HOSPITAL-ACQUIRED (NOSOCOMIAL) PNEUMONIA
Bacteriology and Pathogenesis
Hospital-acquired (nosocomial) infections are those occurring after at least 48 hours of hospitalization; they affect 5% to 6% of all hospitalized patients. This rate is at least three times higher for adult patients in ICUs. In most studies, pneumonia is the second or third most common type identified, accounting for 13% to 18% of all cases, occurring at rates of 4 to 7/1,000 hospitalizations. Nosocomial pneumonia develops in up to 25% of patients in an ICU, and the incidence of pneumonia in mechanically ventilated patients may approach 30/100 patients. If adjusted for “ventilator days,” the rates are approximately 15/1,000 ventilator days for medical or surgical patients. Looked at slightly differently, estimated rates of ventilator-associated nosocomial pneumonia are 1% to 3% per day of intubation/ventilation. Mortality from nosocomial pneumonia is 30% to 50%, and pneumonia is the most common fatal nosocomial infection. The problem is especially severe in the elderly and in intubated and ventilated patients. Rates of 17/1,000 patient-days have been noted in this age group, compared with fewer than 2/1,000 patient-days in persons under the age of 50 years.
Approximately 17% of all nosocomial pneumonias occur in the 1% of patients who are ventilated. Patients with endotracheal tubes or tracheostomies are at risk because these devices bypass respiratory tract host defense mechanisms and allow bacteria to be deposited directly into the lower respiratory tract. Additionally, secretions may pool in the subglottic area above the endotracheal cuff and slowly leak into the lower respiratory tract. Best estimates are that approximately 137,000 deaths may be attributable to this condition annually.
Other overlapping high-risk groups for nosocomial pneumonia are patients in intensive care units (ICUs) and persons undergoing thoracic or thoracoabdominal surgery. Reasons include analgesia, sedation, and pain from thoracotomy tubes.
Bacteriology and Pathogenesis
Historically, the most common causes of nosocomial pneumonia were considered to be Enterobacteriaceae (most often Escherichia coli, Enterobacter species, Klebsiella species, and Serratia species), Staphylococcus aureus, and other gram-negative bacilli (e.g., Pseudomonas aeruginosa). Certain species of enteric gram-negative rods often predominate in individual hospitals. Depending on the study and the diagnostic test performed, more than one pathogen may be found in up to 60% of patients. Such observations have important implications for therapy. Data based on transtracheal aspiration techniques and anaerobic bacteriology demonstrate the potentially important role of anaerobes, Haemophilus influenzae, and Streptococcus pneumoniae. In almost 50% of cases, mixed infection was documented and often involved anaerobes. S. pneumoniae and H. influenzae are more likely to be noted early in hospitalization (especially in immunocompetent patients with trauma) and in patients who have not received prior antimicrobials. Legionella species have also been reported as causes of both sporadic and epidemic cases of hospital-acquired pneumonia. Many cases occur in heavy smokers, those with chronic obstructive lung disease, and persons with underlying disorders of cell-mediated immunity. Nosocomial pneumonias associated with viruses, such as respiratory syncytial virus and influenza virus, among others, have been less thoroughly investigated, in part because of difficulties with cultures.
Up to 25% of ICU patients become colonized with resident gram-negative bacilli within 24 hours, and this increases to approximately 50% after 4 to 5 days. Sources of bacteria are often the patient’s own upper respiratory tract flora. Historically, respiratory therapy equipment was incriminated, but this problem is now less important. Exogenous factors associated with nosocomial pneumonia include the hands of hospital personnel and occasionally tap water and other liquids. Contaminated hands were a contributory factor in a recent outbreak of nosocomial legionellosis. Factors associated with colonization include antimicrobial therapy, underlying illnesses, acidosis, coma, and endotracheal intubation. The percentage of colonized patients in whom clinical pneumonia developed was 23%, versus only 3% for those who were not colonized. Other sources of organisms that may predispose to nosocomial pneumonia include the stomach.
