Infections, Antibiotic Prevention, and Antibiotic Management
Philip S. Barie
Soumitra R. Eachempati
I. Epidemiology
Incidence. The incidence of infection following injury approaches 25%. Although most trauma-related deaths within the first 24 hours after injury are from exsanguination or injury to the central nervous system, the leading cause of post-traumatic death after 24 hours is infection, often through the multiple organ dysfunction syndrome (MODS). The high risk of infection is due to the host immune response to injury and stress; challenges of infection control under emergency conditions; direct inoculation of wounds by clothing, dirt, or debris; blood transfusions; catabolism and resultant protein–calorie malnutrition; and poor glycemic control. Appropriate antibiotic prophylaxis reduces the risk, but inappropriate prophylaxis may increase the risk of infection. Early definitive surgical source control (of drainable or resectable foci) and timely, appropriate empiric antibiotic therapy are key to successful management.
Patterns of injury. Infections following injury occur in the injured tissue, the surgical site (incision), or as a health care–associated infection (HAI) such as pneumonia or central line–associated blood stream infection (CLABSI) (Table 14-1). Considered together, HAIs are as common as infections of injured tissues. The likelihood of infection is higher with increasing injury severity score (ISS), increasing number of abdominal organs injured, traumatic brain injury, colon injury, shock, increasing number of blood transfusions, prolonged mechanical ventilation, or creation of an ostomy. Traumatic wounds have devitalized, ischemic tissue, with increased risk of infection if contaminated by enteric contents (e.g., penetrating abdominal trauma), fragments of clothing fabric (e.g., gunshot wounds), dirt or gravel (e.g., motor vehicle or farm injuries), or vegetation (e.g., fall from height into a tree). More wound contamination increases the risk of infection of injured tissue.
Comparison with critically ill surgical patients (non-trauma). The epidemiology of HAI is changing among critically ill patients, with high (but decreasing) incidences of pneumonia and CLABSI, and fewer urinary tract infections (UTIs) and surgical site infections (SSIs). The epidemiology of infection following trauma differs from other critically ill surgical patients. Trauma patients are both more likely to become infected (Table 14-1) and develop infection earlier post-injury. Pneumonia is the most common HAI following injury. The timing of onset of infection may influence the choice of antimicrobial therapy.
II. Risk Factors
The host is put at risk of invasion by microbial pathogens whenever a natural epithelial barrier (e.g., skin, respiratory tract mucosa, gastrointestinal mucosa) is breached. Colonization of natural epithelial barriers occurs even in healthy hosts. However, invasion does not occur unless injury or some other mechanism of inoculation occurs. Injury, catheterization, or incision creates a portal for tissue invasion by pathogens. Potential pathogens are ubiquitous in the environment. Innate immunity provides continuous surveillance against invasion by foreign antigens, and stimulates a repair response (inflammation), which may result in counterproductive augmentation of the inflammatory response that is destructive to the host. Prolonged or severe inflammation (e.g., the systemic inflammatory response syndrome, SIRS) (Table 14-2) is associated with the MODS.
Injury severity. Shock and higher ISS increase the risk of infection globally. Thoracoabdominal penetrating injury is associated with a higher risk of infection than either abdominal or thoracic injury alone. The risk of intra-abdominal infection is higher with increasing numbers of abdominal organs injured. Several “local” injuries induce systemic immune, inflammatory, and coagulation responses, including pulmonary contusion and traumatic brain injury (TBI), the latter being the injury most associated with infection, especially pneumonia.
Table 14-1 Rates of Health Care–associated Pneumonia and Catheter-related Bacteremia Among Various ICU Types
ICU type
CVC usea Infection rate
TT useb Infection rate
Mean/median
Mean/median
Medical
0.52
5.0/3.9
0.46
4.9/3.7
Pediatric
0.46
6.6/5.2
0.39
2.9/2.3
Surgical
0.61
4.6/3.4
0.44
9.3/8.3
Cardiovascular
0.79
2.7/1.8
0.43
7.2/6.3
Neurosurgical
0.48
4.6/3.1
0.39
11.2/6.2
Trauma
0.61
7.4/5.2
0.56
15.2/11.4
aNumber of days of catheter placement/1,000 patient-days in ICU.
bNumber of days of indwelling endotracheal tube or tracheostomy/1,000 patient-days in ICU.
