Chapter 18: Infections

Death after traumatic injury has classically been described in terms of a trimodal distribution. Immediate and acute (<24 hours) deaths usually result from uncontrolled hemorrhage, but infections and multiple organ dysfunction syndrome, which often arise from infection, are responsible for a significant proportion of late deaths. Indeed, infection is responsible for most deaths in patients who survive longer than 48 hours after trauma.¹ Trauma-related infections can be divided into those that result directly from the injury (eg, due to contamination that occurs in conjunction with the traumatic injury) and nosocomial infections that arise in the health care setting in conjunction with treatment of the injury. The pathogens involved can be exogenous or endogenous bacteria, depending on the mechanism of injury and/or the iatrogenic cause.

Most post-traumatic infections are polymicrobial, involving a mixture of aerobic and anaerobic organisms.² In one series, 37–45% of all trauma patients experienced infectious complications during their initial hospitalization.³ In that same study, 80% of trauma patients who were in the intensive care unit (ICU) at least 7 days met the criteria for the systemic inflammatory response syndrome (SIRS). All caregivers need to understand the principles of surgical infections in the context of trauma patients. This chapter will review the following: factors that normally prevent infection, how trauma disrupts or overwhelms normal host defenses, recognition and treatment of the most common infectious complications after traumatic injury, principles of infection prevention, and how prophylaxis and prevention principles can be applied chronologically during the treatment of trauma patients.

Humans have evolved mechanisms to avoid infection despite the ubiquitous presence of bacteria in our environment and throughout our bodies. A recent review documented the magnitude and diversity of the human microbiota.⁴ Under normal circumstances there is a balance between our microbiota, intact environmental barriers, and host defenses (Fig. 18-1). Invasive surgical procedures or traumatic injury disrupts this balance and significantly increases the probability of developing an infection (Fig. 18-2). Microbes are abundant on the surface of the skin, within the oral cavity, and increase in numbers down the length of the gastrointestinal tract. Bacterial numbers differ at various locations, and the pathogenic species and their respective numbers at different anatomic sites are summarized in Table 18-1. Once inoculation of bacteria into normally sterile sites occurs, infection will ensue if bacteria proliferate faster than the host defense mechanisms eradicate them. Trauma generally results in a much greater disruption of normal barriers than occurs with elective surgery. Trauma may also have concomitant hypoperfusion (shock), devitalized tissue, and retained foreign bodies.

Environmental Barriers

Entry of microbes is normally limited by the integrity of environmental barriers, such as intact skin and the mucosa of the respiratory, gastrointestinal, and genitourinary tracts.⁵ The importance of intact skin is clearly evident when one considers the potential for microbial infection seen in burn patients or in patients with toxic epidermal necrolysis.² Many traumatic injuries are associated with an alteration in the integrity of the skin. Even minor lacerations and abrasions have the potential to disrupt crucial environmental barriers. Interventions made in the process of caring for trauma patients, such as insertion of intravenous or urinary catheters, tube thoracostomy, etc., likewise disrupt the integument and may permit skin bacteria to access sterile sites. Studies have shown the quantitative number of microbes required to produce clinical infection is significantly decreased in the presence of foreign bodies, blood, or devitalized tissue, which are often present with traumatic wounds.²

Microbial Factors

The bacteria species responsible for clinical infections in surgical or trauma patients constitute a minority of the skin or gastrointestinal flora. These pathogenic bacterial species generally possess one or more virulence factors that facilitate infection and increase their pathogenicity. In contrast, the vast majority of endogenous and environmental bacteria are relatively nonpathogenic. For example, more than 99% of the colonic microbiota are nonpathogenic anaerobes that never cause clinical infections. Lactobacilli, which do not cause clinical infection, are the most abundant skin bacteria. In contrast, the most common pathogen associated with surgical site infections (SSI), Staphylococcus aureus, has numerous virulence factors that facilitate invasion and thwart host defenses. In the abdominal cavity, Escherichia coli and Bacteroides fragilis are the prototypical organisms associated with intra-abdominal infection, yet they account for only 0.01% and 1% of colonic bacteria, respectively. Under normal circumstances the overwhelming numbers of nonpathogenic bacteria constitute a robust “defense” against infection, because infection is proportionately less likely if >99% of the inoculum is incapable of producing infection. This concept of adherent resident bacteria preventing invasion has been termed colonization resistance.⁵ This is an important point because the skin and gastrointestinal microbial flora change considerably when trauma patients are hospitalized, both in terms of number and proportion of virulent bacteria and in terms of susceptibility to antibiotics, should an infection develop.

The skin microflora is relatively homogeneous, although bacterial numbers are higher in the axilla and groin areas. The endogenous skin bacteria are predominately gram-positive aerobic Staphylococcus and Streptococcus species, along with Corynebacterium and Propionibacterium.⁵ As noted above, S. aureus is the most common pathogen present on the skin, but increasingly isolates from trauma patients and community-acquired infections are methicillin resistant S. aureus (MRSA).^6,7 This fact needs to be taken into account in terms of appropriate empiric or prophylactic antibiotic selection for these patients.⁷

The oral and nasopharynx harbor large numbers of bacteria, most frequently streptococcal species. Much smaller numbers of bacteria, typically 10²–10³ CFU/mL, are present in the normal stomach, because the normal acid pH of the stomach inhibits bacterial growth. Gastric bacterial numbers increase in the absence of gastric acid, as in patients receiving proton pump inhibitors. Bacterial numbers are much higher in the small intestine and the density of bacteria increases from the duodenum to the terminal ileum. Bacterial counts in the proximal gastrointestinal tract are in the range of 10⁴–10⁵ CFU/mL, whereas numbers in the terminal ileum are close to colonic densities (10⁸–10¹⁰ CFU/mL). Bacterial numbers in the colon are even higher (10¹¹–10¹² CFU/mL) although most colonic bacteria are nonpathogenic. The presence of such large numbers is associated with very low oxygen tension, and 99.9% of bacteria present are anaerobes. The urogenital, biliary, pancreatic ductal, and distal respiratory tracts are “sterile” in healthy individuals.⁵

Host Defense Mechanisms

Host defense refers to endogenous factors that counteract microbial invasion. In addition to the environmental factors and colonization resistance described earlier, crucial humoral and cellular host defense mechanisms exist that eliminate bacteria within sterile spaces. Initially, several primitive and relatively nonspecific host defenses including proteins such as lactoferrin, fibrinogen, and complement begin to act against invading microbes. Lactoferrin sequesters the critical microbial growth factor iron, thereby limiting microbial growth. Fibrinogen within the inflammatory fluid has the ability to trap large numbers of microbes during polymerization into fibrin.⁵ Complement is activated on contact with bacteria and viruses, from tissue damage, or when IgG/IgM antibodies recognize microbial agents. Activation of complement releases C3a and C5a, which are potent chemotaxins that recruit neutrophils and macrophages. These components enhance endothelial adhesiveness and increase vascular permeability. Complement activation can directly destroy microbial agents via formation of a membrane attack complex (composed of complement proteins C5–C9) and enhance microbial phagocytosis by way of C1q and C3bi subunits. In vitro studies have shown that 50–70% of a moderate inoculum is eliminated prior to the influx of phagocytic host cells.

Many different tissues also contain resident innate immune cells. These include macrophages, dendritic cells, Kupffer cells, glial cells, mesangial cells, and alveolar macrophages.⁸ These innate immune cells express a wide variety of pathogen-associated molecular pattern (PAMP) receptors on their surface.^9,10,11 The best known examples of PAMPs are the toll-like receptors (TLRs) of which there are now many, well-described receptor molecules.¹¹ TLRs bind to ligands on bacteria (or damaged host tissue), and TLR binding results in activation of these cells. Activated macrophages secrete an array of substances in response leading to amplification and regulation of the acute proinflammatory response (Fig. 18-3). Sequential release of protein cytokines, including tumor necrosis factor-alpha (TNF-alpha), interleukin-1 (IL-1), IL-6, IL-8, and interferon-gamma (INF-γ), follows. These mediators produce the signs and symptoms that we associate with infection (fever, tachycardia, tachypnea, leukocytosis, etc.). IL-8 is a very potent chemoattractant for neutrophils, the innate immune cells primarily responsible for ongoing microbial phagocytosis and intracellular microbial killing. Unfortunately, the same process that recruits neutrophils and stimulates phagocytosis and oxidative killing may also contribute to damage of host tissues. Simultaneous with this proinflammatory response there is also production of anti-inflammatory mediators, such as IL-10 and transforming growth factor-β (TGFβ).¹² Some of these mediators may contribute to the immune “hypo-responsiveness of trauma” over the ensuing days (Table 18-2).

Specific anatomic locations have additional unique factors that defend against infection.¹³ For example, the peritoneal cavity has lymphatic channels on the undersurface of the diaphragm that facilitate removal of bacteria.¹⁴ The sub-diaphragmatic surface is a lower-pressure area due to the effect of respiratory excursion and this serves to move free fluid within the peritoneal cavity to this location. Movement of the diaphragm “pumps” this fluid into the thoracic duct and from there it gains rapid access to the systemic circulation. Experimental studies show that labeled bacteria inoculated into the peritoneal cavity appear in the thoracic duct within 6 minutes and in the bloodstream within 12 minutes.¹³ The respiratory tract has unique host defenses that help to ensure the sterility of the lung parenchyma, as well. Goblet cells within the respiratory mucosa secrete mucin that helps to trap bacteria. Ciliated respiratory epithelial cells move the mucus centrally where it, and the bacteria trapped within it, can be expectorated by coughing. The presence of endotracheal tubes, smoking, inhaled toxins, and some anesthetic agents interfere with mucocilliary clearance mechanisms, and this may predispose to pneumonia. Bacteria or other microbes that gain access to the alveoli are normally phagocytized by alveolar macrophages, although the macrophage activation that may accompany this process has been proposed as one possible pathogenic mechanism for acute lung injury (ALI) or adult respiratory distress syndrome (ARDS).^15,16,17

To a very large extent the microbial agents responsible for infections or infectious complications after trauma are the same agents that cause most other surgical or ICU-associated infections. Table 18-1 shows the most common infectious agents that cause trauma-associated infections at various anatomic sites. Generally, Staphylococcus spp. and Streptococcus spp. are the most common pathogens responsible for infections in which the traumatic injury or operative intervention needed to treat the injury did not transgress a mucosal surface. For traumatic injuries that involve the aerodigestive tract the most common isolates are E. coli (43.4%), S. aureus (18.9%), Klebsiella pneumoniae (14.4%), and Enterococcus faecalis (5.6%).¹ Hospitalized trauma patients develop nosocomial bacterial infections from the usual ICU-associated pathogens (Table 18-3). A few infectious agents that can be associated with trauma are seldom encountered in other settings including rabies virus, Clostridium tetani, and Vibrio spp. We will also discuss some of the unique challenges and concerns that have been brought to the attention of surgeons during the Ebola outbreak of 2014.