The diagnosis of nosocomial pneumonia is often difficult, not least because communication with the patient may be impossible. In the presence of endotracheal tubes or tracheostomy, purulent secretions may be indicative of local irritation rather than true infection, and for similar reasons, polymorphonuclear leukocytes may be present on Gram’s-stained endotracheal material. Cultures of lower respiratory tract secretions obtained through endotracheal tubes or tracheostomies (blind bronchial sampling) may reveal potential pathogens that possibly reflect only colonization or tracheobronchitis (a clinical entity with far less severe implications). Fever and leukocytosis can develop in a severely ill patient for many reasons other than lower respiratory tract infection. Alternatively, the patient with pneumonia may fail to have a fever or elevated WBC count for reasons that include old age, renal failure, or effects of medication. Well-defined infections may be unaccompanied by fever in up to 30% of cases. Similarly, roentgenography of the chest may prove unreliable. Among the numerous other causes for pulmonary infiltrates are infarct, congestive heart failure, and malignancy. Autopsies performed on patients with adult respiratory distress syndrome suspected of having nosocomial pneumonia have failed to corroborate clinical and radiographic findings in up to 30% of cases.
The diagnosis of nosocomial pneumonia in the intubated/ventilated patient is controversial. Recommendations have ranged from empiric antibiotic therapy through initial invasive procedures that include flexible fiberoptic bronchoscopy with protected brush (FFPB) and quantitative bacteriology or bronchoalveolar lavage (BAL) with quantitative bacteriology. Endpoints that differentiate true pathogens from contaminants are more than 103 (FFPB) and more than 104 (BAL) organisms per milliliter. An endotracheal aspirate is a relatively simple specimen to obtain; however, its role in the diagnosis of nosocomial pneumonia has been controversial. Although some data demonstrate a failure to identify pathogens in up to 40% of cases, others consider the absence of gram-negative bacilli or gram-positive cocci, which is consistent with S. aureus, to be reliable information in helping to determine the bacteriology of ventilator-associated pneumonia. Additionally, two recent studies have employed quantitative bacteriology of endotracheal aspirates, with more than 105 to 106 organisms per milliliter as a cutoff. Corroboration with “gold standards” was approximately 70%. A recent prospective investigation of ventilator-associated pneumonia utilized FFPB, BAL, and quantitative endotracheal aspirates, with mortality as the endpoint. There was no mortality advantage when data from FFPB or BAL were utilized in comparison with data available from quantitative endotracheal aspirates. Furthermore, antibiotics were changed more often in patients in whom FFPB and BAL were carried out. It is the opinion of the author that endotracheal aspiration with quantification is evolving into the most cost effective method of diagnosing nosocomial pneumonia in ventilated patients and should be performed before changes in antibiotics are made.
Blood cultures should be obtained from all patients, although positive results may be noted in only about 5% of cases. In general, a diagnosis of nosocomial pneumonia should be strongly considered when the following are present: (a) new or worsening pulmonary infiltrates of uncertain origin; (b) purulent pulmonary secretions in association with polymorphonuclear leukocytes and organisms seen with Gram’s stain; (c) an appropriate clinical course, usually associated with respiratory decompensation; and (d) fever and leukocytosis. In practice, it is often necessary to treat critically ill patients as if they had pneumonia even if not all the criteria are fulfilled. Endotracheal aspiration with quantitative bacteriology may evolve into the simplest and most valuable study to identify likely pathogens.
Respiratory support, treatment of ancillary conditions (congestive heart failure, adult respiratory distress syndrome), and antimicrobial therapy are the cornerstones of management. Antimicrobials should be administered in full therapeutic doses, generally intravenously, with consideration of any organ failure. Recent recommendations from the American Thoracic Society suggest that antibiotics be chosen based on risks for “severe” pneumonia, underlying host conditions, and time of onset of nosocomial pneumonia. Patients at risk for severe nosocomial pneumonia include those who require care in a critical care unit, show rapid radiologic progression of pneumonia, have acute renal failure requiring dialysis, or are in shock or respiratory failure. Early pneumonia is that which occurs within the first 5 days of hospitalization. Underlying conditions that need to be assessed include risk for aspiration, underlying structural lung disease, and risks for legionellosis. Table 59-1 provides antibiotic recommendations based on core pathogens. Table 59-2 expands antibiotic recommendations when mitigating factors are present.