Infection rates are indexed per 1,000 patient-days.
Based on the National Nosocomial Infection Surveillance System, US.
Centers for Disease Control and Prevention. (From Bercker S, Weber-Carstens S, Deja M, et al. Critical illness polyneuropathy and myopathy in patients with acute respiratory distress syndrome. Crit Care Med 2005;33:711–715.)
Table 14-2 Immune Dysfunction after Trauma
Specific immunity
- Lymphopenia
- Helper: Suppressor T-cell ratio,1
- Downregulated:
- –T-, B-cell proliferation
- –NK cell activity
- –IL-2 receptor expression
- –IL-4, −10 production
- –HLA-DR expression
- –DTH skin test response
- –T-, B-cell proliferation
Nonspecific immunity
- Monocytosis
- Upregulated:
- –Acute-phase proteins
- –TNF, IL-6 production
- –Eicosanoid production
- –Acute-phase proteins
- Downregulated:
- –Neutrophil function
NK, natural killer cell; IL, interleukin; HLA-DR, human leukocyte antigen; DTH, delayed topical hypersensitivity; TNF, tumor necrosis factor.
- Lymphopenia
Immune dysfunction
The immune response to injury is immediate and complex (Table 14-3). The consequences are immediate activation of:
Coagulation as a result of endothelial dysfunction and activation of platelets
Mononuclear and polymorphonuclear leukocytes causing release of both pro- and anti-inflammatory cytokines and activation of host defenses against microbial invasion
Eventual depression of innate and adaptive immunity with host immunosuppression and predisposition to later HAI
Inflammation and the stress response. The stress hormone response that characterizes the “fight or flight” response:
Augments cardiovascular function through the sympathetic nervous system
Enhances glycogenolysis
Mobilizes peripheral lean muscle and fat as fuel (catabolism)
Enhances coagulation to stanch hemorrhage
Stimulates a pro-inflammatory cytokine response to begin the process of tissue repair (Table 14-4). Innate and adaptive immunity are depressed in large part by the actions of cortisol (Table 14-5).
Table 14-4 Overview of the Stress Response to Injury
Activation of the sympathetic nervous system
Activation of hypophyseal–pituitary–adrenal axis
Peripheral insulin resistance
Production of pro- and anti-inflammatory cytokines
Acute-phase changes of hepatic protein synthesis
Recruitment and activation of neutrophils, monocyte/macrophages, and lymphocytes
Upregulation of procoagulant activity
Table 14-5 Principal Hormonal Responses to Surgical Stress
Endocrine gland
Hormones
Change in secretion
Anterior pituitary
ACTH
Increased
Growth hormone
Increased
TSH
Variable
FSH/LH
Variable
Posterior pituitary
AVP
Increased
Adrenal cortex
Cortisol
Increased
Aldosterone
Increased
Pancreas
Insulin
Decreased
Glucagon
Increased
Thyroid
Thyroxine
Decreased
Triiodothyronine
Decreased
ACTH, adrenocorticotropic hormone; TSH, thyroid-stimulating hormone; FSH, follicle-stimulating hormone; LH, luteinizing hormone; AVP, arginine vasopressin.
Medical comorbidity. Both very young and elderly patients are at increased risk of infection (Table 14-6). Obesity, malnutrition, diabetes mellitus, hypocholesterolemia, hypothermia, and chronic kidney disease also increase risk of infection.
Table 14-6 Conditions Known to Increase the Risk of Infection
- Extremes of age
- Malnutrition
- Obesity
- Diabetes
- Prior site irradiation
- Hypothermia
- Hypoxemia
- Remote infection
- Corticosteroid therapy
- Recent operation, especially of the chest or abdomen
- Chronic inflammation
- Hypocholesterolemia
- Extremes of age
Transfusion
Blood transfusion cannot be avoided in many, but trauma patients are more than five-fold more likely to develop infection if a blood transfusion is administered. Several theories exist as to why transfusion predisposes to infection:
Transfusion may be immunosuppressive through leukocyte antigen–mediated decreases in innate immunity, specifically a shift to the Th2 (immunosuppressive) response phenotype.