Rabies

Rabies is a rare, potentially fatal, clinical disease caused by the rabies virus. Rabies is an RNA virus present in the saliva of mammals and transmission to humans generally occurs following a bite from a rabid animal. Prior to the development of a vaccine by Louis Pasteur, bites from a rabid animal were uniformly fatal. In North America, raccoons, skunks, bats, foxes, coyotes, and bobcats are the primary reservoirs. Most patients with rabies have no documented exposure to a rabid animal, and the majority of these are associated with bat bites. Many victims underestimate the importance of a bat bite and a substantial portion do not even recall being bitten. Bats (Carnivora and Chiroptera) represent the ultimate zoonotic reservoir for the virus, as well. The rabies virus is highly labile and can be inactivated readily by ultraviolet radiation, heat, desiccation, and other environmental factors.

The word “rabies” derives from the Latin rabere meaning “to rage” and refers to the clinical manifestations of the disease that include hyperactivity, disorientation, hallucinations, and bizarre behavior. The rabies virus is neurotropic and causes an acute encephalitis. Other hallmarks of the disease include an intense fear of suffocation (eg, hydrophobia and aerophobia) secondary to intense laryngeal and pharyngeal spasm. Once the patient begins manifesting symptoms death is nearly certain. With increased vaccination and postexposure prophylaxis (PEP) over the past 50 years, the clinical disease is becoming increasingly uncommon. According to the Centers for Disease Control and Prevention (CDC) 49 cases of human rabies were reported in the United States between 1995 and 2011, with only 3 survivors. That said, it is important for the practitioner of emergency medicine/surgery to be knowledgeable about rabies since animal bites are encountered frequently in clinical practice.

Humans are not routinely vaccinated against rabies, while domestic animals receive routine rabies vaccinations. If a human is bitten by a rabid animal, rabies can be prevented by PEP before the virus enters the central nervous system during the incubation period. The diagnosis of rabies can be made rapidly by identification of rabies virus in the brain of a potentially infected animal. This procedure can be performed most expeditiously by euthanizing the suspected animal. If the rabies test is negative, then no postexposure vaccination or prophylaxis is needed. The incidence of positive rabies tests ranges from as high as 6–10% in wild animals down to levels of ~1% in domestic pets. An acceptable alternative approach, if the suspected source is a domestic pet (dog, cat, ferret, etc.), is for the offending animal to be quarantined and observed for 10 days. If the animal exhibits signs of rabies, the exposed person should begin PEP immediately and the animal should be euthanized and its brain tissue tested for rabies. If the animal is confirmed to have rabies, PEP must be completed; however, if the test results are negative PEP can cease.

Immediate measures that should be taken to decrease the risk of rabies transmission include thorough washing of bite and scratch wounds with soap and water, followed by application of povidone–iodine or alcohol. Human rabies immune globulin (HRIG) and rabies vaccine should be given in all cases except in persons who have been immunized previously.¹⁸ Immune globulin should never be delivered in the same syringe as the vaccine, as this will cause precipitation. The Advisory Committee on Immunization Practices (ACIP) of the CDC and the American Academy of Pediatrics recommend a single dose (20 IU/kg) of HRIG be given to provide protection for the first 2 weeks until the vaccine elicits an antibody response. Detailed and up-to-date information for rabies exposure is available on the CDC’s Web site (http://www.cdc.gov/rabies/index.html), and this site should be consulted for the latest information. The ACIP recommends a regimen of human diploid cell vaccine (Imovax^®) for PEP on days 0, 3, 7, 14, and 28 along with a single dose of HRIG on day 0. Once initiated, rabies prophylaxis should not be interrupted or discontinued because of local or mild systemic reactions to the vaccine.

Tetanus

Tetanus is a rare, life-threatening condition that is caused by toxins produced by Clostridium tetani, a spore-forming, gram-positive bacillus.¹⁹ Clostridial spores can survive indefinitely, and they are ubiquitous in soil and feces. Under anaerobic conditions the spores can germinate into mature bacilli that elaborate the neurotoxins tetanospasmin and tetanolysin. Tetanospasmin produces most prominent clinical symptoms by interfering with motor neuron release of the inhibitory neurotransmitters gamma-aminobutyric acid (GABA) and glycine. The resulting loss of inhibition produces muscle spasm (usually spasm of the masseter muscle) and severe autonomic over-activity manifested by high fever, tachycardia, and hypertension. Historically, tetanus was highly fatal, but intensive medical therapy with neuromuscular blockade, mechanical ventilation, and ICU monitoring has lowered the case fatality rate to 11–28%. According to the CDC there were a total of 233 tetanus cases reported in the United States between 2001 and 2008, with 71.7% arising from acute trauma. Clinicians and trauma surgeons must remain alert for the potential of Clostridial contamination and provide appropriate, timely tetanus prophylaxis.¹⁹

The diagnosis of tetanus is made on clinical grounds alone, as there are no laboratory tests that can diagnose the condition or rule it out. Tetanus immunization is accomplished as a component of standard early childhood immunizations (diphtheria–pertussis–tetanus [DPT]), with administration of tetanus toxoid (TT) every 5–10 years to maintain immune memory (see, http://www.cdc.gov/tetanus/vaccination.html). No deaths have been reported in individuals who have been fully immunized. The CDC recommendations for tetanus prophylaxis depend on the wound characteristics and the prior immunization status of the patient. A wound with extensive contamination, one that is poorly vascularized, or with extensive soft tissue trauma is considered to be a “tetanus-prone wound.” A tetanus booster should be administered to patients who have received primary immunization, but who have not received TT during the past 10 years, or the past 5 years for tetanus-prone wounds.¹⁸ In patients who have never undergone primary immunization, human tetanus immune globulin (HTIG) should be administered along with TT at a different site. Antitetanus antibody binds to exotoxins and neutralizes their toxicity. High-risk groups such as the elderly, individuals with human immunodeficiency virus (HIV), and intravenous drug users (IVDU) who had received primary vaccination may not have tetanus antibodies, and more liberal use of HTIG should be considered in these groups.¹⁸

Infections Associated with Marine Trauma

Vibrio vulnificus is a gram-negative rod present in seawater that can result in atypical, necrotizing soft tissue infections when traumatic injuries occur in the ocean.^20,21,22 V. vulnificus is common in warm seawater and thrives in water temperatures greater than 68°F (20°C). The organism is not associated with pollution or fecal waste. Approximately 25% of V. vulnificus infections are caused by direct exposure of an open wound to warm seawater containing the organism. Exposure typically occurs when the patient is participating in water activities such as boating, fishing, or swimming. Infections are occasionally attributed to contact with raw seafood or marine wildlife. The risk of developing Vibrio infection is much higher in immunocompromised patients or patients with preexisting hepatic disease or diabetes mellitus.²¹ Established infection with V. vulnificus can be highly invasive with mortality rates of 30–40% and a mortality greater than 50% in immunocompromised patients. A published report from 2009 documented a 37% mortality rate even after implementation of a specific treatment guideline for necrotizing Vibrio infections.²⁰

Patients with wound infections caused by V. vulnificus develop painful cellulitis that progresses rapidly.^21,22,23 Physical examination will often reveal marked swelling and painful, hemorrhagic bullae surrounding traumatic wounds. In some cases, there can be rapid progression and associated systemic symptoms. Marked local tissue swelling with hemorrhagic bullae is characteristic. Systemic symptoms include fever and chills, and bacteria are present in the bloodstream in more than 50% of patients. Hypotension or septic shock may be an early symptom, and alterations in mental status occur in approximately one-third of patients. Table 18-4 summarizes clinical symptoms present in patients with Vibrio infection. It is important for trauma surgeons to be aware of the potential for Vibrio infections in the appropriate clinical setting, because antibiotic treatment is distinctly different from the agents typically employed for trauma patients. Aggressive surgical debridement, incision and drainage of purulent collections, and even amputation may be crucial adjuncts for management of these often severe soft tissue infections.²¹ Recent experience in 30 patients with documented Vibrio infection found that fasciotomy was needed in all patients, and 17% required amputation.²⁰ Recommended antibiotics include doxycycline (100 mg iv/po bid), ceftazidime (2 g q 8 hours), cefotaxime (2 g q 8 hours), or ciprofloxacin (750 mg po bid or 400 mg iv q 12 hours).^21,24

Traumatic injuries that occur in freshwater conditions may develop infections from Aeromonas hydrophila.²³ A. hydrophila is a gram-negative anaerobic rod that is a common pathogen of fish and amphibians. Cutaneous inoculation of the organism can result in cellulitis, abscesses and, occasionally, necrotizing soft tissue infections. Like the situation with Vibrio infections, patients with hepatic disease and immune-compromise have a greater risk of developing generalized disease. A. hydrophila can be recovered from the bloodstream in a significant proportion of patients and this fact, along with a history of injury in fresh water, will aid in alerting clinicians to the correct diagnosis. Antibiotic agents active against A. hydrophila include third-generation cephalosporins, fluoroquinolones, doxycycline, or trimethoprim–sulfamethoxazole.²³

Ebola Virus Disease

On September 30, 2014, in the midst of the worst ebola virus disease (EVD) outbreak on record,²⁵ the Centers for Disease Control (CDC) confirmed the first documented case of EVD in the United States in Dallas, Texas. Within 2 weeks a nurse who cared for that initial patient was also diagnosed with the disease. While EVD is not a surgical disease per se, surgeons knowledge and skills in resuscitation, critical care, and disaster management may be invaluable in the treatment of EVD. Therefore all surgeons should be familiar with the fundamentals of EVD management.

