Prevention

Prevention of surgical site infections involves meticulous aseptic technique, appropriate timing of prophylactic antimicrobials, tight perioperative glucose control, and strict adherence to a hand hygiene policy.

Diagnosis

Surgical site infections typically present with localized cellulitis (erythema, warmth, and tenderness) and purulent discharge. With deep infections there may be sternal instability, chest pain, and systemic upset. Wound swabs and blood cultures (two sets) should be obtained before antimicrobials are given. If mediastinitis is suspected, a contrast computed tomography (CT) scan of the chest may be performed to confirm the diagnosis.¹⁸ However, most surgeons elect to proceed directly to reexploration, as it is diagnostic, provides tissue or pus for culture, and is therapeutic.

Microbiology

Staphylococci, either S. aureus or a coagulase-negative staphylococcus (mainly S. epidermidis), are responsible for the majority of postoperative sternal wound infections, with gram-negative bacilli accounting for most of the remainder.^19–21 More than 50% of leg wound infections are polymicrobial; multiple enteric gram-negative organisms are commonly identified.¹⁹ Polymicrobial mediastinal infection implies a serious breach of aseptic technique, such as that which may occur during lifesaving chest reopening in the ICU. Mediastinitis due to perforation of the esophagus or trachea is also likely to be polymicrobial, commonly involving enteric gram-negative bacilli and oropharyngeal streptococci, gram-negative coccobacilli, and anaerobes.

Treatment

For superficial cellulitis, oral antimicrobial therapy for 5 to 7 days is usually sufficient. If pus is draining from the wound, the skin sutures may be removed to facilitate drainage. If there is copious pus, sternal instability, severe pain, or systemic upset, intravenous antimicrobials should be given and reexploration considered. At the completion of reexploration, the sternum is usually rewired over multiple mediastinal drains. The drains may be continuously irrigated with antimicrobial or antiseptic solution for several days.²² If there is extensive necrosis of the sternum and mediastinal tissues, closure of the sternum may be delayed. Once infection has settled and granulation tissue has formed, the wound may be closed directly or, more commonly, via a muscle flap reconstruction.²³ Right ventricular laceration is a rare but usually fatal complication of open sternal wounds.²⁴ A vacuum-assisted closure device may be used to promote the healing of open sternal (and leg) wounds and is an effective bridge between débridement and delayed definitive closure.²⁵ Intravenous antimicrobials should be continued for at least 2 weeks; a longer course may be indicated if there are ongoing signs of infection or osteomyelitis or if a prosthetic valve is in situ. Infected leg wounds occasionally require surgical débridement and either delayed closure or closure by secondary intention (i.e., by the formation of granulation tissue).

Definitions and Causes

Pleural space infections include complicated exudative pleural effusions and empyema. Pleural effusions may be classified as transudative or exudative on the basis of their biochemistry (Table 35-3).²⁶ Transudative effusions are due to disturbances in the capillary Starling forces (see Equation 1-12) caused by conditions such as heart failure and liver cirrhosis. Transudative fluid is typically clear, low in protein, and acellular. Exudative effusions are due to inflammation or infection and are associated with conditions such as bacterial pneumonia (parapneumonic effusions), tuberculosis, connective tissue disorders, lung or pleural malignancy, pulmonary embolism, and postpericardectomy syndrome (see Chapter 20). Exudative effusions may be further classified as simple or complicated (see Table 35-3). Simple exudative effusions result from increased permeability of the pleura to protein and fluid, and they do not necessarily imply infection. Complicated exudative effusions result from invasion of the pleural space by microorganisms. They are characterized by a high neutrophil count, high lactate dehydrogenase concentration (due to neutrophil lysis), and a low glucose concentration and pH (due to anaerobic glucose utilization by bacteria and neutrophils).²⁷ The fluid is cloudy because of the presence of cellular debris. Activation of the coagulation cascade leads to fibrin deposition and the formation of loculations within the effusion. Undrained, complicated effusions can lead to empyema, which is defined as pus within the pleural space. There is marked fibroblast activity, which leads to thickening and fibrosis of the pleura.

Table 35-3 Chemistries of Transudative Pleural Effusions, Exudative Pleural Effusions (Simple and Complicated), and Empyema

Rights were not granted to include this table in electronic media. Please refer to the printed book.

