Published May 29, 2002Justin
L.
|
In 1995, the American Thoracic Society published a consensus statement defining hospital-acquired pneumonia (HAP) as a pneumonia that is not incubating at the time of hospital admission and begins more than 48 hours after admission. HAP occurs relatively frequently and is associated with a high rate of mortality;1 therefore, it is important to prevent, promptly diagnose, and effectively treat this infection. This chapter provides basic information for the clinician and reviews guidelines for the management of HAP. |
|
EpidemiologyPathogenesisDiagnosisTreatmentPreventionReferences
|
||||||||||||||||||||||||||||||||||
| EPIDEMIOLOGY | |||||||||||||||||||||||||||||||||||||
There are 300,000 cases of HAP annually, and it carries an associated mortality of 30% to 70%.2 It is difficult to determine the fraction of patients with HAP whose mortality is directly attributable to their pneumonia (the attributable mortality), but this rate is estimated to be between 27% and 50%.2,3 This means that of the patients who develop HAP, it will be the proximate cause of death in one quarter to one half of them, while the remaining one half to three quarters will develop HAP but die from some other cause. HAP lengthens the hospital stay by 7 to 9 days and is associated with a higher cost of medical care.2 General risk factors for developing HAP include age more than 70 years, serious comorbidities, malnutrition, impaired consciousness, prolonged hospitalization, and chronic obstructive pulmonary diseases.2 HAP is the most common infection occurring in patients requiring care in an intensive care unit (ICU), with incidence rates ranging from 6% up to 52%,3 much higher than the 0.5% to 2% incidence reported for hospitalized patients as a whole.4 This increased incidence is due to the fact that patients located in an ICU often require mechanical ventilation, and mechanically ventilated patients are 6 to 21 times more likely to develop HAP than are nonventilated patients.3 Mechanical ventilation is associated with high rates of HAP because the endotracheal tube bypasses upper respiratory tract defenses, allows for pooling of oropharyngeal secretions, prevents effective cough, and can be a nidus for infection.2 The development of HAP in mechanically ventilated patients portends a poor prognosis, with a rate of mortality 2 to 10 times greater for this group than for mechanically ventilated patients without HAP.3 The incidence of HAP is even greater in the subset of mechanically ventilated ICU patients who have acute respiratory distress syndrome (ARDS), with reported incidence rates in this group of 34% to 60%.4 One study of 878 mechanically ventilated patients showed that the incidence of HAP rose from 11% in patients without ARDS to 36% in patients with ARDS.5 In contrast to other studies in which patients with ARDS who developed HAP had mortality rates of more than 50%,4 this study showed that HAP prolonged the time that patients required mechanical ventilation, but did not seem to have an effect on overall survival.5 |
|||||||||||||||||||||||||||||||||||||
| PATHOGENESIS | |||||||||||||||||||||||||||||||||||||
HAP is likely to occur when a sufficiently large number of organisms are delivered to the lower respiratory tract such that host defenses are overwhelmed (eg, by aspiration or contaminated respiratory therapy equipment), when host defenses are impaired (eg, by immunodeficiency or steroids), or if particularly virulent organisms are involved.2 Gram-negative bacteria (GNB) account for 55% to 85% of HAP infections, and gram-positive cocci account for 20% to 30%.4 GNB rarely causes community-acquired pneumonia, but hospitalized patients are exposed to organisms that nonhospitalized patients are not, and many factors directly related to hospitalization make patients more likely to develop, and less able to fight, infection. Microaspiration of contaminated oropharyngeal secretions seems to be the most important of these factors, as it is the most common cause of HAP.2 Microaspiration is a common event even in nonhospitalized patients, happening in as many as 50% of normal sleeping subjects and 70% of patients subject to some degree of sedation or depressed consciousness. Microaspiration in a hospitalized patient is more serious and more likely to cause infection because the oropharyngeal secretions that are aspirated are likely to contain organisms that are not present under normal circumstances but which frequently cause HAP. The oropharynx of hospitalized patients becomes colonized by GNB in as many as 35% of moderately ill and 73% of critically ill patients, often within the first 4 days of admission.2 With the introduction of these new pathogenic organisms into the oropharynx, the previously benign event of microaspiration now becomes a mechanism by which virulent organisms are introduced into the lower respiratory tract and cause pneumonia.2 |
|||||||||||||||||||||||||||||||||||||
| DIAGNOSIS | |||||||||||||||||||||||||||||||||||||