Transmission

Ebola virus is a zoonosis belonging to the Filoviridae family. Five species have been identified as follows: Zaire, Sudan, Ivory Coast, Bundibugyo, and Reston.²⁶ Ebola is a single-stranded RNA virus that sporadically spreads to humans from a presumptive animal or bat reservoir in the wild. Human spread only occurs through direct contact with infected bodily fluid, and the most infectious sources are blood, feces, and vomit.²⁷ Ebola virus can remain viable on surfaces for 1–6 days, although the risk of transmission from a surface contact is considered very low.^28,29 Humans are only contagious while they are symptomatic. Blood from patients with untreated late EVD harbor 10⁹ virus particles per mL, making it highly infectious. This value can be compared to the much lower viral loads in patients with untreated HIV (10⁵/mL blood) or Hepatitis C (5–20 × 10⁶/mL blood), both of which are recognized as being highly transmissible.³⁰ EVD is not an airborne pathogen, but, since it can be transmitted via large droplets, health care providers in close contact with symptomatic patients should take necessary droplet precautions (See Prevention).³¹

Pathophysiology

Ebola virus enters the body through breaks in the skin, mucus membranes or ingestion and then infiltrates cells, especially lymph tissues, liver, and spleen. The virus continues to replicate until cells become necrotic and lyse, spilling more viral particles into the circulation. Ebola virus elicits a profound proinflammatory cytokine and chemokine response producing a vigorous SIRS reaction.^32,33 Significant endothelial injury ensues, with loss of vascular tone and increased vascular permeability leading to hypotension and shock.^34,35 At the end stages of disease the virus is found in all body fluids and skin, making handling of bodies of EVD victims extremely hazardous. In contrast to prevailing biases, bleeding does not occur in all patients and only manifests late in the disease process as bleeding from skin, gums, and the gastrointestinal tract.³⁶ Mortality rate remains high at 50–90% overall with a poorer prognosis seen in those with older age, diarrhea, hemorrhagic conjunctivitis, shortness of breath, confusion/coma, and hemorrhage.³⁷

Symptoms

Patients in early stages of the disease present with the acute onset of flu-like symptoms. Fever is the most common symptom (present in 87%), with fatigue (76%), abdominal pain (44%), and nausea and vomiting (65%) frequently seen, also.^37,38,39 Late symptoms are representative of a fulminant infection and SIRS with symptoms of severe vomiting and diarrhea, both of which may be bloody. Other late signs include a coagulopathy identified by diffuse oozing from mucosal surfaces and sites of intravenous lines. Death is often the result of recalcitrant multiorgan system failure.²⁵

Treatment

Currently there is no vaccine against or pharmacologic treatment for EVD, and the mainstay of therapy is supportive care. Many patients will have severe diarrhea, losing up to 15 L/day and require aggressive resuscitation efforts focused on replenishing circulatory volume. Significant electrolyte abnormalities are common, and correction of these must be done in order to prevent cardiac dysrhythmias. In the early stages of the disease patients can often tolerate oral fluids, antiemetics, and antidiarrheals. Patients with protracted vomiting or in those later stages of the disease should have intravenous fluids administered. Blood transfusions may be necessary, especially for those with hemorrhagic symptoms, and should follow current recommendations for transfusion triggers. Respiratory support in the form of mechanical ventilation may be necessary, but caution should be exercised with noninvasive support, as it may lead to aerosolization.⁴⁰ Some novel antivirals known to have in vitro or animal anti-Ebola activity were used during the 2014 outbreak, but data regarding their effectiveness are inconclusive at this time. Experimental therapies used during the 2014 EVD outbreak included plasma from convalescent or immune patients and experimental monoclonal antibodies, but it is not clear that either method provided a survival benefit.^41,42 Several vaccines are currently undergoing accelerated human trials, but none as yet are commercially available.⁴³

Prevention

The 2014 EVD outbreak clearly demonstrated how globalization has enhanced the capacity for a previously isolated tropical disease to rapidly spread. Prevention relies on maintaining a level of preparedness of providers and facilities that can receive patients with EVD, provide quality critical care, protect the providers, and assuage the fears of the community. Preventing EVD transmission is divided into mitigating transmission of the primary disease and screening for early identification of the disease as part of planning for a disaster.

Strict adherence to protocols for use of personal protective equipment (PPE) and adoption of practices to limit exposure to body fluids constitute the mainstays of primary prevention. Practiced use of full PPE, as described by the CDC (http://www.cdc.gov/vhf/ebola/healthcare-us/ppe/guidance.html) that is donned and doffed in the presence of a trained observer, is paramount when treating a patient with EVD.⁴⁴ All practitioners who will come into contact with a patient with EVD should be trained and certified in their ability to properly wear full PPE prior to their first encounter with the patient. The body should be covered from head to toe with single use hood, gown, double gloves, gown (or coveralls), impermeable boot covers, and a respirator or N95 mask. Ideally extended cuff cloves, gowns with thumb hooks, and/or tape securing the gowns and gloves should be used. During contact with the patient, the practitioner should limit contact with surfaces with body fluid, keep their hands away from their face, and employ frequent alcohol-based scrubs on their gloved hands. There should be no rush or shortcuts taken while placing the PPE, so time consideration should be given if plans for procedures or intubation are required. Total time spent in the care of the patient while wearing PPE should be monitored because providers can become fatigued and dehydrated. Elective surgical procedure should be postponed in a patient with EVD and nonoperative alternatives should be strongly considered for urgent conditions (ie, appendicitis, perforated ulcer, etc). A risk-benefit analysis considering the risk of exposure to the operative team must be considered prior to operating for an emergent condition bearing in mind that, in late EVD disease, the patient is unlikely to survive.

Disaster Management

Health care preparedness for a real or potential Ebola outbreak mandates the presence of an effective screening procedure that will capture any potential carrier of EVD. A robust program that contains multiple contingency plans and has been rehearsed by all the team members is essential, as disaster leaders should be prepared to operate without significant assistance for at least 72 hours. A simple screening method that can be utilized at any first encounter with a patient (including clerical areas) should focus on the following: (1) recent travel (last 21 days) to an affected area; (2) recent (21 days) direct contact with a patient infected with Ebola.⁴⁵

Positive responses to these simple questions lead to an escalation of care with an immediate separation and isolation of the patient. After proper isolation, a more focused history identifying exposure risk can be performed. Once a diagnosis of EVD is confirmed, the CDC recommended plans should be in place to ensure immediate isolation is available which includes a private room with a door, a private bathroom, and separate areas for donning and doffing PPE. Other pragmatic plans for handling of lab specimens, environmental waste management, and restriction of nonessential personnel must be in place.⁴⁶ The institutional management of threatened or actual Ebola outbreak requires trained, professional emergency managers and disaster management experts to be involved in incident command. Incident control should designate a team of site managers who have a constant presence to oversee implementation of safety precautions, monitor adequacy of necessary supplies, and evaluate care in isolation areas.

The high infectious risk and mortality rate of EVD evokes strong emotions of fear, both in healthcare workers and the general public. The incident command should remain the voice of calm and reason during the disaster and encourage constant communication between the front-line providers, the incident commander, and the community.

General Principles

As in all other aspects of surgical care, it is preferable to try to prevent infections wherever possible. A number of interventions and practices have been demonstrated to be highly effective in preventing infections after elective operations, and many of those techniques have specific application in the care of injured patients. In this section, current evidence-based interventions that are applicable to trauma patients to prevent infection will be reviewed.

Prophylactic Antibiotics

Prophylactic antibiotics are intended to prevent development of infection. The concept of prophylaxis presupposes that infection is not present. The decision to use prophylactic antibiotics and the choice of agents are based on the risk of developing a SSI. There are very good data regarding SSI rates for elective surgery, and the incidence of SSI by wound class for elective operations is shown in Table 18-5. Traditionally, Class I or clean wounds are those that do not violate the respiratory, alimentary, or genitourinary tracts. The wound infection rate is approximately 2%. Class II, or clean-contaminated wounds, refers to elective operations on potentially contaminated organs, such as the gastrointestinal tract, biliary tract, genitourinary tract, and respiratory tree (the procedure will violate a mucosal surface, which can never be completely sterile). The usual incidence of infection for these types of wounds is 5–10%. Contaminated wounds (Class III) differ from Class II wounds by the degree of spillage, with an incidence of infection of 15–30%. Finally, Class IV or dirty-infected wounds are characterized by frank pus or extensive and prolonged contamination. These wounds are characterized by an infection rate of >30% if primary closure is attempted.⁴⁷ A recent report using ACS-NSQIP risk-adjusted data suggested substantially lower superficial and deep SSI rates, although the definitions used for class III and IV wounds are questionable.⁴⁸ Emergent operative interventions increase the wound class by one step, so it is clear that higher rates of wound infection will be encountered in dealing with patients who have acute traumatic injuries that require operative intervention.

In trauma surgery the majority of wounds encountered will be Class III or IV, and the luxury of a preinoculation dose of antibiotics, as recommended by the Surgical Care Improvement Project (SCIP), is usually unavailable.⁴⁹ With this in mind, it is prudent to administer a single dose of an agent with activity against community-acquired aerobic and anaerobic pathogens as soon as possible for all patients requiring operation in the thorax or abdomen. Evidence-based guidelines for antibiotic prophylaxis of other surgical interventions or different anatomic sites are summarized in Table 18-6.

The issue of postoperative continuation of prophylactic antibiotics in penetrating abdominal trauma has been investigated extensively. The Eastern Association for the Surgery of Trauma (EAST) has published guidelines derived from an evidence-based review.⁵⁰ A single preoperative dose of a prophylactic antibiotic with broad-spectrum aerobic and anaerobic coverage is recommended for trauma patients sustaining penetrating abdominal wounds. Absence of an injury to a hollow viscus requires no further administration. If, however, a hollow viscus injury is present, there are sufficient Class I and Class II data to recommend continuation of prophylactic antibiotics for only 24 hours. Timely discontinuation of prophylactic antibiotics is important because the practice of prolonged administration has been linked to increased rates of subsequent nosocomial infections with resistant organisms.⁵¹ To maintain adequate tissue and serum levels of antibiotics in the face of ongoing hemorrhage and vasoconstriction, the administered dose may be increased 2- or 3-fold and repeated after every 10 units of transfusion of blood products, although there is not strong evidence to support this practice.

Surgical Scrub

For many years povidone–iodine scrub was the standard disinfectant used for preparation of surgical sites and preoperative scrubbing by the surgical team. This dominance has been challenged by several well-designed studies performed in elective surgery showing significantly lower SSI rates with the use of chlorhexidine–alcohol compared with iodine (9.5% vs 16.1%).⁵² The fact that chlorhexidine–alcohol begins bacterial killing immediately on contact and does not require drying for antimicrobial effectiveness makes it potentially attractive for use in emergent surgery. One caveat regarding use of alcohol-based disinfectants is that it is imperative that the solutions be dry if electrocautery is used during surgical procedures to avoid intraoperative fires.

Double Gloving

Glove perforation is an underappreciated phenomenon that may adversely impact the sterility of an operative procedure. Microperforation rates as high as 16% have been reported.⁵³ When two pairs of gloves are used, inner glove perforation rate is substantially lower. In addition to patient outcome, the surgeon must consider personal safety. The CDC estimates that more than 2.7 million US citizens (~1%) have chronic HCV (http://www.cdc.gov/hepatitis/HCV/HCVfaq.htm) and that 1.2 million Americans are living with HIV (http://www.cdc.gov/hiv/basics/livingwithhiv/index.html). One epidemiologic study has reported the prevalence of HIV or hepatitis C to be as high as 20–65% and 10–45%, respectively, in an urban university hospital population for patients undergoing lymph node biopsy or drainage of a soft tissue infection.⁵⁴ More recently, Brady et al.⁵⁵ reported that the seroprevalence of undiagnosed hepatitis C virus (HCV) infection was 7.9% with another 7.8% of the population having preexisting HCV infection.

