UTI.
Prolonged mechanical ventilation is associated with nosocomial pneumonia. These patients present with more severe disease, are more likely to be infected with drug-resistant pathogens, and suffer increased mortality compared to patients who develop community-acquired pneumonia. The diagnosis of pneumonia is established by presence of a purulent sputum, elevated leukocyte count, fever, and new chest X-ray abnormalities, such as consolidation. The presence of two of the clinical findings, plus chest X-ray findings, significantly increases the likelihood of pneumonia.75 Consideration should be given to performing bronchoalveolar lavage to obtain samples for Gram’s stain and culture. Some authors advocate quantitative cultures as a means to identify a threshold for diagnosis.76 Surgical patients should be weaned from mechanical ventilation as soon as feasible, based on oxygenation and inspiratory effort, as prolonged mechanical ventilation increases the risk of nosocomial pneumonia.
Infection associated with indwelling intravascular catheters has become a common problem among hospitalized patients. Because of the complexity of many surgical procedures, these devices are increasingly used for physiologic monitoring, vascular access, drug delivery, and hyperalimentation. Among the several million catheters inserted each year in the United States, approximately 25% will become colonized, and approximately 5% will be associated with bacteremia. Duration of catheterization, insertion or manipulation under emergency or nonsterile conditions, use for hyperalimentation, and the use of multilumen catheters increase the risk of infection. Use of a central line insertion protocol that includes full barrier precautions and chlorhexidine skin prep has been shown to decrease the incidence of infection.77 Although no randomized trials have been performed, peripherally inserted central venous catheters have a catheter-related infection rate similar to those inserted in the subclavian or jugular veins.78
Many patients who develop intravascular catheter infections are asymptomatic, often exhibiting solely an elevation in the blood WBC count. Blood cultures obtained from a peripheral site and drawn through the catheter that reveal the presence of the same organism increase the index of suspicion for the presence of a catheter infection. Obvious purulence at the exit site of the skin tunnel, severe sepsis syndrome due to any type of organism when other potential causes have been excluded, or bacteremia due to Gram-negative aerobes or fungi should lead to catheter removal. Selected catheter infections due to low-virulence microbes such as Staphylococcus epidermidis can be effectively treated in approximately 50% to 60% of patients with a 14- to 21-day course of an antibiotic, which should be considered when no other vascular access site exists.79 The use of antibiotic-bonded catheters and chlorhexidine sponges at the insertion site have been associated with lower rates of colonization.77 Use of ethanol or antimicrobial catheter “locks” have shown promise in reducing incidence of infection in dialysis catheters.80 The surgeon should carefully consider the need for any type of vascular access device, rigorously attend to their maintenance to prevent infection, and remove them as quickly as possible. Use of systemic antibacterial or antifungal agents to prevent catheter infection is of no utility and is contraindicated.
Severe sepsis is increasing in incidence, with over 1.1 million cases estimated per year in the United States with an annual cost of 24 billion dollars. This rate is expected to increase as the population of aged in the United States increases. One third of sepsis cases occur in surgical populations and sepsis is a major cause of morbidity and mortality.81 The treatment of sepsis has improved dramatically over the last decade, with mortality rates dropping to under 30%. Factors contributing to this improvement in mortality relate both to recent randomized prospective trials demonstrating improved outcomes with new therapies, and to improvements in the process of care delivery to the sepsis patient. The “Surviving Sepsis Campaign,” a multidisciplinary group that worked to develop treatment recommendations has published guidelines incorporating evidence-based treatment strategies most recently in 2013.13 These guidelines are summarized in Table 6-9.
Table 6-8Quality improvement organizations in the United States of interest to surgeons ||Download (.pdf) Table 6-8 Quality improvement organizations in the United States of interest to surgeons
Table 6-9Summary of Surviving Sepsis Campaign guidelines ||Download (.pdf) Table 6-9 Summary of Surviving Sepsis Campaign guidelines
Initial Evaluation and Infection Issues
Initial resuscitation: Begin resuscitation immediately in patients with hypotension or elevated serum lactate with resuscitation goal of central venous pressure (CVP) 8 to 12 mm Hg, mean arterial pressure of ≥65 mm Hg, urine output of ≥0.5 mL/kg/h, and mixed venous oxygen saturation of 65%.