From Maskell NA, Butland RJ: BTS guidelines for the investigation of a unilateral pleural effusion in adults. Thorax 58(suppl 2): ii8-ii17, 2003; Chapman S, Davies RJ: Recent advances in parapneumonic effusion and empyema. Curr Opin Pulmon Med 10:299-304, 2004; Light RW, MacGregor MI, Luchsinger PC, et al: Pleural effusions: the diagnostic separation of transudates and exudates. Ann Intern Med 77: 507-513, 1972.

Even when there is active infection, bacteria are not commonly seen on gram stain or grown from fluid sampled from a complicated effusion or empyema. Neutrophils and other inflammatory cells rapidly clear bacteria from the pleural space, which may reduce the number of bacteria to below the threshold visible on microscopy. Treatment with antimicrobials prior to fluid sampling may affect staining of bacteria for microscopy and may also inhibit microorganism growth in vitro. Some pathogenic microorganisms, notably anaerobes, are fastidious (i.e., have complicated nutritional requirements), which makes it difficult to grow them in vitro.

Community-acquired bacterial pneumonia is the most common cause of empyema. Sometimes empyema arises spontaneously, commonly in patients with lung disease but occasionally in young, healthy people. In the latter case, the causative organism is often one of the Streptococcus milleri group. Other causes of empyema include subdiaphragmatic infection (e.g., cholecystitis); postoperative infection (e.g., mediastinitis or bronchopleural fistula); esophageal perforation (e.g., following surgery or perforation by a foreign body); and infections related to chest trauma or intercostal drain sites.

Diagnosis

Patients with complicated effusions and empyema are typically systemically unwell and show signs of sepsis. They may have productive cough, chest pain, and respiratory distress. The appearances of chest radiographs are variable. Free-flowing effusions produce characteristic changes on erect and supine films (see Chapter 6). Loculated effusions or pleural thickening may produce nondependent pleural-based opacities (Fig. 35-1). There may be signs of the underlying cause of empyema, such as consolidation of the lung (pneumonia) or an air fluid level (bronchopleural fistula). Ultrasound scanning is useful to confirm the presence and volume of fluid in the pleural space, to identify septations, and to guide diagnostic thoracocentesis.²⁸ Thickening and enhancement of the parietal pleura in conjunction with pleural collection as seen on contrast-enhanced CT scan are strongly suggestive of empyema (Fig. 35-2).²⁹ CT scanning may also identify the cause. Septations are not easily seen on CT scans.

Figure 35.1 Erect chest radiograph of a left-sided nondependent pleural collection consistent with empyema.

Figure 35.2 Transverse CT scan of a right-sided, loculated empyema. The parietal pleura (arrows) is thickened. Both an anterior and a posterior pleural collection are seen, along with a thinner collection laterally.

Microbiology

Causative microorganisms include Streptococci species (S. pneumoniae, S. milleri group, viridans group); S. aureus; enteric gram-negative bacilli (particularly Klebsiella pneumoniae and Escherichia coli); Pseudomonas aeruginosa; Haemophilus influenzae; and anaerobes.³⁰ With hospital-acquired empyema, similar microorganisms are identified, but it is more common to find Enterococcus species, MRSA, S. milleri, and multi-drug-resistant gram-negative bacilli (particularly P. aeruginosa).³⁰ Occasionally, Candida species are responsible for empyema.

Treatment of Exudative Effusions

Simple parapneumonic effusions usually resolve spontaneously with treatment of the underlying pneumonia. In general, thoracocentesis should be performed in all parapneumonic effusions because clinical differentiation between simple and complicated effusions is unreliable.³¹ Effusions in patients with hypoalbuminemia, heart failure, or atelectasis are at low risk for infection and do not require sampling (but may need draining if they are large). Effusions in patients with either pneumonia or sepsis and effusions that are associated with thickening of the parietal pleura on CT scan or loculations on ultrasound scanning should be sampled.³¹