Temperature Control

Hypothermia has been shown to be a strong prognostic indicator of poor outcome when considered in the context of the trauma “triad of death” (hypothermia, acidosis, coagulopathy).⁵⁶ Hypothermia slows enzymatic activity needed for blood clotting and can induce an acquired coagulopathy. Hypothermia can also adversely impact SSI. In elective colorectal surgery a prospective, randomized study compared a group in whom intraoperative normothermia (36.6°C) was maintained to a control group with mild hypothermia (34.7°C). The normothermic group had a significantly lower rate of SSI (6% vs 19%) and a 20% shorter hospital stay.⁵⁷ The precise mechanism for the beneficial effects of normothermia remains unclear, but may relate to tissue perfusion and improved host defense. Therefore, every reasonable effort should be made to maintain normothermia in all surgical patients.

Supplemental Oxygen

The role of oxygen in development of SSI was examined in another prospective randomized study in patients undergoing elective colorectal operations. The authors observed a significantly lower SSI rate when 80% versus 30% inspired oxygen was delivered during the operation.⁵⁸ It is noteworthy that other studies have failed to replicate these results⁵⁹ and that this intervention has never been studied in a trauma population; however, an alternative interpretation of this study in elective patients is that it is best to avoid lower intraoperative F_iO₂.

Suture Material

There have now been several randomized controlled trials demonstrating a decreased incidence of SSI when antibiotic-coated sutures were employed for facial closure. The most rigorous study has involved triclosan-coated sutures, where a 30% decrease in SSI was seen in a large meta-analysis.⁶⁰ Sutures decrease the inoculum of bacteria needed to establish infection and can serve as a foreign body within potentially infected wounds, so there is a sound physiologic basis for a potential benefit. An unanswered additional consideration is whether this intervention is cost-effective.

Blood Transfusion

Autologous blood transfusion can be lifesaving for an exsanguinating patient, but numerous authors have reported worse infectious complications with increased blood utilization both in the immediate resuscitation^61,62,63,64 and when used in a delayed fashion.^65,66,67 Transfusion results in a multitude of immunosuppressive effects, including the following: (1) decreased CD3+, CD4+, and CD8+ cells; (2) overall reduced T-cell proliferation to mitogenic stimuli; (3) decreased natural killer cell activity; (4) defective antigen presentation; and (5) impaired cell-mediated immunity.⁶⁸ The increased risk of infection associated with blood transfusion appears to be dose dependent,^67,69 and logistic regression analyses report that the risk of infection increases 13% per unit transfused.⁷⁰ Taylor et al.⁷¹ reported that for each unit of packed red blood cells (PRBCs) transfused, the odds of developing a nosocomial infection were increased by a factor of 1.5. The age of the transfused blood is an additional risk factor for infectious complications.^72,73,74 As blood ages in the blood bank, it undergoes predictable changes that affect its ability to deliver oxygen. This “storage lesion” includes the following: (1) an increased affinity of hemoglobin for oxygen and reduced oxygen release to tissues; (2) depletion of 2,3-diphosphoglyerate (2,3-DPG) with resultant inadequacy of oxygen transport by red blood cells; (3) reduction in deformability, altered adhesiveness, and aggregability; and (4) accumulation of bioactive compounds with proinflammatory effects. In trauma patients, Offner et al. estimated that each transfused unit of RBC older than 14 days increased the risk of major infection by 13%.⁷⁴

To minimize infectious risk blood transfusion in nonbleeding patients should be avoided. A large multicenter randomized study reported the safety of a restrictive transfusion strategy (trigger of 7.0 g/dL hemoglobin) compared to a liberal strategy (trigger of 10.0 g/dL). In fact, mortality rates patients who were younger (age <55) and less sick (APACHE <20) were more than double with liberal transfusion. When the subgroup of trauma patients (n = 203) was reviewed in a secondary analysis, McIntyre et al. confirmed the safety of the restrictive strategy.⁷⁵ Additionally, practices to be avoided include transfusing multiple units of blood in stable nonbleeding patients, using blood as a volume expander, and transfusing blood preemptively in anticipation of future operative blood loss. Advanced age should not be used as a sole criterion to transfuse a patient. Several recent guidelines describe transfusion of autologous RBCs in trauma patients in the postresuscitation period.^76,77

Nutritional Support

The timing, adequacy, and route of administration of nutrition to trauma patients have definite implications for infectious complications. Adequate nutrition is essential for patient recovery, healing of traumatic wounds, and post-trauma rehabilitation. This is because trauma causes increased metabolism and protein turnover that results in a catabolic state characterized by breakdown of skeletal muscle, impaired healing, and immunosuppression. Once acute resuscitation is complete, nutritional support should be instituted and the enteral route is preferred. Numerous trials have compared enteral nutrition (EN) with parenteral nutrition (PN). Advantages of the enteral route include lower cost, maintenance of function of the gut mucosal barrier, and more physiologic delivery of nutrients. The only real advantage of PN is the consistency of adequate calorie provision when the enteral route is unavailable. Based on numerous high-quality studies, the evidence strongly favors the use of EN over PN in regards to infectious complications.^78,79,80 Traditional markers of nutrition (albumin, transferrin, and retinal-binding protein) are also restored better using EN.⁸¹ An additional factor that has been implicated in infectious complications is the issue of glycemic control. Current recommendations relating to glycemic control⁸² are in flux, but it appears clear that results are improved when hyper- and hypoglycemia are avoided. Enthusiasm for “very tight” glucose control has waned after several studies showed no benefit and increased complications with attempts to maintain blood glucose <110 mg/dL.^83,84,85 Newer consensus documents favor continuous glucose monitoring and dynamic insulin protocols to optimize glucose control.⁸⁶

Tracheostomy

Although several studies comparing early and late tracheostomy have been performed, there is still no consensus regarding whether earlier tracheostomy impacts development of ventilator-associated pneumonia (VAP).^87,88,89 The decision to perform tracheostomy is often institution and surgeon specific. A recent Cochrane Database meta-analysis showed that mortality was decreased slightly (relative risk, RR = 0.83) and that the likelihood of discharge from the intensive care unit (ICU) was higher at day 28 (RR = 1.29) with early versus late tracheostomy.⁹⁰

Diagnosis of Infection

Almost by definition, infectious problems are rarely the presenting complaint of a patient who has sustained an acute traumatic injury. Recognition of an infection in a patient recovering from traumatic injuries, however, continues to be a common, sometimes challenging, clinical problem. Frequently, many of the signs and symptoms commonly associated with infections are also present in trauma patients. The immediate physiologic and immunologic response to tissue injury is initiation of the inflammatory response. Acute traumatic injuries cause the cardinal signs of inflammation including pain (dolor), edema (turgor), heat (calor), redness (rubor), and loss of function (functio laesa). Trauma victims will often manifest several of the SIRS criteria (eg, tachycardia, elevated temperature, elevated WBC) in a clinical context where they are at increased risk for infection.⁹¹

The diagnosis of infection requires a high clinical suspicion tempered by knowledge of the most likely infectious complications at various time points after injury and filtered by experience with caring for patients with similar injuries. As discussed in the earlier section on “Prevention of Infections,” the preferred management of infections is to prevent their occurrence. When prevention measures have been ineffective, the diagnosis of infection is based on clinical, laboratory, and imaging methods. Most trauma patients, especially those requiring operative interventions, those with open fractures, or those who have sustained penetrating trauma, will be treated with a course of empiric antibiotics (Table 18-7). The choice of specific antimicrobial agents is determined by the endogenous pathogens or likely exogenous contamination at the anatomic site of injury (eg, exposure to Vibrio spp. with marine trauma, exposure to Clostridium sp. with farm injury).

Hospitalized patients are at risk for development of nosocomial infections, and the diagnosis is made on the basis of a high suspicion coupled with laboratory and/or radiologic confirmation. In most cases, bacterial culture constitutes the “gold standard” for diagnosis of infection, although at times it can be impractical or impossible to obtain adequate samples. The most specific culture information, if available, is obtained with quantitative or semi-quantitative methods (eg, burn wound biopsy, bronchoscopic alveolar lavage [BAL], or nonBAL). In cases where cultures cannot be obtained, empiric treatment is initiated based on the most likely pathogenic organisms and adjusted based on clinical response. Evidence-based recommendations have been developed for diagnosis and antimicrobial treatment of most hospital-acquired infections,^{92,93,94,95,96,97} albeit not specifically addressing trauma patients. The trauma Glue Grant (www.gluegrant.org) proposed a standard operating procedure (SOP) to identify the source of infection in critically ill trauma patients.⁹⁸ An updated adaptation of the SOP is reproduced in Fig. 18-4. This approach emphasizes and prioritizes the most likely infections and suggests acceptable antimicrobial agents that are otherwise consistent with evidence-based guidelines from infectious disease societies for a wide range of different infections. Diagnostic imaging, particularly cross-sectional imaging, and/or ultrasound, can be helpful to identify and access potentially infected fluid collections in deep locations. Percutaneous aspiration and/or drainage has largely replaced the need for operative management of deep infections and abscesses.^99,100

Intra-Abdominal Infections

Intra-abdominal infections, particularly an abscess, are infectious complications of both blunt and penetrating abdominal trauma. The presence of very high numbers of bacteria within the gut, coupled with the impaired perfusion present in shock states and the immune alterations associated with trauma, result in a 10–25% incidence of intra-abdominal infection. Appropriate resuscitation and empiric antibiotics, along with sound intraoperative decision making, minimize the risk for infectious complications. Early source control (see section on Operating Room) is crucial to limit the likelihood for postoperative infections. Primary infection is relatively rare, but may be the presenting complaint in patients in whom an injury to a hollow viscus was unrecognized. Clinical examination may provide sufficient information to warrant abdominal exploration, and the presence of peritoneal signs or other signs of an acute abdomen should not be ignored in trauma patients. Peritonitis and acute abdominal signs may develop if bacterial contamination has been present for >12 hours. Diagnostic studies such as an abdominal CT scan will often suggest the diagnosis (eg, unexplained free fluid or free intra-abdominal air is highly suspicious for injury to a hollow viscus). Diagnostic imaging is much more important to identify postoperative abdominal infections, since physical examination may be equivocal in awake patients and it may be impossible to adequately examine sedated trauma victims in the ICU.