Target resuscitation to normalize lactate in patients with elevated lactate levels.
Diagnosis: Obtain appropriate cultures prior to antibiotics but do not delay antibiotic therapy. Use rapid antigen assays in patients with suspected fungal infection. Imaging studies should be performed promptly to confirm a source of infection.
Antibiotic therapy: Begin IV antibiotic therapy as early as possible: should be within the first hour after recognition of severe sepsis/septic shock. Use broad spectrum antibiotic regimen with penetration into presumed source, reassess regimen daily with deescalation as appropriate. Discontinue antibiotics in 7–10 d for most infections, stop antibiotics for noninfectious issues.
Source control: Establish anatomic site of infection as rapidly as possible, implement source control measures immediately after initial resuscitation. Remove intravascular access devices if potentially infected.
Infection prevention: Selective oral and digestive tract decontamination.
Hemodynamic Support and Adjunctive Therapy
Fluid therapy: Fluid resuscitate using crystalloid, using fluid volumes of 1000 mL (crystalloid), target CVP of 8 to12 mm Hg.
Vasopressors/Inotropic Therapy: Maintain MAP of ≥65 mm Hg, centrally-administered norepinephrine is first-line choice. Dopamine should not be used for “renal protection,” insert arterial catheters for patients requiring vasopressors. Phenylephrine is not recommended in treatment of septic shock. Dobutamine infusion can be used in setting of myocardial dysfunction. Do not use strategy of targeting supranormal cardiac index.
Steroids: Consider intravenous hydrocortisone (dose ≤300 mg/d) for adult septic shock when hypotension responds poorly to fluids and vasopressors.
Other Supportive Therapy
Blood product administration: Transfuse red blood cells when hemoglobin decreases to <7.0 g/dL.
Mechanical ventilation: Target an initial tidal volume of 6 mL/kg body weight and plateau pressure of ≤30 cm H2O in patients with acute lung injury. Use positive end-expiratory pressure to avoid lung collapse. Use a weaning protocol to evaluate the potential for discontinuing mechanical ventilation. Pulmonary artery catheter is not indicated for routine monitoring.
Sedation: Minimize sedation using specific titration endpoints.
Glucose control: Use protocolized approach to blood glucose management targeting upper blood glucose target of 180 mg/dL.
Prophylaxis: Use stress ulcer (proton pump inhibitor or H2 blocker) and deep venous thrombosis (low-dose unfractionated or fractionated heparin) prophylaxis.
Limitation of support: Discuss advance care planning with patients and families and set realistic expectations.
Patients presenting with severe sepsis should receive resuscitation fluids to achieve a central venous pressure target of 8-12 mm Hg, with a goal of mean arterial pressure of ≥ 65 mHg and urine output of ≥ 0.5 mL/kg/h. Delaying this resuscitative step for as little as 3 hours until arrival in the ICU has been shown to result in poor outcome.82 Typically this goal necessitates early placement of central venous catheter.
A number of studies have demonstrated the importance of early empirical antibiotic therapy in patients who develop sepsis or nosocomial infection. This therapy should be initiated as soon as possible with broad spectrum antibiotics directed against most likely organisms, since early appropriate antibiotic therapy has been associated with significant reductions in mortality, and delays in appropriate antibiotic administration are associated with increased mortality. Use of institutional and unit specific sensitivity patterns are critical in selecting an appropriate agent for patients with nosocomial infection. It is key, however, to obtain cultures of appropriate areas without delaying initiating antibiotics so that appropriate adjustment of antibiotic therapy can take place when culture results return.
Additionally, early identification and treatment of septic sources is key for improved outcomes in patients with sepsis. Although there are no randomized trials demonstrating this concept, repeated evidence in studies of patients who develop intraabdominal infection, necrotizing soft tissue infection, and other types of infections demonstrate increased mortality with delayed treatment. As discussed earlier, one exception is that of infected pancreatic necrosis.