If thoracocentesis reveals a complicated effusion, it should be drained to prevent progression to empyema.³² If the effusion is relatively free-flowing, tube thoracostomy using a large-bore tube (e.g., 36F) is appropriate. If there are numerous septations or there has not been substantial radiographic resolution of the effusion within 24 hours of tube thoracostomy, intrapleural fibrinolytic therapy may be considered. With this technique, 100,000 IU of urokinase or 250,000 IU of streptokinase in 100 ml of 0.9% saline is instilled into the chest cavity and the tube clamped for 1 to 4 hours. This can be repeated once or twice a day for 2 to 4 days. Systemic fibrinolysis does not occur. The value of intrapleural fibrinolytic therapy, in terms of radiographic and clinical improvement, has been demonstrated in a number of small studies. However, in a recent large, randomized trial, intrapleural streptokinase was not associated with a reduced need for surgery or with improved survival rates.³³ Thus, if the technique is not rapidly effective, patients should undergo surgery. Antimicrobial therapy for complicated effusions is the same as for empyema, described subsequently.

Treatment of Empyema

Once a fluid sample has been obtained by needle thoracocentesis, intravenous antimicrobials should be commenced empirically. For community-acquired empyema, amoxicillin/clavulanic acid or a secondgeneration cephalosporin (e.g., cefuroxime) combined with metronidazole may be used.³⁴ An alternative for patients allergic to penicillins is clindamycin plus a fluoroquinolone (e.g., ciprofloxacin). For hospital-acquired empyema, a carbapenem plus vancomycin is appropriate.³⁴ Antimicrobial therapy should be adjusted on the basis of culture and sensitivity results, but empiric cover against anaerobes should continue because these organisms are commonly present but frequently are not isolated on culture. Aminoglycosides have poor penetration into empyemas and are best avoided. Following drainage of an empyema, antimicrobial therapy should continue for a minimum of 3 weeks.³⁴

Effective drainage of an organized empyema in which there are extensive loculations and pleural thickening requires surgery. There are two main approaches: (1) open thoracotomy and decortication; (2) video-assisted thoracoscopic débridement. The latter is a relatively new technique that appears to be safe and effective in experienced hands.^35,36 During surgery, loculations must be broken down, pus drained, and any thickened visceral pleura excised (decortication) to allow the collapsed lung to reexpand. Specimens for culture should be obtained at the time of surgery and their results used to guide antimicrobial therapy. Decortication of a chronically organized empyema can result in substantial perioperative blood loss. Postoperatively there may be marked systemic inflammatory response, including fever, tachycardia, third-space fluid losses, and vasodilatation. Positive-pressure ventilation may be difficult because of air leaks, acute lung injury, and poor reexpansion of the affected lung.

Postpneumonectomy Empyema

Empyema that occurs within the first few weeks after pneumonectomy is usually caused by a bronchopleural fistula. Patients present with fever, cough, purulent or hemorrhagic sputum, respiratory distress and, occasionally, sepsis and acute respiratory failure. Chest radiographs typically demonstrate increased air content and reduced fluid levels within the pleural space. There may be soiling of the nonoperative lung demonstrated by pulmonary infiltrates on the chest radiograph (see Fig. 12-6). Empyema that occurs more than 3 months after pneumonectomy is usually due to hematogenous spread of infection into the pleural space. Patients typically present subacutely with malaise, flulike symptoms, low-grade fever, and weight loss.

The diagnosis of a bronchopleural fistula is usually confirmed by bronchoscopy or a contrast-enhanced CT scan of the chest. Alternatively, methylene blue can be instilled into the pleural space and recovered from the sputum. Needle aspiration of the thoracic space is rarely helpful in the acute setting, but if it is undertaken, it should be guided by ultrasound. Two sets of blood cultures should be obtained before starting antimicrobial therapy. Causative organisms and empiric antimicrobial therapy are as described for hospital-acquired empyema. In particular, antimicrobials against MRSA and P. aeruginosa must be administered until culture results are available. Surgical treatment of a bronchopleural fistula is described in Chapter 12.

Pathophysiology and Risk Factors

The initiating event of infective endocarditis is the formation of fibrin-platelet thrombus on damaged endothelium or prosthetic material.³⁸ With bacteremia, microorganisms seed on the thrombus. Bacterial replication and further thrombosis lead to the formation of infected vegetations (see Fig. 7-11). Clinical disease may take several weeks to develop following a bacteremic episode. Infection may result in localized tissue destruction, which can cause valve regurgitation, abscesses, or fistulae. Abscesses (see Fig. 7-14) and fistulae typically involve the aortic root. Very occasionally, large vegetations cause valve stenosis. There may be systemic emboli, including cerebral and pulmonary emboli. Pulmonary emboli arise from right-sided vegetations, whereas cerebral emboli arise from left-sided vegetations. Systemic embolization resulting from right heart lesions can occur in a patient with an intracardiac shunt or a patent foramen ovale. Right-sided lesions are associated with injection drug use. The risk for infective endocarditis with various cardiac lesions is summarized in Table 10-7; congenital cardiac lesions and rheumatic heart disease underlie most cases.