Augmentation of the SIRS response.
Mechanically, the “storage lesion” that develops after 2 weeks of storage in the blood bank depletes erythrocyte 2,3-diphosphoglycerate and cell membrane stores of adenosine triphosphate, decreasing oxygen delivery (DO2) and RBC deformability, respectively. The latter lesion impairs the ability of the RBC to transit the microcirculation, resulting in rouleaux formation and mechanical obstruction of the microcirculation.
Each unit of RBC infusion results in the administration of about 200 mg free iron, which is vital for microbial growth. However, it is unclear whether iron administration per se increases the risk of infection.
Transfusion principles are detailed elsewhere in this text.
Hyperglycemia
Hyperglycemia was viewed as inevitable following critical illness and injury as a consequence of the counter-regulatory stress hormone response, the mobilization of glucose through glycogenolysis, and the subsequent mobilization of amino acids for glucose synthesis through catabolism of lean tissue. We recognize that hyperglycemia (>200 mg/dL) is associated with an increased incidence of SSI following cardiac or major general surgery.
Despite these observations, little priority was assigned to prevention of hyperglycemia until the publication in 2001 by van den Berghe et al. of a prospective trial of tight glucose control (80 to 110 mg/dL) in critically ill surgical patients
(mostly cardiac surgery) by continuous infusion of insulin. However, the NICE-SUGAR trial noted that tight glucose control as above increased mortality in critically ill patients when compared to a target glucose of <180 mg/dL (different from “no glucose control”), primarily through excess cardiac events.
Glucose dyshomeostasis has several manifestations during stress (Table 14-7).
Peripheral glucose uptake and utilization are increased.
Glycogenolysis is depressed after initial, short-term mobilization of hepatic glycogen stores.
Gluconeogenesis is increased.
Peripheral insulin resistance.
Importantly, hyperglycemia impairs immune cell function.
Neutrophils may activate spontaneously, with increased generation of adhesion molecules, impairing microcirculatory flow.
Insulin-stimulated chemokinesis is decreased.
Phagocytes manifest decreased respiratory burst and thus impaired microbial killing.
Given that the stress response is stereotypical and pervasive among critically ill and injured surgical patients, there is indirect evidence that glucose control with insulin therapy seeking <180 mg/dL is as important for trauma patients as it is for surgical patients.
Table 14-7 Glucose Dyshomeostasis During Stress and Effects on Cellular Immunity | |
|---|---|
|
III. Prevention of Infection
Principles. Infection prevention tactics must be utilized as an ensemble; the list of options is not a buffet from which the clinician may select. Infection control is paramount, but often under-emphasized. Traumatic wounds must be cleansed thoroughly and debrided to remove devitalized tissue. Surgical incisions must be handled gently, inspected daily, and dressed if necessary using aseptic technique. Avoid drains and catheters when possible and remove as soon as practical. Use antibiotics in a focused manner to minimize emergence of antibiotic multi–drug-resistant (MDR) pathogens.
Infection control
Infection control is an individual responsibility as well as a responsibility of the trauma team and trauma unit. Hand hygiene is the single most effective means to reduce the spread of infection. Yet, if adherence to hand washing is studied, it is invariably lacking. To be effective, hand cleansing with soap and water requires a minimum of 30 to 45 seconds. Alcohol gel hand cleansers are as effective as soap and water (the notable exception being spores of Clostridium difficile, against which alcohol is ineffective), and compliance is
higher. Use universal precautions (i.e., cap, mask, gown, gloves, and protective eyewear) whenever there is a risk of splashing of body fluids (at all times in the trauma bay, and commonly in the ICU).
The sources of the bacteria causing infection are the patients’ endogenous flora, and skin surfaces, airways, gut lumen, wounds, catheters, and inanimate surfaces within the patient’s room (e.g., bed rails, bedside commodes, and computer terminals do become colonized).
Whether infection develops is determined primarily by the response of host defenses, as many organisms that cause infection following injury are inherently avirulent (e.g., Candida, Enterococcus, Pseudomonas). The fecal–oral route is the most common manner by which auto-infection develops, but health care workers hasten the transmission of pathogens around a unit, with colonizing pathogens recovered from the surfaces of personal communication devices, stethoscopes, and neckties, although defective hand hygiene is the key offender.