Multiple recent trials have evaluated the use of vasopressors and inotropes for treatment of septic shock. The current first-line agent for treatment of hypotension is norepinephrine. It is important to titrate therapy based on other parameters such as mixed venous oxygen saturation and plasma lactate levels as well as mean arterial pressure to reduce the risk of vasopressor-induced perfusion deficits. Several recent randomized trials have failed to demonstrate benefit with use of pulmonary arterial catheterization, leading to a significant decrease in its use.
A number of other adjunctive therapies are useful in treatment of the patient with severe sepsis and septic shock. Low-dose corticosteroids (hydrocortisone at ≤300 mg/day) can be used in patients with septic shock who are not responsive to fluids and vasopressors. However, a recent randomized trial failed to show survival benefit. Patients with acute lung injury associated with sepsis should receive mechanical ventilation with tidal volumes of 6 mL/kg and pulmonary airway plateau pressures of ≤30 cm H2O. Finally, red blood cell transfusion should be reserved for patients with hemoglobin of <7 grams/dL, with a more liberal transfusion strategy reserved for those patients with severe coronary artery disease, ongoing blood loss, or severe hypoxemia.
In the 1940s, penicillin was first produced for widespread clinical use. Within a year of its introduction, the first resistant strains of Staphylococcus aureus were identified. There are two major components that are responsible for antibiotic resistance. First, there may be a genetic component innate to the organism that prevents an effect of a particular antibiotic. For instance, if an organism does not have a target receptor specific to the mechanism of action of a particular antibiotic, the antibiotic will not be effective against this organism. A good example is penicillin and Gram-negative organisms, as thesemicrobes lack penicillin-binding proteins. The second component driving resistance is that related to antibiotic selection. Over generations of exposure to a particular antibiotic, selection pressure will drive proliferation of more organisms resistant to that antibiotic. It is this mechanism that leads to antibiotic resistance in the world today, given that there are millions of kilograms of antibiotics used annually in people, in agriculture, and for animal use. This has led to antibiotic resistance described in all classes of antibiotics in common use today. Antibiotic resistance comes at a high cost, with a significant increase in mortality associated with infection from resistant organisms, and an economic cost of billions of dollars per year.
Resistance mechanisms are varied, and include one of three routes. Resistance can be intrinsic to the organism (natural resistance), can be mutational and mediated by changes in the chromosomal makeup of the organism, and finally can be mediated by extrachromosomal transfer of genetic material via transposons or plasmids. Resistance due to mutation includes mechanisms mediated by target site modification, reduced permeability/uptake, metabolic bypass, or derepression of multidrug efflux systems. Genes transferred via plasmid or transposon include those that cause drug inactivation, increases in antibiotic efflux systems, target site modification, and metabolic bypass.
There are several drug resistant organisms of interest to the surgeon. MRSA occurs as a hospital-associated infection more common in chronically ill patients receiving multiple courses of antibiotics. However, recent strains of MRSA have emerged in the community among patients without preexisting risk factors for disease.72 These strains, which produce a toxin known as Panton-Valentin leukocidin, make up an increasingly high percentage of surgical site infections since they are resistant to commonly employed prophylactic antimicrobial agents.83 Extended spectrum β-lactamase (ESBL)-producing strains of Enterobacteraceae, originally geographically localized and infrequent, have become much more widespread and common in the last decade.84 These strains, typically Klebsiella or E coli species, produce a plasmid-mediated inducible β-lactamase. Commonly encountered plasmids also confer resistance to many other antibiotic classes (multidrug resistance). A common laboratory finding with ESBL is sensitivity to first-, second-, or third- generation cephalosporins with resistance to others. Unfortunately, use of this seemingly active agent leads to rapid induction of resistance and failure of antibiotic therapy. The appropriate antibiotic choice in this setting is a carbepenem. While Enterococcus used to be considered a low virulence organism in the past, infections caused by E faecium and faecalis have been found to be increasingly virulent, especially in the immunocompromised host. The last decade has seen increased isolation of a vancomycin-resistant strain of Enterococcus.85 This resistance is transposon-mediated via the vanA gene and is typically seen in E faecium strains. A real concern in this setting is transfer of genetic material to S aureus in a host coinfected with both organisms. This is thought to be the mechanism behind the half dozen recently described cases of vancomycin resistance in S aureus.