Microbiology

Gram-positive microorganisms are responsible for the majority of cases of infective endocarditis. They include (1) streptococci, particularly viridans streptococci (e.g., Streptococcus sanguis, S. mutans, S. mitis); β-hemolytic streptococci (e.g., S. pyogenes, S. agalactiae); and nonhemolytic streptococci (e.g., S. bovis); (2) staphylococci, particularly Staphylococcus aureus and coagulase-negative staphylococci (primarily S. epidermidis); (3) enterococci (e.g., Enterococcus faecium, E. faecalis). Viridans streptococci and S. aureus cause most cases of native valve endocarditis, whereas S. aureus and coagulase-negative staphylococci are important causes of prosthetic valve endocarditis occurring early after implantation (Table 35-4). Less common (<10%) causes of infective endocarditis include gram-negative bacilli, fungi, and the HACEK group of microorganisms. The HACEK group are fastidious, slow-growing, gram-negative coccobacilli from the oropharynx which, although constituting only 3% to 4% of infective endocarditis, are increasing in prevalence. They are composed of Haemophilus species, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, and Kingella kingae. As much as one third of cases of infective endocarditis are culture-negative.

Diagnosis

The diagnosis of infective endocarditis rests on clinical, microbiologic, and echocardiographic criteria, as outlined in Table 35-5.^37,39 Infective endocarditis should be suspected in any at-risk patient who develops a febrile illness. Three sets of blood cultures should be obtained between 30 and 60 minutes apart, and an echocardiogram should be performed (see material under Investigations). Blood samples should be incubated for at least 1 week because fastidious bacteria and fungi can occasionally be grown after the standard incubation period of 5 days.

Table 35-5 Modified Duke Criteria for Diagnosing Infective Endocarditis

TEE, transesophageal echocardiography.

Note: A definite diagnosis requires: (1) two major criteria; (2) one major criterion plus three minor criteria; or (3) five minor criteria. A possible diagnosis requires: (1) one major criterion and one minor criterion; (2) three minor criteria.

* See text for details.

** The term persistently positive blood cultures is defined as at least two blood cultures drawn more than 12 hours apart, or three of three or a majority of four or more blood cultures drawn over the span of at least 1 hour.

† An echocardiogram is positive when any of the following are seen: (1) a discrete, echogenic, oscillating intracardiac mass located at the site of endocardial injury; (2) a periannular abscess; (3) a new dehiscence of a prosthetic valve.

From Mylonakis E, Calderwood SB: Infective endocarditis in adults. N Engl J Med 345:1318-1330, 2001; Li JS, Sexton DJ, Mick N, et al: Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis 30:633-638, 2000.

Clinical Features

Infective endocarditis may have an acute or a subacute course. The acute form is commonly caused by S. aureus. Signs of sepsis are common, and valvular destruction and abscess formation (with aortic valve involvement) occur early. More commonly, infective endocarditis has a subacute course with symptoms of fever, malaise, and musculoskeletal pain extending over several weeks. Acute valvular regurgitation is less common with subacute endocarditis.

Signs of cardiac involvement include a new or changing murmur, abnormalities of cardiac conduction, or clinical features of heart failure and valvular regurgitation. First-degree heart block with a progressively lengthening PR interval strongly suggests an aortic root abscess.

Neurologic complications occur in about one third of patients. They may take the form of a generalized encephalopathy—thought to be due to multiple microemboli—or of a focal neurologic deficit due to a macroembolus with or without cerebral abscess formation. Brain CT scanning may reveal asymptomatic cerebral abscesses. Mycotic aneurysms may involve the large arteries.