Table 59-1. Management of patients with mild/moderate nosocomial pneumonia, no unusual risk factors, onset anytime, or patients with severe nosocomial pneumonia and early onset
Table 59-2. Management of patients with mild/moderate nosocomial pneumonia and risk factors
In most conditions, a single effective antibiotic suffices for treatment. Two agents may be indicated for pneumonia caused by P. aeruginosa or Enterobacter species (when treatment includes selected cephalosporins), and for empiric therapy when precise organism identification is not available. When legionellosis is considered, either a macrolide or quinolone should be included in the antibiotic regime. Third-generation cephalosporins, imipenem-cilastatin, ticarcillin-clavulanate, and parenteral quinolones are probably all effective against susceptible gram-negative pathogens. Aminoglycosides remain valuable agents for many enteric gram-negative bacilli and P. aeruginosa. Regimens of single daily doses of 5 to 7 mg/kg are generally more efficacious and may be safer than the more classic regimens in which doses are administered thrice daily. For considerations of P. aeruginosa (e.g., severe nosocomial pneumonia) two effective agents are advised. Clinicians may employ double b-lactam therapy or regimens that include a quinolone. Gram-positive nosocomial pneumonias can be treated with a single appropriate antimicrobial.
The optimal length of therapy for nosocomial pneumonia is unknown. At least 10 days of effective therapy is generally employed. Patients with lung abscess, empyema, and other suppurative complications will require longer courses. Those with pneumonia caused by Legionella pneumophila are generally treated for up to 21 days. Use of qui-nolones may allow selected patients to complete prolonged courses with oral medications. Staphylococcal pneumonia should be treated parenterally for at least 2 weeks. Patients who are bacteremic with S. aureus often require at least 4 weeks of parenteral therapy.
Endotracheal administration of aminoglycosides (generally gentamicin or tobramycin) is valuable in selected patients. It should be considered for persons with gram-negative bacillary or P. aeruginosa pneumonia who are receiving parenteral gentamicin or tobramycin. This adjunctive mode of administration provides excellent levels of drug within bronchial secretions and is generally well tolerated. Dosage is typically 40 mg of gentamicin or tobramycin administered endotracheally every 8 hours. Patients are positioned to allow distribution to the area of pneumonia. A recent blinded, placebo-controlled trial demonstrated enhanced microbiologic eradication of the offending pathogen(s). Other studies suggest improved clinical outcomes as well.
Endotracheal tubes and other equipment that bypass normal respiratory defenses should be used judiciously. Postoperative patients should be given cough and deep-breathing exercises, and it is sensible to position patients so as to decrease vomiting. Somnolent patients should be turned regularly. Endotracheal or tracheostomy tubes should be suctioned as necessary. Routine treatment of colonizing bacteria in the absence of clinical or radiographic lower respiratory tract infection is not indicated. Such therapy enhances the emergence of resistant organisms and the potential for clinical infection. Adherence to infection control procedures, including hand washing and appropriate equipment management, is mandatory.
An association between upper gastrointestinal tract colonization and nosocomial pneumonia is now established. Organisms located within the stomach may be aspirated. Colonization with gram-negative bacilli is enhanced by elevations in gastric pH, as may be noted with the use of medications (antacids and histamine2 antagonists) that prevent stress bleeding. Sucralfate (which does not alter gastric pH) appears more effective in preventing colonization and respiratory infection than antacids and histamine2 antagonists. Reasons include a preserved lower pH and a direct antibacterial action of the drug. Other measures that attempt to decrease aspiration of upper gastrointestinal tract contents include the use of jejunostomy tubes for long-term feeding and acidification of enteral feedings.
Selective digestive tract decontamination has been studied, with conflicting results. Intraoral plus nasogastric mixtures of amphotericin B, polymyxin E, and tobramycin (or variants) have been employed in selected patients at high risk for nosocomial pneumonia. These products are administered several times daily throughout the ICU stay. IV cefotaxime has also been used during the first 3 days. Although many data demonstrate decreased oropharyngeal and upper gastrointestinal tract colonization as well as decreased rates of nosocomial pneumonia, mortality rates do not generally change, and this procedure is not currently considered a standard of care.
Occasional outbreaks of nosocomial pneumonia caused by a specific pathogen may occur in hospital areas such as an ICU. Prophylactic administration of aerosolized polymyxin B has been shown to be effective in decreasing both transmission and infection with P. aeruginosa during respiratory outbreaks with this pathogen. All at-risk persons entering the ICU should receive the aerosol for several weeks after the last isolate. However, prolonged prophylaxis results in higher mortality rates from pneumonia caused by pathogens resistant to polymyxin. Thus, the routine use of this aerosol as a preventive measure is not encouraged.