Contact isolation is an important part of infection control and should be used in targeted patients with methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci (VRE), or MDR gram-negative bacilli (GNB).
Appropriate catheter care includes:
Avoidance of insertion when non-essential.
Appropriate hand cleansing and barrier protection of the caregiver at all times.
Appropriate skin cleansing and barrier protection of the patient during insertion.
Selection of the proper catheter.
Proper dressings while catheters are indwelling.
Removal as soon as possible when no longer needed, or if inserted under less than ideal circumstances. The benefit of the information gained by catheterization or the relief afforded the patient must always be weighed against the risk of infection.
Any indwelling catheter carries a risk of infection, but non-tunnelled central venous catheters pose the highest risk, including local site infections and blood stream infections (Table 14-1). Other catheters that have a substantial risk of infection include:
Thoracostomy tubes or catheters (particularly if inserted as an emergency procedure).
Ventriculostomy catheters for monitoring of intracranial pressure.
Urinary bladder catheters.
Each day of endotracheal intubation and mechanical ventilation increases the risk of pneumonia (by about 3%/day for the first week, and 1%/day thereafter. It is controversial whether tracheostomy to facilitate pulmonary toilet and decreased work of breathing decreases the risk of infection.
In terms of preventing infection, abdominal drains are the most superfluous.
Whenever possible, skin preparation should be with chlorhexidine solution with or without alcohol, which is viricidal and fungicidal as well as bactericidal. Extensive evidence-based guidelines exist for prevention of catheter-related infection.
Chlorhexidine is superior to povidone–iodine solution for skin preparation prior to central venous catheter insertion. When povidone–iodine solution is used, it must be allowed to dry; it is not bactericidal when wet.
Full barrier precautions (i.e., cap, mask, sterile gown, sterile gloves, eye protection, and a large field drape) are mandatory for all bedside catheterization procedures except arterial and urinary bladder catheterization, for which sterile gloves and a sterile field suffice.
The site of central venous catheter insertion influences the risk of infection; the femoral vein site carries a higher risk of infection that either the subclavian or internal jugular vein sites, and should be avoided whenever possible.
Anytime a deep catheter is inserted under less than ideal conditions as described (e.g., a central venous catheter placed during a trauma or cardiac
resuscitation) it should be removed and replaced elsewhere (if needed) within 24 hours.
A single dose of a first-generation cephalosporin (e.g., cefazolin) may prevent some infections following emergency tube thoracostomy or all ventriculostomy placements, but is not indicated for vascular or bladder catheterizations. Topical antiseptics placed post-procedure at the insertion site are of no benefit.
The choice of catheter may play a role in decreasing the risk of infection with endotracheal tubes, central venous catheters, and urinary catheters.
An endotracheal tube with an extra lumen that opens to the airway just above the balloon, to facilitate the aspiration of secretions that accumulate in an area that cannot be reached by routine suctioning, below the vocal cords but above the balloon on the endotracheal tube (subglottic secretions), can decrease the incidence of ventilator-associated pneumonia (VAP) by one-half. An endotracheal tube coated with ionic silver reduces the incidence of VAP.
Antibiotic- (e.g., minocycline/rifampin) or antiseptic-coated central venous catheters (e.g., chlorhexidine/silver sulfadiazine) are effective in reducing the incidence of CLABSI; the catheter coated with minocycline/rifampin appears to be most effective.
Urinary bladder catheters coated with ionic silver reduce the incidence of catheter-related bacterial cystitis.
Dressings must be maintained clean, dry, and intact. Maintaining an intact dressing is difficult when the patient is agitated or the body surface is irregular (e.g., the neck [internal jugular vein catheterization] as opposed to the chest wall [subclavian vein catheterization]]).
A simple gauze dressing is best. Occlusive transparent dressings can accumulate moisture beneath that is a usable growth medium for residual skin flora, which re-colonize the skin anyway within a few hours.
Mark the dressing clearly with the date and time of each change.
Dressing carts or trays should not be brought from patient to patient; rather, sufficient supplies should be kept in each patient’s room. Inanimate objects (e.g., stethoscopes, scissors) are potential transmission vectors if not cleansed thoroughly after contact with each patient.