A variety of peripheral embolic, vascular, and immunologic phenomena may be identified. Petechiae are small purplish spots that may be present in the conjunctiva, the mucous membranes of the mouth, and on the skin—commonly on the fingertips. Splinter hemorrhages are red or brown linear marks under the fingernails or toenails. Osler nodes are small painful pink nodules on the pads of the fingers and toes. Janeway lesions are nontender red-blue macular lesions on the palmar surface of the hands and feet. Roth spots are superficial retinal hemorrhages that may be seen on fundoscopy. Circulating immune complexes can cause polyarthritis or glomerulonephritis. However, renal dysfunction is more likely to be due to heart failure or sepsis than to glomerulonephritis. Splenomegaly is common. Pain localized to a specific region (e.g., liver, bone, soft tissue) may represent an embolic abscess.

Investigations

Normochromic normocytic anemia, leukocytosis, and elevated acute-phase reactants (fibrinogen, C-reactive protein, and the erythrocyte sedimentation rate) are common findings. In addition to blood cultures and an echocardiogram (see earlier discussion under Diagnosis), a chest radiograph, and an electrocardiogram should be obtained in all patients. Transesophageal echocardiography may be required to diagnose small vegetations and to fully delineate aortic root abscesses or fistulae. A contrast CT scan of the brain, abdomen, or chest should be performed if there are clinical signs of systemic embolization. Suspected bony emboli may be confirmed by a radionuclide bone scan.

Treatment

If infective endocarditis is suspected clinically and the presentation is acute, empiric antimicrobial treatment (see subsequent material) should be started as soon as blood cultures have been obtained. If the presentation is subacute, it may be reasonable to await the results of blood cultures before starting antimicrobials. Continued pyrexia despite appropriate antimicrobial therapy suggests an undrained abscess. This should be investigated by a repeat echocardiogram, chest, abdominal and pelvic CT scan and, if indicated, a radionuclide bone scan.

Surgery is indicated in the following circumstances:

Antimicrobial therapy alone is usually successful in infective endocarditis due to Streptococcus species. However, when due to S. aureus, fungi, or gram-negative microorganisms such as Serratia species and P. aeruginosa, surgery is commonly required.

Risk factors for systemic embolization include vegetations greater than 10 mm diameter or highly mobile lesions attached to the anterior leaflet of the mitral valve.³⁸ Given that more than 50% of systemic emboli involve the central nervous system, early surgery should be considered in patients with large or mobile vegetations, particularly if they are caused by S. aureus or a fungus. If medical therapy is planned for high-risk patients (i.e., those with large vegetations or valve dysfunction or if there is an aggressive causative microorganism), frequent echocardiographic assessment (e.g., every 48 to 72 hours) should be performed, and if vegetations are growing or valve function is deteriorating, surgery is indicated.

The timing of surgery for patients with central nervous system emboli is problematic. If surgery is performed shortly after an embolic stroke, there is the potential for an intracerebral bleed during the perioperative period, but if surgery is delayed, there is the potential for further emboli or other complications to develop. In one study, surgery performed within 1 week of a cerebral event was associated with neurologic deterioration in 44% of patients.⁴⁰ This risk fell progressively to 2.3% at 4 weeks. The risk of death fell from 31% to 7% during the same period. Thus, ideally, surgery should be delayed for 4 weeks following a neurologic event. However, in the face of recurrent emboli, persistent sepsis, or acute heart failure, surgery may have to occur earlier.

Repair, rather than replacement, of an incompetent valve is preferred because it reduces the risk of recurrence of endocarditis. Otherwise, when replacement is required, the usual criteria for selection of valve type should be applied (see Chapter 10). An exception to this is prosthetic aortic valve endocarditis, when the use of an aortic homograft, rather than a mechanical or a stent-mounted xenograft, is associated with a reduced incidence of recurrence.⁴¹ The incidence of prosthetic-valve endocarditis is highest in the first 2 months following implantation; it decreases progressively over the first year.⁴²

Prevention

In addition to the general strategies outlined earlier, the incidence of VAP is reduced by early extubation; maintaining an endotracheal cuff pressure above 20 cm H₂O; nursing patients in a semirecumbent, 45-degree, head-up position; kinetic therapy (lateral rotation); and the use of noninvasive ventilation.⁴⁴ Postpyloric feeding, compared with gastric feeding, has not been shown to reduce the risk of aspiration or pneumonia and should be introduced only if there is intolerance of gastric feeding.⁴⁵ Continuous subglottic suctioning using custom-designed endotracheal tubes is also effective but is not widely practiced. Pharmacologic strategies to reduce bacterial colonization of the lower respiratory tract also decrease the incidence of VAP. They include application to the oropharynx of topical antimicrobials or disinfectants (chlorhexidine), selective decontamination of the digestive tract (see earlier discussion), avoidance of unnecessary antimicrobial therapy, and avoidance of antacids.⁴⁴ The indications and drug choices for prophylaxis of stress ulcers are listed in Chapter 34. Sucralfate, compared with ranitidine, is associated with a reduced incidence of VAP; however, this advantage is offset by a slightly increased incidence of stress ulcer bleeding.⁴⁴