While alarming to contemplate, the risk of human immunodeficiency virus (HIV) transmission from patient to surgeon is low. As of May 2011, there had been six cases of surgeons with HIV seroconversion from a possible occupational exposure, with no new cases reported since 1999. Of the numbers of health care workers with likely occupationally acquired HIV infection (n = 200), surgeons were one of the lower risk groups (compared to nurses at 60 cases and nonsurgeon physicians at 19 cases).86 The estimated risk of transmission from a needlestick from a source with HIV-infected blood is estimated at 0.3%. Transmission of HIV (and other infections spread by blood and body fluid) from patient to health care worker can be minimized by observation of universal precautions, which include the following: (a) routine use of barriers (such as gloves and/or goggles) when anticipating contact with blood or body fluids, (b) washing of hands and other skin surfaces immediately after contact with blood or body fluids, and (c) careful handling and disposal of sharp instruments during and after use.
Postexposure prophylaxis for HIV has significantly decreased the risk of seroconversion for health care workers with occupational exposure to HIV. Steps to initiate postexposure prophylaxis should be initiated within hours rather than days for the most effective preventive therapy. Postexposure prophylaxis with a two- or three-drug regimen should be initiated for health care workers with significant exposure to patients with an HIV-positive status. If a patient’s HIV status is unknown, it may be advisable to begin postexposure prophylaxis while testing is carried out, particularly if the patient is at high risk for infection due to HIV (e.g., intravenous narcotic use). Generally, postexposure prophylaxis is not warranted for exposure to sources with unknown status, such as deceased persons or needles from a sharps container.
The risks for surgeons of acquiring HIV infection have recently been evaluated by Goldberg and coauthors.87 They noted that the risks are related to the prevalence of HIV infection in the population being cared for, the probability of transmission from a percutaneous injury suffered while caring for an infected patient, the number of such injuries sustained, and the use of postexposure prophylaxis. Annual calculated risks in Glasgow, Scotland, ranged from one in 200,000 for general surgeons not utilizing postexposure prophylaxis to as low as one in 10,000,000 with use of routine postexposure prophylaxis after significant exposures.
Hepatitis B virus (HBV) is a DNA virus that affects only humans. Primary infection with HBV generally is self-limited, but can cause fulminant hepatitis or progress to a chronic carrier state. Death from chronic liver disease or hepatocellular cancer occurs in roughly 30% of chronically infected persons. Surgeons and other health care workers are at high risk for this blood-borne infection and should receive the HBV vaccine; children are routinely vaccinated in the United States.88 This vaccine has contributed to a significant decline in the number of new cases of HBV per year in the United States, from approximately 250,000 annually in the 1980s to 3,350 in 2010.89,90 This is truly one of the unsung victories in vaccination strategy in the last 20 years.
Hepatitis C virus (HCV), previously known as non-A, non-B hepatitis, is a RNA flavivirus first identified specifically in the late 1980s. This virus is confined to humans and chimpanzees. A chronic carrier state develops in 75% to 80% of patients with the infection, with chronic liver disease occurring in three-fourths of patients who develop chronic infection. The number of new infections per year has declined since the 1980s due to routine testing of blood donors for this virus. Fortunately, HCV is not transmitted efficiently through occupational exposures to blood, with the seroconversion rate after accidental needlestick approximately 1.8%.91 To date, a vaccine to prevent HCV infection has not been developed. Experimental studies in chimpanzees with HCV immunoglobulin using a model of needlestick injury have failed to demonstrate a protective effect, and no effective antiviral agents for postexposure prophylaxis are available. Treatment of patients who develop HCV infection includes ribavirin and pegylated gamma interferon.92