Evaluate indwelling catheters daily for its continued utility and remove as soon as possible.
Most abdominal drains do not decrease the risk of infection. On the contrary, the risk is probably increased because the catheters hold open a portal for invasion by bacteria and soon become a “two-way street.” Other than for hepatic or pancreatic injuries or discrete collections of purulent fluid, abdominal drains are seldom useful. Closed suction drains should not be left in proximity to intestinal suture lines; the negative pressures generated, particularly when such drains are “stripped,” may cause disruption.
Antibiotic prophylaxis
Shock, hypoperfusion, and hemorrhage complicate the pharmacokinetics of prophylactic antibiotic administration immediately following trauma. Shock, hypovolemia, and hypoperfusion also increase the risk of organ dysfunction caused by antibiotics (e.g., aminoglycosides and renal injury).
With modern β-lactam antibiotics, the drug needed in the tissues is easily attained with conventional prophylactic doses (e.g., cefazolin 1 g, cefoxitin 1 to 2 g) unless the patient is morbidly obese or bleeding briskly.
IV. Duration
Antibiotics with short elimination half-lives (e.g., cefazolin and cefoxitin) should be re-dosed intraoperatively (e.g., every 3 hours for cefoxitin and every 4 hours for cefazolin) to ensure that tissue concentrations remain adequate during the vulnerable period when the incision is open. SSI and only SSI is prevented by antibiotic prophylaxis. Antibiotic prophylaxis more than 24 hours beyond injury
increases the risk of nosocomial infection. Thus, antibiotic prophylaxis in trauma must not extend beyond 24 hours except perhaps for grade III open fractures.
Numerous randomized prospective trials have shown that 12 to 24 hours of antibiotic prophylaxis for penetrating abdominal trauma is equivalent to 5 days of prophylaxis, even when a colon injury is present, provided surgery is performed within 12 hours of injury. Penetrating abdominal trauma with no intestinal injury requires only a single preoperative dose of antibiotic prophylaxis.
Vascular catheter insertion procedures do not require antibiotic prophylaxis. Other catheter insertion procedures require only a single dose of prophylactic antibiotics, except perhaps for emergency tube thoracostomy (up to 24 hours). Indwelling catheters otherwise should never receive prolonged antibiotic prophylaxis.
V. Specific Injuries
Abdominal injury
The data are unequivocal that prophylaxis of no more than 24 hours of a second-generation cephalosporin (e.g., cefoxitin) is equivalent to a longer course (e.g., 5 days) for penetrating abdominal trauma with injury to a hollow viscus, provided the surgery is performed within 12 hours of injury. Penetrating trauma that does not injure a hollow viscus needs only a single dose of antibiotics given prior to operation.
Although not as well studied, the principle is similar for blunt abdominal trauma; if managed non-operatively, no antibiotics are required. If surgery is performed, the duration of prophylaxis (a single dose or 24 hours of prophylaxis) is determined by the pattern of injury.
The abdomen may be left open temporarily as part of damage control or to prevent or manage the abdominal compartment syndrome. There is no evidence that the open abdomen requires antibiotic prophylaxis even if a prosthesis is employed as part of the temporary closure. Another dose of prophylactic antibiotic aimed against skin flora (e.g., gram-positive cocci [GPC]) is appropriate at the time of abdominal wall closure or reconstruction.
There is no evidence that prophylactic antibiotics are required if the liver and spleen are embolized as part of the non-operative management of blunt trauma to those organs.
The infection risk associated with the late post-splenectomy period (including bacteremia from encapsulated GPC) is genuine for children but low for adults. Experts often recommend prophylaxis with oral penicillin until age 18 years for splenectomized children, but adults do not require long-term antibiotic prophylaxis.
All individuals who undergo splenectomy should receive the polyvalent pneumococcal vaccine, with booster doses at 5-year intervals. Some experts advocate co-administration of vaccines against Haemophilus influenzae and Neisseria meningitidis; optimal timing of booster doses is unknown. Also unknown is whether patients who have undergone splenic embolization or the increasingly rare splenorrhaphy procedure should be vaccinated against post-splenectomy sepsis. Vaccination of the embolized patient following splenic injury is our practice, especially for children.Related posts:
Full access? Get Clinical Tree