Diagnosis

VAP is usually diagnosed clinically and radiographically. Clinical signs include deteriorating gas exchange, purulent secretions, pyrexia, an abnormal leukocyte count including immature forms, and increased levels of C-reactive protein. The cardinal radiographic feature is the presence of a new pulmonary infiltrate. Commonly, only a few features of VAP are present, and it is difficult to make the diagnosis with confidence. This uncertainty has led to the development of clinical scoring systems for VAP, one of which is shown in Table 35-7; a Clinical Pulmonary Infection Score⁴⁶ greater than 6 is considered diagnostic of VAP.⁴⁷

Table 35-7 Clinical Pulmonary Infection Score for Ventilator-Associated Pneumonia

Rights were not granted to include this table in electronic media. Please refer to the printed book.

From Pugin J, Auckenthaler R, Mili N, et al: Diagnosis of ventilator-associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic “blind” bronchoalveolar lavage fluid. Am Rev Respir Dis 143:1121-1129, 1991; Singh N, Falestiny MN, Rogers P, et al:

If VAP is suspected, a sputum specimen should be obtained prior to starting antimicrobial therapy. It may be an aspirate obtained by blind tracheobronchial suctioning or a washing obtained during bronchoscopy. For the diagnosis of hospital-acquired pneumonia in nonventilated patients, a sputum specimen should be obtained from deep coughing. The presence of squamous cells indicates that the sample contains oropharyngeal rather than tracheobronchial secretions, and any positive culture results should be interpreted with caution.

Microorganisms that are commonly isolated in tracheal secretions but rarely cause pneumonia in immunocompetent patients include the viridans group of streptococci, commensal Neisseria species, Corynebacterium species, Candida species, and coagulase-negative staphylococci. Potentially pathogenic microorganisms include S. aureus, S. pneumoniae, enteric gram-negative bacilli, P. aeruginosa, H. influenzae, and Moraxella catarrhalis.

Bacterial colonization of the lower respiratory tract is common in hospitalized patients, so the presence of microorganisms on gram stain or culture of tracheobronchial secretions is in itself not indicative of nosocomial pneumonia and is not an indication for antimicrobial therapy. An exception is quantitative culture of bronchial washings or protected brush specimens obtained during bronchoscopy. Pneumonia may be diagnosed on the basis of the presence of potentially pathogenic microorganisms: more than 10⁴ or 10⁵ colony-forming units/ml in broncoalveolar lavage and more than 10³ colony-forming units/ml in protected brush specimens.⁴⁴ Detection of 5% or more neutrophils or macrophages with intracellular organisms on a Wright-Giemsa stain of a smear of cytocentrifuged bronchoalveolar lavage fluid is also diagnostic of VAP. Culture of blood and, if appropriate, pleural fluid should also be performed.⁴⁴

Microbiology and Antimicrobial Therapy

The most likely causative microorganism is determined by whether a patient has risk factors for colonization by multidrug-resistant organisms (Table 35-8). Of these risk factors, duration of hospital stay is the most important: beyond 5 days, the causative microorganism is more likely to be multidrug-resistant. In this situation, MRSA and multiresistant Pseudomonas aeruginosa are commonly responsible, especially in the setting of a high prevalence of these microorganisms. The likely causative microorganisms and the empiric antimicrobial therapy recommended by the American Thoracic Society⁴⁴ are listed in Table 35-9. Note that in contrast to empyema, VAP is rarely caused by anaerobes. These recommendations advocate an aggressive antimicrobial strategy, which ensures coverage of all likely pathogens. Within institutions, empiric treatment should take into account local patterns of prevalence and resistance. Therapy should be adjusted when culture results become available. The duration of antimicrobial treatment should be 7 to 8 days, assuming the following criteria are met⁴⁴: (1) the patient has received appropriate therapy from the start; (2) there is a good clinical response to treatment; (3) there is no evidence of infection by Pseudomonas species, Acinetobacter species, or Stenotrophomonas species. If these criteria are not met, a 10- to -14 day course is appropriate.⁴⁴

Table 35-8 Risk Factors for Colonization With Multidrug-Resistant (MDR) Microorganisms

Rights were not granted to include this table in electronic media. Please refer to the printed book.