Intra-abdominal infections identified later (>4 days postinjury) in the hospital course are much more likely to be caused by hospital-associated, rather than community-associated, organisms. This is because most patients will have received one or several doses of (frequently relatively broad-spectrum) antibiotics that alter the bacterial flora.¹⁰¹ Late intra-abdominal infections may grow Pseudomonas spp., Serratia spp., or Candida spp. Percutaneous aspiration and/or drainage has become the mainstay for treatment of late intra-abdominal abscesses and fluid collections. Whenever possible, material from intra-abdominal collections should be sent for Gram stain, culture, and sensitivity determination because of the high incidence of resistant bacteria in recent series.¹⁰² The Surgical Infection Society (SIS) and the Infectious Disease Society of American (IDSA) published updated evidence-based guidelines to inform antibiotic choices for intra-abdominal infections.¹⁰³ These guidelines differentiate between community- and hospital-acquired infections (Tables 18-8 and 18-9). As a general rule, antimicrobial coverage directed against community abdominal pathogens may be given within the first 3 days after injury, whereas antibiotic choices ≥4 days after injury should anticipate hospital organisms.

Empyema

Under normal circumstances the pleural cavity has a net negative pressure and a very small (<20 mL) volume of fluid.¹⁰⁴ Current understanding of fluid flux within the pleural cavity implicates parietal pleural lymphatics as the main route through which fluid is removed.¹⁰⁵ Pleural fluid normally turns over at a rate of ~0.15 mL/(kg h) (10–12 mL/h for a 70-kg individual), and the maximum capacity for lymphatic drainage is estimated to be 700 mL per day (~30 mL/h). When there is increased production or decreased clearance of pleural fluid, a pleural effusion develops. Pleural effusions are frequently associated with fluid overload, but can also arise in the setting of acute inflammatory processes of the lung including pneumonia (para-pneumonic effusion), ALI, or ARDS. Blunt and penetrating chest trauma can induce alterations within the pleural cavity, with loss of negative pressure (traumatic pneumothorax), accumulation of blood (hemothorax), or a combination of both (hemopneumothorax). Treatment of either condition generally requires insertion of a thoracostomy tube, and this can introduce bacterial contamination of the fluid or blood present within the pleural cavity. Since lymphatic stomata of the parietal pleural are the means by which blood and/or fluid is normally reabsorbed, the presence of a fibrin clot can obstruct this route of egress and contribute to persistence of fluid within the chest cavity. As discussed above in “Prevention of Infections,” the presence of bacteria, blood, foreign bodies, and unexpanded lung can predispose to infection within the pleural cavity or empyema.

Diagnosis of empyema requires sampling of fluid or tissue from the pleural space.¹⁰⁶ In the setting of empyema analysis of pleural fluid will demonstrate a pH <7.20, glucose <40, the presence of bacteria on Gram stain, or a positive culture.¹⁰⁷ Table 18-10 shows the most common bacterial isolates from post-traumatic empyema (although some obviously purulent collections may not grow bacteria). It is now generally understood that post-traumatic empyema arises from exogenous contamination of the pleural cavity. This was demonstrated clearly by Hoth et al.¹⁰⁸ who obtained simultaneous bronchoalveolar lavage and pleural cultures and noted that there was minimal correlation between pleural cultures and BAL samples. Recognition that skin flora is associated with empyema underscores the importance of using sterile technique during insertion of a thoracostomy tube. Studies examining risk factors for development of empyema have identified the duration of drainage through a chest tube and incomplete evacuation of hemothorax as two of the leading factors.¹⁰⁹ Retained hemothorax and empyema complicate about 4% of patients with a hemothorax that was treated with a thoracostomy tube.¹¹⁰

Several modalities have been employed to treat empyema.¹⁰⁷ Complete evacuation of a hemothorax and re-expansion of the lung at the time the initial chest tube is inserted are important to prevent empyema. Chest CT scans showing pleural thickening or incomplete evacuation may be important adjuncts in managing patients with thoracic trauma. Instillation of fibrinolytic agents has been demonstrated to aid in the evacuation of a retained hemothorax; however, this may be dangerous in the presence of an intra-thoracic injury. There is now considerable enthusiasm for early use of video-assisted thoracoscopy (VATS), which has proven to be safe and more cost-effective than a second thoracostomy tube.^110,111 Ideally, VATS should be performed within 7 days of injury, as the rates of empyema and conversion to thoracotomy are increased after 1 week.¹¹²

Osteomyelitis/Septic Arthritis

Trauma-related osteomyelitis and septic arthritis represent not uncommon complications of musculoskeletal injury.¹¹³ Post-traumatic musculoskeletal infections generally arise from bacteria introduced exogenously, either at the time of injury or during operative repair. This is in contrast with hematogenous dissemination of bacteria with most nontraumatic bone and joint infections. Although a variety of bacterial species have been isolated from post-traumatic osteomyelitis and septic arthritis (Table 18-11), Staphylococcus species are far and away the most common isolates.^{114, 115} S. aureus and S. epidermidis have a number of virulence factors that provide a particular predilection to bone tissue.¹¹⁵ These virulence factors are summarized in Table 18-12, but include adhesive properties and exotoxins and enzymes that facilitate invasion. Another property that has been recognized as being etiologically important to the development and persistence of osteomyelitis is a small colony variant (SCV) phenotype that grows more slowly and has increased resistance to aminoglycosides and decreased hemolytic activity. Clinical use of aminoglycoside beads and broader-spectrum antibiotics may select for these more resilient SCV phenotypes in vivo.

The presence of contaminating bacteria at the site of a bone or joint injury incites a vigorous local inflammatory response. The local source of the inflammatory cytokines TNF, IL-1, and IL-6 is not entirely clear, but osteoblasts may contribute. In any case, high local levels of proinflammatory cytokines have dramatic effects on bone turnover and new bone formation. TNF and IL-1 induce increased maturation of osteoclasts and enhance osteoclastic activity.¹¹⁶ At the same time these mediators inhibit mesenchymal cell differentiation into osteoblasts. Similar processes contribute to destruction of cartilage and bone in chronic arthritic conditions. In such settings anti-TNF and anti-IL-1 therapies have been very successful in preventing inflammation.¹¹⁷ The net impact of increased osteoclast and decreased osteoblast activity is either bone destruction or inhibition of bone healing. Furthermore, contact between osteoblasts and S. aureus can induce TNF-related apoptosis-inducing ligand (TRAIL), which, in the presence of Fas-associated death domain (FADD), commits cells to apoptotic cell death.

Several unique factors associated with bone infection underscore the importance of evidence-based recommendations that emphasize early and continuing aggressive debridement of orthopaedic injuries, particularly those associated with open or contaminated wounds. At sites of bone or joint injury local blood supply to fragments may be interrupted predisposing to bone necrosis. Lack of blood supply also precludes delivery of systemic antibiotics, host inflammatory cells, and the molecular oxygen needed for oxygen-dependent bactericidal activity of such cells. It is axiomatic that all dead bone must be removed to minimize the risk of bone infection. In addition, many orthopaedic injuries require the presence of foreign bodies, such as plates or rods, to stabilize fractures. Experimentally only very low levels of bacteria are needed to produce infection in the presence of foreign material. In the presence of extensive local bony destruction and/or contamination many orthopaedic surgeons utilize external fixation devices to minimize, but not eliminate, the risk of infection and nonunion.

Soccupational Exposure to Environmental Pathogens

It is an unfortunate reality that health care workers in general, and surgeons in particular, are exposed to blood-borne occupational hazards, including hepatitis B virus (HBV), HCV, and HIV. Body fluids considered potentially infectious include blood, cerebrospinal fluid, synovial fluid, pleural fluid, peritoneal fluid, pericardial fluid, and amniotic fluid. Conversely, exposures to nonbloody feces, nasal secretions, saliva, sputum, sweat, tears, urine, and vomitus are not considered potentially infectious, as the risk for transmission of HBV, HCV, and HIV infection from these fluids is extremely low.¹¹⁸ PPE demonstrated to be effective include double gloving, using blunt needles for fascial closure, protective eye shields, impervious surgical gowns, and routine implementation of universal precautions.

While prevention is obviously the best course of action, certain measures taken after an exposure can decrease the risk of sero-conversion. The first step involves treatment of the exposure site. Wounds and skin sites should be washed with soap and water and mucous membranes flushed with water. There is no evidence to support applying antiseptics to the wound or expressing fluid to reduce the risk of transmitting a blood-borne pathogen. The CDC strongly discourages the application of caustic agents (eg, bleach) or the injection of antiseptics or disinfectants into the wound. Second, the exposure source (ie, the patient) should be evaluated for their HBV, HCV, and HIV status. If the status is unknown, the patient should be informed of the incident. The health care practitioner must be aware of applicable state and local laws regarding informed consent for serologic testing. Testing of needles or other sharp instruments involved in the exposure is not acceptable as a replacement or complement to testing the source patient. Additional up-to-date resources and recommendations are available via the National Clinician’s Postexposure Prophylaxis (PEP) Hotline (PEPline, 888-448-4911 or via http://nccc.ucsf.edu/clinical-resources/pep-resources/pep-guidelines/).

When an exposure to HBV occurs, the risk of sero-conversion is dependent on the degree of contact (ie, size of the inoculum) and the hepatitis B e-antigen (HBeAg) status of the source. For example, if the patient is HBeAg positive, the risk of developing HBV infection is about 50%, compared to 25% if the patient is HBeAg negative. PEP includes hepatitis B immune globulin (HBIG) and, possibly, the hepatitis B vaccination series, depending on the hepatitis B antigen status of the patient and the antibody status of the at-risk health care worker. If indicated, HBIG should be given as soon as possible, since early administration after exposure to hepatitis B surface antigen-positive blood can provide an estimated 75% protection from HB infection.¹¹⁸

In contrast to HBV, HCV is not efficiently transmitted via occupational exposure. It is estimated that HCV sero-conversion after accidental percutaneous exposure from an HCV-positive source occurs 1.8% of the time, and some have suggested that transmission occurs only from puncture by hollow-bore needles. Transmission has never been reported after intact or nonintact skin exposure to blood and only rarely occurs after exposure of mucous membranes to blood. Currently, intravenous immune globulin (IVIG) is not recommended after occupational HCV exposure. The rationale is based on several clinical observations including the following: (1) HCV infection does not incite a protective antibody response; (2) studies of IVIG use for PEP in HAV and HBV cannot be extrapolated to HCV; and (3) HCV IVIG use in chimpanzees has failed to prevent HCV sero-conversion after exposure.¹¹⁸ There is no evidence that the administration of INF-α or antiviral agents prevents HCV infection after occupational exposure, and their use is not currently recommended. Currently there is no information available regarding use of new direct-acting oral agents in this setting. The exposed health care worker should be tested for baseline HCV viral status and continue close follow-up for 12 months for the purpose of early identification should sero-conversion occur. Additionally, the health care worker should contact the CDC Hepatitis Information Line (888) 443-7232 or http://www.cdc.gov/hepatitis/hcv/index.htm.¹¹⁸