It is unlikely that the availability of more potent antimicrobials will result in a decline in the morbidity or mortality of nosocomial pneumonia. Improving host defenses, adherence to infection control policies and procedures, and preventing colonization with potential pathogens, however, may prove to be beneficial. (R.B.B.)
American Thoracic Society. Hospital-acquired pneumonia in adults: diagnosis, assessment of severity, initial antimicrobial therapy, and preventative strategies. Am J Respir Crit Care Med 1995;153:1711–1725.
Conclusions of a consensus panel. Provides recommendations for the management of nosocomial pneumonia, including antibiotic decision making based on severity of disease, time of onset of nosocomial pneumonia, and risk factors for specific pathogens. Some of the recommendations are somewhat outdated, as newer antibiotics are now available, such as the newer fluorinated quinolones, meropenem, and cefepime.
Bartlett JG, et al. Bacteriology of hospital-acquired pneumonia. Arch Intern Med 1986; 146:868–871.
One of the few studies of the etiology of nosocomial pneumonia that employed transtracheal aspiration and anaerobic bacteriology. The subjects were a Veterans Administration population having no association with endotracheal intubation and ventilation. A diverse group of organisms was identified, including many anaerobes, common bacteria (S. pneumoniae and H. influenzae), and enteric gram-negative bacilli. At least two organisms were identified in almost 50% of patients.
Brown RB, et al. Double-blind study of endotracheal tobramycin in the treatment of gram-negative bacterial pneumonia. Antimicrob Agents Chemother 1990;34:269–272.
This blinded, prospective investigation assessed the role of endotracheal tobramycin in addition to standardized parenteral antimicrobial therapy for patients with gram-negative bacillary pneumonia. Microbiologic cure was enhanced in patients who received endotracheal antimicrobials, but clinical efficacy was no different. No significant adverse reactions occurred.
Chastre J, Fagon JY, Trouillet JL. Diagnosis and treatment of nosocomial pneumonia in patients in intensive care units. Clin Infect Dis 1995;21(Suppl 3):S226–S237.
An excellent summary article that deals with the problems of diagnosing nosocomial pneumonia in this patient population. The authors prefer FFPB with quantitative bacteriology, attempt to perform the procedure before initiating new antibiotic treatments, and withhold antibiotics if results are negative.
Craven DE, Steger KA. Epidemiology of nosocomial pneumonia. New perspective of an old disease. Chest 1995;108 (Suppl):1S–16S.
The authors present an excellent overview of nosocomial pneumonia, with special reference to bacteriology and risk factors for specific pathogens.
Driks MR, et al. Nosocomial pneumonia in intubated patients given sucralfate as compared with antacids or histamine type 2 blockers. N Engl J Med 1987;317:1376–1382.
An excellent study comparing two methods commonly employed to prevent stress ulceration in ICU patients. It demonstrates that sucralfate, not primarily associated with elevating gastric pH, was less likely to result in gastric colonization with enteric gram-negative bacilli and was associated with a lower incidence of nosocomial pneumonia.
Fish DN, et al. An update on nosocomial pneumonia: treatment, prevention, and future directions. Infect Med 1992;9:27–28,33,37–42,44–46.
This review stresses antimicrobial therapy of nosocomial pneumonia based on likely pathogens. Provides good pharmacokinetic and pharmacodynamic data and gives doses of often-employed antimicrobials and regimens.
Sanchez-Nieto JM, et al. Impact of invasive and noninvasive quantitative culture sampling on outcome of ventilator-associated pneumonia. Am J Respir Crit Care Med 1998;157:371–376.
The authors conducted an open, prospective, randomized study in 51 ventilated patients to determine if they could identify benefits of invasive diagnostic procedures in the management of ventilator-associated pneumonia. In one group, decision making was based on results of either FFPB or BAL (both with quantitative bacteriology), whereas in the second group it was based on endotracheal aspiration with quantitative bacteriology. There was no difference in mortality with the invasive procedures, but there were more changes in antibiotic treatment. The authors conclude that they could not identify benefits of the more invasive procedures in this pilot study.
Tablan OC, et al., and the Hospital Infection Control Practices Advisory Committee. Guidelines for prevention of nosocomial pneumonia. Infect Control Hosp Epidemiol 1994;15:587–627.
This Centers for Disease Control and Prevention document represents an excellent overview of the problem of nosocomial pneumonia and the methods available for prevention. Recommendations are stratified on the basis of strength of available data.