From American Thoracic Society, Infectious Diseases Society of America: Guidelines for the Management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 171:388-416, 2005.

Table 35-9 Pathogens Responsible for Nosocomial Pneumonia in Patients With and Without Risk Factors for Colonization by Multidrug-Resistant Microorganisms

Rights were not granted to include this table in electronic media. Please refer to the printed book.

From American Thoracic Society, Infectious Diseases Society of America: Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 171:388-416, 2005.

Exit Site Infections

Signs of an exit site infection include localized redness, swelling, tenderness, and purulent discharge. As soon as infection is identified, the catheter should be removed, samples obtained for cultures (pus swab of the exit site, catheter tip, and peripheral blood cultures), and empiric antimicrobial therapy considered. In the absence of systemic symptoms or particular risk factors, removal of the catheter may be sufficient. Alternatively, oral treatment with an antistaphylococcal penicillin for 7 days is appropriate. This may be modified by the clinical response to treatment and the results of microbiologic cultures. For patients in whom the consequences of bacteremia are dire, such as those who are immunocompromised or are receiving mechanical cardiac support, empiric treatment should be as for a catheter-related blood stream infection. Similarly, if blood cultures are subsequently positive, treatment as for a catheter-related blood stream infection is indicated.

Catheter-Related Blood Stream Infections

Bacteremia can occur with any intravascular catheter—arterial, peripheral venous, central venous—but it occurs most commonly with central venous catheters.

Clinical Presentation

Patients with sepsis typically have systemic upset with fever, sweating, chills, rigors, and tachycardia. Fever may be absent in patients receiving extracorporeal membrane oxygenation or renal replacement therapy and in patients who are severely immunocompromised. The complete blood count may demonstrate an increased or decreased leukocyte count with immature forms (band neutrophils, metamyelocytes). Inflammatory markers (C-reactive protein and procalcitonin) are typically elevated. In severe sepsis, there are signs of organ dysfunction, such as oliguria, confusion, hypoxemia, hypotension, and lactic acidosis. A persistently elevated lactate level despite appropriate resuscitation measures is an important marker of death in patients with septic shock.^62,63 There may be thrombocytopenia and coagulopathy.

A primary focus of sepsis may be identified; for instance, purulent discharge from the sternal wound, abdominal distension and tenderness, or a new pulmonary infiltrate on the chest radiograph. However, localizing signs are commonly absent in critically ill or immunocompromised patients. Early signs of sepsis are easily missed in sedated patients and when there is another cause for hemodynamic instability such as ventricular dysfunction. Hypotension, in association with hyperdynamic circulation (warm peripheries, bounding pulses), is characteristic of septic shock. However, patients who have not been adequately fluid resuscitated or have severely impaired cardiac function may have cool peripheries and weak pulses. Also, myocardial depression occurs in approximately 50% of patients with severe sepsis or septic shock and is independent of preexisting myocardial dysfunction. Elevated troponin levels in patients with sepsis indicate left ventricular dysfunction and are associated with worse prognoses.⁶⁴ By contrast, there is no consistent relationship between elevated B-type natriuretic peptide and adverse outcome due to sepsis.⁶⁴

Treatment

Evidence-based guidelines for the management of severe sepsis and septic shock have been developed,⁶⁵ and they form the basis of the recommendations that follow. With severe sepsis and septic shock, early institution of antimicrobial therapy, hemodynamic resuscitation, and source control are critical to survival.