Like HCV, transmission of HIV occurs rarely after occupational exposure. The CDC estimates the risk of sero-conversion at approximately 0.3% following a percutaneous exposure to HIV-infected blood and 0.09% after exposure of a mucous membrane. For exposure to fluids or tissues other than HIV-infected blood, the risk of transmission has not been quantified, but is probably much lower. In 2013, the US Public Health Service (PHS) updated the recommendations for HIV PEP to include three agents (Raltegravir plus Truvada”(Tenofovir + emtricitabine).¹¹⁹ To maximize the possibility of protection, PEP should be initiated as soon as possible and the health care worker exposed to HIV should be evaluated within hours. A baseline HIV test should be performed, and HIV antibody testing should be performed for at least 6 months postexposure (at 6 weeks, 12 weeks, and 6 months). A 4-week regimen is advised for most HIV exposures. If the source person’s virus is known to be resistant to the routine PEP regimen, selection of an alternate regimen is highly recommended.¹¹⁸

Resuscitation Bay

Efforts to minimize infection must be initiated as soon as the patient arrives in the trauma bay. Although the initial focus will appropriately center on control of hemorrhage and initiation of resuscitation, these efforts will reduce the risk of infection, as well. Restoration of adequate blood flow and oxygen delivery is the first step in reducing the incidence of infection. It has been clearly shown that the incidence of infection from invasive procedures in the ICU such as insertion of a central venous line or chest tube are dramatically decreased by employing full-barrier precautions.¹²⁰ Full-barrier precautions may not always be practical in severely injured victims, but suspension of proven infection control measures and sterile technique for invasive procedures should be the rare exception. Maintenance of normothermia likewise represents optimal treatment for the trauma patient and has the additional benefit that it will decrease the likelihood for development of infectious complications.⁵⁷ Contaminated wounds should be cleaned and/or formally debrided in an urgent time frame, although more recent reviews of experience in high-energy orthopaedic injuries have downplayed the importance of time to debridement as the most important factor to prevent osteomyelitis.¹²¹ A tetanus booster should be administered to all patients, unless it is known that they received a booster dose within 5 years.

Prophylactic or empiric antibiotics, if indicated, should be initiated in the trauma bay. Antibiotics should be started in all patients with penetrating injuries or open fractures and in any patient in whom there is a high likelihood of injury to a hollow viscus. Due to the hostile environment in which military injuries occur, current recommendations are that injured soldiers start oral or parenteral antibiotics as soon as possible if there is any break in the skin. Patients with blunt mechanisms who require operative interventions should receive perioperative antibiotic prophylaxis according to established evidence-based guidelines. Antibiotic choices for different anatomic regions are compiled in Table 18-7.

Operating Room

The conduct of operative interventions can also significantly impact the likelihood for postoperative infectious complications. The primary factor(s) determining the risk for infection is the nature and magnitude of the traumatic injury requiring surgical intervention. The operating surgeon must remember that the highest priority during an exploratory laparotomy for hemorrhagic shock is control of bleeding. For example, if a patient sustained abdominal gunshot wounds with multiple enterotomies and an injury to the inferior vena cava and right renal hilum, it is highly likely that the surgeon would encounter blood and massive intestinal contamination upon entering the abdomen. The initial focus must be on control of hemorrhage, and this may require resisting the surgical impulse to stop enteric leakage. The analogy is similar to the principle applied to ATLS prioritization of exsanguinating abdominal hemorrhage in a patient with a severe traumatic brain injury. The best way to save the brain (or to minimize infection) is to control the hemorrhage!

After acute hemorrhage is controlled it is appropriate to stop ongoing leakage from the bowel and to try to remove most gross contamination. In addition to the bacterial contamination arising from enteric leakage, disruption of gastrointestinal integrity also releases foreign bodies (eg, undigested food particles), mucin, and bile. Any or all of these so-called adjuvant substances significantly decrease the bacterial inoculum needed for clinical infection. Adjuvant substances enhance bacterial infectivity via two mechanisms. First, some adjuvant substances augment bacterial growth or stimulate bacteria to express virulence factors.¹²² The second mechanism involves interference with host defense mechanisms such as the function of innate immune cells.¹²³ For example, bile’s detergent activity can lyse polymorphonuclear leukocytes and macrophages. Blood and devitalized tissue are two additional adjuvant factors that are frequently present. Blood (specifically hemoglobin) can be converted to a leukotoxin by some enteric bacteria.¹²⁴ Fibrin clots sequester bacteria and make them inaccessible to host phagocytes, and this may predispose to late development of intra-abdominal abscesses.^125,126 Devitalized, ischemic, or necrotic tissues are also potent sources of damage/danger signals that activate host immune cells and exacerbate acute inflammation. At the same time these substances can interfere with phagocytosis and oxidative killing mechanisms of the host defense cells.^9,127 Thus, to the extent possible it’s desirable to remove most blood and blood clot from the peritoneal cavity. This can be accomplished by irrigating the peritoneal cavity with a goal of removing the obvious contamination, foreign material, and blood. It is worth noting that a prospective randomized study showed no benefit to formal meticulous debridement to remove fibrinous debris from the peritoneal cavity in established peritonitis.¹²⁸ While it is a popular aphorism that “the solution to pollution is dilution,” there is scant evidence to back up this bias. Experimentally, the best approach involves using the least amount of irrigation sufficient to remove gross contamination and adjuvant material.

Evidence-based guidelines for antimicrobial prophylaxis of trauma recommend broad-spectrum agents with activity against the anticipated pathogens that are likely to be encountered at the anatomic area of injury.¹²⁹ The guidelines are based on the principle that, “it is always best to anticipate the worst case scenario.” With respect to the infectious risk after blunt or penetrating abdominal injury, this requires coverage against colonic bacterial flora. In the case of abdominal trauma, agents with activity against aerobic and anaerobic bacteria are recommended.^129,130,131 With injuries to the extremities the most likely pathogens will be aerobic gram-positive bacteria, particularly Staphylococcus species. Injury to maxillofacial structures requires antibiotic prophylaxis with activity against normal oral flora, and neurosurgical procedures most often employ agents similar to those used for the extremities.

Little is known about the pharmacology of antibiotics in the acute resuscitative phase of trauma. Most available pharmacologic data derive from healthy individuals and, therefore, cannot be applied directly to injured patients. Buijk et al.¹³² described a cohort of 89 critically ill patients who received aminoglycosides. In these septic patients the volume of distribution was significantly higher than in those without septic shock and the maximum antibiotic concentration achieved was significantly lower. In a study of patients who required significant resuscitation with fluids and blood during a laparotomy, the volume of distribution was expanded and correlated with the degree of fluid resuscitation. Antibiotic elimination was also more rapid in these injured patients compared with normal controls.¹³³

With massive blood loss antibiotic prophylaxis requires frequent re-dosing to maintain plasma and tissue levels above the mean inhibitory concentration (MIC). Animal models of experimental infection after hemorrhagic shock report better prophylaxis with increasing doses of appropriate intraoperative antibiotics.¹³⁴ Large volume resuscitation and altered endothelial permeability with trauma or burns result in an expanded volume of distribution. Renal dysfunction from hypovolemia, myoglobinuria, or radiologic contrast often accompanies severe injury, but the potential risk of a nephropathy has no impact on acute antibiotic dosing. The greatest risk for subsequent infectious complications arises from under rather than over-dosing in acute trauma.

The conduct of the operation itself significantly impacts the chances for survival and the risk of infection. Abundant data from elective surgery underscores the importance of maintaining normothermia, avoiding shock, and minimizing use of blood transfusion.^{57,65,135,136} Damage control resuscitation, emphasizing acute management of immediate life-threatening injuries, has the collateral benefit of positively impacting the incidence of postoperative infectious complications.¹³⁷ In most cases vascular reconstruction and bowel anastomoses, if needed, should be delayed until the patient is warm, adequately resuscitated, and hemodynamically stable. While the emphasis is on acute management of life-threatening injuries, this approach has been shown to decrease intra-abdominal complications. Most surgeons continue empiric prophylactic antibiotics if wounds are temporarily closed, although this approach has not been formally evaluated. The evidence is insufficient to recommend for or against the use of prophylactic antibiotics in the management of an open abdomen.

If the patient is hemodynamically stable, euvolemic, and normothermic, then there is no adverse impact to definitively managing abdominal, vascular, neurosurgical, or orthopaedic injuries during the first operation. In a grossly contaminated wound primary closure is associated with an unacceptably high wound infection rate. Delayed primary closure (DPC), a practice dating back to Ambrose Pare, was advocated by surgical pioneer John Hunter in the 1700s and popularized during World War I.¹³⁸ It is based on development of fine granulations within the wound prior to definitive closure. Thus, DPC combines the infective resistance of healing by secondary intention with the cosmetic benefits of primary closure. Randomized prospective trials have reported significantly lower rates of wound infections compared to primary closure of grossly contaminated wounds.^139,140,141

The management of penetrating wounds to the colon has evolved since World War II, when the diverting ostomy reduced mortality rates to about 30% in the preantibiotic era. With improvements in trauma resuscitation and accumulating experience with antibiotics, investigators began to question whether fecal diversion was necessary after colonic repair. Initially, Stone and Fabian¹⁴² published the first prospective randomized trial of colostomy versus primary repair. They excluded patients with “high-risk” criteria such as shock, hemorrhage, greater than two organs injured, gross contamination, operative delay greater than 8 hours, injury requiring resection, and loss of the abdominal wall. Subsequent investigations have reported that primary colon repair, even in the face of “high-risk” criteria, is associated with a decreased incidence of infectious complications compared to diverting ostomy.^143,144,145

Open pelvic fractures are associated with extremely high morbidity and mortality, mostly from septic complications. Diverting colostomies are often placed in such patients, but the evidence supporting this approach is weak and generally derived from small retrospective studies. One systematic review found no difference in the overall infection rate with or without colostomy, with the exception of a lower complication rate when colostomy was used for perineal/rectal wounds.¹⁴⁶ More studies are required to provide definitive recommendations.