Bacteriostatic Versus Bactericidal Activity

Antimicrobials may be bacteriostatic (i.e., they inhibit microorganism growth) or bactericidal (i.e., they kill microorganisms). This distinction is blurred because it is partly dependent on the antimicrobial effect-site concentration, the organism involved, and whether the drug is used in combination with other antimicrobials. For instance, vancomycin is bacteriostatic against many enterococci but bactericidal against other gram-positive microorganisms. For most infections, bacteriostatic activity is satisfactory because it permits normal host defenses to destroy pathogens. However, bactericidal agents offer an advantage when host defenses are impaired (e.g., in immunocompromised patients) or when access to the infection site by host defenses is reduced (e.g., in infective endocarditis). Antimicrobials that are bacteriostatic include the tetracyclines and erythromycin; other agents commonly used in the ICU are, if given in appropriate doses, bactericidal.

Concentration-Dependent Versus Time-Dependent Inhibition

Antimicrobials exhibit either concentration-dependent or time-dependent inhibition of pathogens.⁸⁴ With concentration-dependent inhibition, the higher the plasma concentration of drug, the greater the inhibition of pathogens. Concentration-dependent inhibition is characteristic of the aminoglycosides. Maximum drug concentrations greater than 10 times the MIC are needed for effective therapy. The optimal dosing regimen for antimicrobials that display concentration-dependent inhibition is a large, once-daily dose. Once-daily dosing of aminoglycosides is associated with increased efficacy and reduced toxicity compared with the same total dose given as multiple, divided doses.⁸⁵ With time-dependent inhibition, the goal is to maximize the time during which the plasma concentration is above the MIC. Time-dependent killing is a characteristic of β-lactams and vancomycin. Theoretically, the ideal dosing regime of drugs that display time-dependent killing is a continuous infusion or frequent intermittent boluses; the latter is the usual strategy for β-lactam dosing. Some agents, notably the fluoroquinolones, display both time- and concentration-dependent inhibition.

Combination Treatment

Combination antimicrobial therapy is useful in three situations: (1) empiric therapy against severe infection; (2) treatment of mixed infections; (3) provision of synergistic antimicrobial therapy. An important example of the last circumstance is the treatment of infective endocarditis due to penicillin-sensitive enterococci. The addition of an aminoglycoside provides an enhanced antibacterial effect. The issue of combination therapy for Pseudomonas aeruginosa was discussed previously.

Different antimicrobial classes act by means of different mechanisms. For instance, β-lactams and glycopeptides inhibit bacterial cell wall synthesis; aminoglycosides, erythromycin, and clindamycin (and others) bind to ribosomes and inhibit protein synthesis; and fluoroquinolones affect nucleic acid synthesis. Combination therapy works best when antimicrobials with different mechanisms of action are used. Effective combinations include β-lactams and an aminoglycoside or vancomycin and an aminoglycoside, although renal toxicity is a very real risk with the latter combination.

Pharmacokinetic Considerations

To avoid toxicity, many antimicrobials require dose adjustments in the presence of renal or hepatic dysfunction. Antimicrobials that require dose adjustment in renal dysfunction are listed in Table 35-15. Aminoglycosides and vancomycin are nephrotoxic, but they can be used safely in patients with renal dysfunction as long as trough blood levels are carefully monitored. In patients with renal impairment requiring aminoglycosides, (which display concentration-dependent inhibition) a standard or a slightly reduced dose (e.g., gentamicin 3 to 5 mg/kg) may be given less frequently (e.g., every 36 to 48 hr), based on the trough drug level. if a trough level is high it should be measured every 12 hours until a satisfactory concentration is reached. At which time a further lower dose may be given. With vancomycin, which displays time-dependent inhibition, a smaller dose (e.g., 500 mg instead of the usual 1000 mg) may be given with the same dosing interval (i.e., every 12 to 24 hr), depending on trough blood levels. When renal function is severely impaired, alternatives to these agents (e.g., aztreonam for aminoglycosides, teicoplanin for vancomycin) should be used.

Renally eliminated drugs may also have nonrenal toxicity; for instance, some β-lactams and fluoroquinolones cause central nervous system toxicity, and their dosages must be reduced in patients with severe renal dysfunction.

Dose adjustment is less well defined for hepatic dysfunction. This is because most drug-induced hepatotoxicity is idiosyncratic (see Table 34-6). Antimicrobials that should be administered at reduced dosages or avoided entirely in patients with hepatic dysfunction are listed in Table 34-7. Certain antimicrobials (notably the macrolides, azole antifungals, and rifampicin) function as substrates, inhibitors, and inducers of the CYP3A hepatic enzyme system (see Table 4-3) and therefore can cause important drug interactions.