Recommendations for management of extremity fractures continue to evolve. A large study found that the most important factor in outcome was early transfer to a trauma center for definitive management.¹²¹ Specifically, this study called into question the benefit of early operative debridement of open fractures, inasmuch as time to operative debridement did not confer a statistically significant benefit. The concept of damage control has also been applied to orthopaedic injuries.¹⁴⁷ The military experience in Iraq and Afghanistan also underscore the utility of this concept, coupled with more liberal use of external fixation devices to minimize infectious complications in this hostile environment.¹⁴⁸ With proper debridement and acute management of the fracture, bacteria inoculated into fractures at the time of injury are rarely isolated from postoperative infections. Rather, hospital-acquired flora are almost universally responsible for infections in fractures.⁹³

ICU and Early Postoperative Period

Infectious complications are commonly encountered in the early postoperative period and are even more likely to be seen in patients who require ongoing critical care (Table 18-13). Clinicians should have a high index of suspicion and consider potential sources of infection in the context of the injuries, characteristics of the patient, likely offending organisms, and length of hospital stay. The immediate postoperative course of a severely injured patient typically involves a vigorous SIRS response.⁹¹ Clinical signs such as mild to moderate fever and tachycardia are almost universal. Furthermore, it is now known that danger signals released from injured tissue activate innate immune responses via the same TLR pathways stimulated by bacterial infection. Leukocytosis is a common manifestation of the SIRS response, so monitoring the white blood cell count immediately after injury has little utility. It is useful to keep the overall trajectory of the patient in mind before automatically initiating a series of expensive and low-yield investigations (eg, blood cultures, urinalysis, x-rays). Identifying infection is particularly difficult in the ICU setting.^97,149 Injured patients are at increased risk for development of nosocomial infections such as pneumonia, central line associated bloodstream infections (CLABSI), catheter-associated urinary tract infections (CAUTI), Clostridium difficile infection (CDI), and SSI; however, critically ill trauma victims are susceptible to severe sepsis and septic shock, as well. The Surviving Sepsis Campaign has emphasized the importance of prompt recognition, aggressive resuscitation, and early institution of broad-spectrum antimicrobial agents in their recently updated evidence-based guidelines (Table 18-14).⁹⁷ In the sections that follow the diagnosis and treatment of common postinjury/postoperative infections that are frequently encountered in trauma patients are discussed.

Pneumonia

The discussion regarding the definition of pneumonia is voluminous and the result is controversial and changing. Currently, pneumonia can be divided into community-acquired pneumonia (CAP) and hospital- or health care-associated pneumonia (HCAP), the former being present on admission and the latter manifesting later in the hospital course. HCAP can further be divided into early (<4 days) and late (≥4 days). These distinctions are more than merely pedantic, as pneumonia that develops later is more likely caused by multidrug resistant (MDR) organisms. Patients with CAP or early HCAP who have no known risk factors should be treated with ceftriaxone or levofloxacin (or equivalent) or ampicillin/sulbactam, or ertapenem.¹⁵⁰ in accordance with local prevalence patterns.

Within the broader category of HCAP an important subgroup are patients on mechanical ventilation who develop pneumonia. Ventilator-associated pneumonia (VAP) is a leading cause of morbidity and mortality in the injured population and remains a daunting diagnostic challenge.^138,139 The definition of VAP is very nonspecific, does not correlate well with histopathologic findings of pneumonia, and is a poor quality measure in the ICU. In 2013, in an effort to improve the surveillance of noninfectious etiologies of pulmonary dysfunction in mechanically ventilated patients, the NHSN released the guidelines for detecting ventilator-associated events (VAE).¹⁵¹ This new guideline replaces the older definition of ventilator-associated pneumonia (VAP) and is intentionally broad as to encompass all pulmonary issues affecting critically patients including ARDS, atelectasis, and cardiogenic sources of pulmonary edema. The new definition consists of a three-tiered progression from ventilator-associated condition (VAC) to infection-related ventilator associated condition (IVAC) to probable ventilator-associated pneumonia (PrVAP). This is a shift away from reporting only on the infectious complications that affect mechanically ventilated patients and is thought to improve comparability, facilitate automation, and minimize gaming in the reporting of pulmonary dysfunction in ICU patients.

The first tier of VAE is a VAC. This occurs when a patient (ventilated for at least 2 days) has >2 days of worsening oxygenation (defined by >0.2 increase in the minimum daily F_iO₂ or PEEP increase >3 cm H₂O). The second-tier VAE is the infectious ventilator-associated condition (IVAC) that occurs when the criteria for VAC have been met PLUS the presence of (1) fever, (2) abnormal WBC (>12 or <4 × 10³/mm³) AND the initiation of antibiotics for >4 days. The third VAC tier consists of possible VAP (PoVAP) and probable VAP (PrVAP). PoVAP is considered when there is the presence of purulent secretions and/or positive cultures. PrVAP must meet the specific diagnostic requirements for culture growth based on the method in which the samples were obtained. The first two VAE tiers of VAE are intended as quality metrics while the third tier is meant to be used for internal quality improvement only.

Despite the intentions of the improved surveillance and reporting structure of VAEs, there has been little data to support its effectiveness. Indeed, several studies have elucidated the inherent issues within this definition structure. In a retrospective analysis of over 1300 patients from 11 North American ICUs, Muscadere et al. found a 10.5% VAC rate, 4.9% IVAC rate, but an 11.2% VAP rate (based on traditional definitions).¹⁵² Further, the kappa (inter-rater reliability) was very low (VAP and VAC 0.18 and VAP and IVAC 0.19). Additionally, the VAC rates did not change with the increase in preventative measures. They did find that VAC and IVAC were associated with a significant increase in morbidity and mortality.

In the intensive care unit there are abundant data focusing on the utility of measures designed to prevent VAP.¹⁵³ All postoperative patients are at risk for pulmonary complications and, therefore, aggressive mobilization and pulmonary toilet should be employed when possible. Unfortunately, injuries such as rib fractures that interfere with coughing and deep breathing, orthopaedic injuries that limit mobility, or traumatic brain injuries result in an altered mental status that complicates pulmonary toilet. Simple measures such as routine hand hygiene and elevation (30–45°) of the head of bed to decrease the likelihood of reflux have resulted in dramatic decreases in VAP rates. Reverse Trendelenberg positioning is an acceptable alternative in patients who must remain supine. Subglottic suctioning and chlorhexidine oral hygiene have been shown to be effective though to a lesser degree, and all of these measures have been incorporated into “ventilator bundles.”¹⁵³ The duration of mechanical ventilation is the greatest risk factor for development of VAP, with an estimated incidence of 1.2–3.5% risk of developing pneumonia per day of mechanical ventilation.¹⁵⁴ Every effort should be made to achieve liberation from the ventilator as soon as possible. Patients should be assessed daily for a trial of spontaneous breathing, and extubation should occur if the patient successfully passes the trial.^155,156,157

The diagnosis of pneumonia can be challenging in critically ill trauma patients because physical examination is often limited in the obtunded or sedated patient. In addition, lung contusions or atelectasis complicate interpretation of chest x-rays and bacterial colonization of the endotracheal tube and trachea is universal after a few days of mechanical ventilation. The clinical pulmonary infection score (CPIS) initially described by Pugin to compare invasive and noninvasive pulmonary sampling techniques for VAP¹⁵⁸ has been used to assist in the diagnosis of VAP, but subsequent trials have shown that CPIS alone is not sufficiently accurate to diagnose or rule out VAP.¹⁵⁹

Diagnostic sampling of the respiratory tract will greatly assist in differentiating tracheal colonization from true infection and in guiding antibiotic therapy. Three modalities of diagnostic sampling of the lower respiratory tree, BAL, bronchoscopic protected specimen brushing, and blind “mini-BAL,” are currently employed and the specimens are sent for quantitative or semi-quantitative microbiologic culture.¹⁶⁰ Some controversy remains as to the optimal mode for sampling, and the quantitative criteria differ depending on the diagnostic method of diagnosis. A positive confirmatory culture is defined as a BAL or mini-BAL culture >10⁴ CFU/mL or protected specimen brush >10³ CFU/mL. A true pneumonia will also manifest as a new radiographic infiltrate, increased oxygen requirements, and the presence of thick, copious secretions. These diagnostic criteria must be satisfied within a 48-hour period.

When pneumonia is identified, it is important to institute therapy as soon as possible. Current recommendations for antimicrobial coverage for pneumonia, and especially VAP, emphasize beginning with broad-spectrum coverage and then de-escalating or narrowing the coverage once culture results are obtained. Inadequate initial antibiotic therapy has repeatedly been shown to be associated with worse outcomes and increased mortality.^{97,161,162,163} Again, knowledge of local antibiograms should be used when choosing initial therapy. Patients with late HCAP or VAP are at risk of having MDR bacteria including Pseudomonas, Acinetobacter, and extended-spectrum beta-lactamase producers such as Klebsiella and E. coli, especially if previously exposed to antibiotics. Gram-negative coverage should include antipseudomonal agents such as piperacillin/tazobactam, cefepime, imipenem, or meropenem. The addition of an aminoglycoside for “synergistic effect” against Pseudomonas is not supported by current guidelines. With the increasing prevalence of MRSA, vancomycin should always be included as initial empiric coverage. Once pneumonia has been diagnosed and therapy initiated, the next question is the appropriate duration of therapy. An oft-quoted study by Chastre et al.¹⁶⁴ reported that an 8-day antibiotic course was equivalent to 15-day therapy for VAP. The foremost consideration should be the clinical status of the patient, with a good response consisting of resolution of elevated temperature, resolving leukocytosis, decreased respiratory secretions, and improved oxygenation. Radiographic resolution often lags behind clinical improvement.

Catheter-Associated Urinary Tract Infection (CAUTI)

Catheter-associated urinary tract infection (CAUTI) is the most common hospital-acquired infection and is almost universally associated with an indwelling urinary catheter.^165,166,167 Most injured patients in the ICU require urinary drainage, and differentiation between urinary colonization and infection can be difficult. The diagnosis of CAUTI requires a quantitative urine culture with >10⁵ CFU/mL along with at least one of the following clinical criteria: (1) temperature > 38.5°C; (2) WBC >10K or <3K; and (3) urinary urgency, dysuria, or suprapubic tenderness. Diagnosis can be made with >10³ CFU if there is also pyuria (>3 WBC/hpf) or a positive leukocyte esterase. These factors must be present within a contiguous 48-hour period. Clinically significant CAUTI usually occur in the setting of urinary trauma and/or repair. Responsible organisms in hospitalized patients include E. coli, Pseudomonas, Proteus, Enterobacter, Serratia, and Citrobacter.¹⁶⁸ Gram-negative coverage is indicated. Fortunately, many systemic antibiotics are excreted via the kidney and will achieve urinary levels that far exceed the MICs for most of these pathogens.

Central Line-Associated Bloodstream Infection (CLABSI)

Central line-associated bloodstream infections (CLABSI) have decreased in incidence in recent years, but they remain a source of serious morbidity, increased hospital and ICU stay, increased costs, and potential death.^169,170 The decision to insert a central line should not be taken lightly. Indications for central access include poor peripheral access, administration of TPN, and administration of high-dose vasoactive medications. Large-bore introducers are also commonly inserted to assist in the acute resuscitation of unstable patients. Often, the clinician is faced with the difficult decision to remove necessary central access or maintain the catheter in the face of potential line sepsis.

In trauma patients in whom catheters were placed emergently using nonsterile technique, the lines should be removed and replaced, if needed, with aseptically inserted catheter(s) at a new site(s). Femoral venous catheters are associated with unacceptably high rates of both infection and deep vein thrombosis, and catheters at these sites should always be removed as soon as possible. The use of peripherally inserted central catheters (PICC) is associated with a lower incidence of CLABSI. Routine guidewire exchange at predefined intervals does not decrease the rate of CLABSI and may even increase the incidence. Using an evidence-based approach Pronovost et al.¹⁷¹ reported a 66% sustained reduction in the incidence of CLABSI by removing unnecessary catheters, avoiding the femoral site, using full-barrier precautions during insertion, hand washing, and cleaning the skin with chlorhexidine. The data regarding routine antibiotic impregnated catheter use is not clear; however, if the local CLABSI rate exceeds the National Health care Safety Network surveillance data, then consideration should be given to using these types of catheters.¹⁷²

The diagnosis of CLABSI is one of exclusion and requires the presence of bacteremia or fungemia in a patient in whom there is no alternate source of infection. Unless there are local signs of infection (ie, redness and purulence), a search for other infectious etiologies should always be performed first. Infected catheters are frequently thrombogenic and the first sign of infection may be an inability to aspirate blood through the port. In addition to meeting SIRS criteria, the patient must have microbiologic evidence of catheter infection as follows: (1) positive semi-quantitative culture (>15 CFU/cm catheter segment); (2) quantitative (>10³ CFU/cm catheter segment) culture with the same organism isolated from blood cultures; and (3) simultaneous quantitative blood cultures with greater than 3:1 CVC to blood ratio of the same bacterial species.

SSI are infections arising at the site of a previous surgical procedure, defined as a location in which an incision had been made or a procedure performed.^173,174 Classification of SSI is based on the anatomic depth of the infection and whether the infection is present in the wound or within an organ space. Superficial and deep incisional SSI are differentiated based on whether the infection extends below the fascial layers (deep SSI). By convention, SSI are infections identified within 30 days of the initial surgical procedure.

Superficial SSI are not infrequent in trauma patients based on the wound classification, disruption of environmental barriers, bacterial inoculation at the time of injury, and dysfunction of host defenses seen with injury. Diagnosis of a superficial SSI is determined by the presence of at least one of the following clinical criteria at the site of a surgical procedure: (1) purulent drainage from the surgical incision; (2) culture of organisms from an aseptically obtained fluid or tissue sample from the incision; (3) clinical signs or symptoms of infection; or (4) clinical diagnosis of infection by the surgeon (eg, the need to open wound). Conditions such as stitch abscesses and erythema or serous drainage at external fixator pin sites do not constitute superficial SSI. In contrast to superficial SSI, a deep SSI involves the deeper soft tissues (eg, fascial and muscle layers) at the site of the surgical incision and ≥1 of the following: (a) purulent drainage deep to the fascia or muscle layers; (b) spontaneous fascial dehiscence; or (c) identification of a deep abscess on direct examination, during reoperation, or radiologic examination. Finally, an organ space SSI involves anatomic structures (eg, organs or spaces) that were manipulated during the surgical procedure. In addition, diagnosis of an organ space SSI is based on at least one of the following: (1) purulent drainage from a drain placed into the organ or space (either incisional or percutaneously drained); (2) culture-positive fluid or tissue; or (3) the presence of an abscess during reoperation or radiologic evaluation.

Treatment of SSI depends on the location and depth of infection.^98,175,176 Superficial SSI are treated by opening of the surgical incision followed by local wound care. There is no demonstrated benefit to systemic antibiotic treatment for superficial SSI in the absence of systemic symptoms. For deep SSI and organ space infections, it is advisable to try to obtain cultures of any purulent drainage, since prior antibiotic selection pressure and the increasing incidence of resistant strains within hospitals makes it difficult to predict antimicrobial responses.⁹⁸ In the face of negative cultures, a not infrequent occurrence, clinicians should base antibiotic choices on the anticipated pathogens at the anatomic site. It is well to keep in mind that infections that arise after treatment with a longer course of antibiotics will almost certainly be resistant or, at best, only partially sensitive to the initial agent used. Thus, while awaiting culture and sensitivity results, it is wise to employ a different antibiotic agent or even class.

Clostridium difficile Infection

C. difficile is a gram-positive anaerobic bacterium that is a frequent cause of infectious colitis.^177,178 It is a part of the normal colonic flora in 2–5% of the healthy population and is normally nonpathogenic. The incidence in the trauma population has been reported to be 3%.¹⁷⁹ Disruption of the colonic microbiota by antibiotics causes relative overgrowth of this organism. Although the occurrence has been reported with all antibiotics, the association is highest with clindamycin and third-generation cephalosporins. The organism reproduces via spore formation and should be considered highly contagious, as the spores are heat and stomach acid resistant. Additionally, alcohol-based hand sanitizers are ineffective in eradicating the spores. Soap and water hand washing is mandatory after contact with a contagious patient. Since 2001, there has been a significant increase in the incidence of C. difficile infection to approximately 84 per 100,000 and this has coincided with an increased number of serious or fatal infections. The higher failure rates being reported (18.2%) with metronidazole therapy have prompted several professional societies to recommend vancomycin as a first-line agent for patients with severe infection (at a dose of 125 mg four times per day, per os). Vancomycin can also be administered per rectum by enema (500 mg four times daily). For milder infections, oral metronidazole remains the preferred treatment because of its lower cost. For both severe and mild infections, intravenous or oral metronidazole (500 mg four times daily) may be given.

Surgical therapy for C. difficile infections was historically limited to performing a subtotal colectomy in patients who developed severe-complicated infection unresponsive to medical therapy. The result was often associated with a high mortality. More recently there has been considerable interest in a different surgical approach using laparoscopic placement of a loop ileostomy for antegrade flushing of the colon. Intraoperatively the colon is lavaged with warm polyethylene glycol and postoperatively, vancomycin is instilled via the ileostomy antegrade through the colon. In a small trial the mortality of this approach was 19% versus a 50% in historic controls of subtotal colectomy. Furthermore, in the loop ileostomy group, 83% of the operations were done laparoscopically and there was over 90% salvage rate of the colon.¹⁸⁰ There are still no clear indications for determining which patients should undergo this type of surgical intervention; however, contraindications include colonic necrosis or perforation as well as the presence of abdominal compartment syndrome.¹⁸¹

Recurrence rates after initial treatment with metronidazole or vancomycin are 20% and may be the result of re-infection with a different strain of C. difficile or persistence of the initial strain. If the latter, the recurrent episode may be treated with the same agent used to treat the initial episode. After the first recurrence, the risk of a second recurrence is 40% and 60% after two or more recurrences. There are no standard recommendations for treatment of multiple recurrences; however, one randomized, double-blind placebo-controlled trial of monoclonal antibodies to C. difficile toxins A and B reported significantly lower recurrence rates among the treatment arm when compared with placebo (7% vs 25%, P >0.001).¹⁶⁵ Another therapeutic modality for recurrent C. difficile infection that is gaining popularity is fecal microbiota transplantation (FMT) to repopulate the normal colonic flora. The concept for FMT is not new, and several retrospective reports have demonstrated significant cure rates for recurrent infections with FMT.^182,183 A prospective clinical trial was initiated to attempt to assess and compare three groups as follows: (1) vancomycin followed by colonic lavage and then FMT delivered via a nasoduodenal tube; (2) vancomycin followed by colonic lavage; and (3) vancomycin alone.¹⁸⁴ That study found such a dramatically superior cure rate with FMT that the trial was stopped early. FMT is currently not approved by the Federal Drug Administration, but multiple methods are in development to more easily deploy the FMT in oral or pill form.

Postsplenectomy Vaccinations

The optimal timing of administration of vaccines after traumatic splenectomy is unknown, but data suggest an increasing trend for elevated functional antibody activity with a delay in vaccination and improved immune antibody response to vaccination at 14 days after surgery.¹⁸⁵ The Advisory Committee on Immunization Practices (ACIP) recommends the use of the 23-valent polysaccharide pneumococcal vaccine for persons 2–64 years of age who have functional or anatomic asplenia. The CDC suggests a single booster dose for those older than 2 years of age who are at high risk for serious pneumococcal infection and those most likely to have a rapid decline in antibody titers. A single revaccination should be given at least 5 years after the first dose, with further dosing not recommended routinely. H. influenzae vaccination with ActHIB conjugate vaccine is recommended. The available meningococcal vaccine protects against serotypes A, C, Y, and W-135 of Neisseria meningitidis.¹⁸⁶ There currently is no recommendation to revaccinate for H. influenza type b or meningococcus. Most trauma patients will be discharged from the hospital much earlier, and given that the population of trauma patients is characterized by poor follow-up, it is probably best to vaccinate the patient prior to discharge rather than wait 14 days (Table 18-15).

Overwhelming Postsplenectomy Infection (OPSI)

Historically, the spleen was considered expendable and the prevailing opinion was that it could be removed with relative impunity. Our current understanding is that the spleen plays an important role in the production of immune mediators that aid in the clearance of bacteria and viruses. Splenic mediators (opsonins) coat circulating bacteria and viruses and convert them into immune complexes, facilitating clearance.^187,188 It is difficult to estimate the current incidence of OPSI, as most of the published data on OPSI antedate the widespread availability of the pneumococcal and H. influenzae vaccines. The time since splenectomy is an important risk factor as 50–70% of admissions to the hospital for serious infections occur within the first 2 years. A recent retrospective cohort study of 20,132 patients who underwent splenectomy for a variety of indications between 1979 and 2009 found an overall incidence of hospital admission for sepsis of 5.7% (95% CI: 5.6–6.0). These investigators found that the incidence of sepsis and mortality was lower if splenectomy occurred secondary to trauma versus a hematologic malignancy. Interestingly they did not observe a major impact following initiation of routine vaccination, except in the subjects undergoing splenectomy for hematologic conditions. Pneumococcal infections account for the majority of reported cases while H. influenzae type b, N. meningitides, and Group A Streptococcus account for an additional 25%.¹⁸⁷

Surgical Site Infection After Discharge

In the modern health care environment there is increased pressure for earlier hospital discharge. While this practice is reported to be safe and clearly decreases costs, it complicates identification of SSI. In a prospective study of 268 patients, 33% of all patients with SSI were diagnosed after hospital discharge.¹⁸⁹ Trauma caregivers should educate patients about signs and symptoms of SSI at the time of discharge and request that they be notified if infections are identified. In the future it is conceivable that shared electronic health records may aid in identification of late SSI.