Diagnosing HAP is difficult because there is no method for obtaining a diagnosis that is reliable in all cases.6 The diagnosis is initially made on clinical grounds by the finding of a new infiltrate on chest radiograph, fever, purulent sputum, or other signs of clinical deterioration. Unfortunately, this clinical method was shown to be specific for HAP in only 27 of 84 patients in a series reported by Fagon et al 7 because many other conditions such as congestive heart failure, pulmonary embolism, atelectasis, ARDS, pulmonary hemorrhage, or drug reactions may mimic pneumonia, particularly in critically ill patients.1 Lack of specificity in the clinical diagnosis gives rise to the need for more reliable diagnostic tools so that fewer patients will be treated with antibiotics for noninfectious causes. While there are many different testing modalities that may be employed to this end, all have their limitations and none is sufficiently sensitive and specific to be considered a "gold standard" test.3 Blood cultures have diagnostic and prognostic value but their reported sensitivity is only 8% to 20%, and their role is therefore limited. Likewise, examination of expectorated sputum is neither sensitive nor specific and should not be routinely used.1 The most useful noninvasive test is the examination of tracheobronchial aspirates (TBA). This method has a high degree of sensitivity, as demonstrated in a recent study where the offending organism was recovered from tracheal secretions in 29 of 31 patients.4 The weakness of this test is its inability to differentiate between the organism responsible for causing the pneumonia and harmless colonizers. Because of this limitation, the use of TBA lies in its negative predictive value, its ability to exclude the presence of resistant organisms, and thus to narrow antibiotic coverage.1 Invasive bronchoscopic techniques are able to take samples directly from the lower respiratory tract without contamination from upper airway or oral secretions and would seem to provide an advance in identifying the responsible pathogen.8 Surprisingly, when bronchoscopic techniques such as bronchoalveolar lavage (BAL) or the use of protected specimen brushes (PSB) have been compared with less invasive methods, they do not appear to differ significantly in terms of sensitivity, specificity or, more importantly, patient morbidity and mortality.9 There is currently a lack of consensus on the role of invasive diagnostic testing for HAP, and it is the subject of ongoing debate. One study regarding this issue compared the results of bronchoscopically obtained PSB with those of TBA in 76 mechanically ventilated patients who were already receiving empiric antibiotic therapy.9 In this study, more patients who received bronchoscopy with PSB had a change in their antibiotic regimen, but there were no significant differences in length of stay, days requiring mechanical ventilation, or mortality between the two groups. This study concluded that outcome was "not influenced by techniques used for microbial investigation." 9 Another study compared the use of invasive testing, such as PSB and BAL, with the use of noninvasive TBA among 413 patients. This study showed an initial decrease in mortality, antibiotic use, and organ dysfunction at 14 days among patients in whom invasive techniques were used, but at a 28-day analysis the difference in mortality could not be similarly demonstrated.10 These and other studies have led many to the conclusion that noninvasive and invasive tools achieve similar diagnostic performance and therefore the use of invasive techniques cannot be justified in every patient with HAP.8 Others argue that if invasive testing is done within the first 12 hours after diagnosis and before antibiotics are administered, the improvement in diagnostic yield may be sufficient to merit its use.3 A recent review of this issue by Ewig and Torres8 stated that invasive and noninvasive techniques do not differ significantly, that both are less sensitive than specific, and that the false-negative rate for these tests ranges from 30% to 40% and the false-positive rate from 20% to 30%. The review also stated that invasive diagnostic testing should not be performed early in the course of HAP, and the best way to make adjustments to the empiric antibiotic regimen is by TBA rather than invasive techniques. Further, they stated that due to poor sensitivity associated with invasive methods, empiric coverage should not be stopped on the basis of negative diagnostic testing alone, and that the potential role for invasive diagnostic evaluation lies in cases of nonresponse to initial treatment.8 |
|||||||||||||||||||||||||||||||||||||
| TREATMENT | |||||||||||||||||||||||||||||||||||||
After HAP is diagnosed, it is imperative that antimicrobial therapy begin promptly because delays in administration of antibiotics have been associated with worse outcomes.3 One study in support of this notion reported a mortality of 30% in patients who received early, appropriate therapy compared with a rate of 91% among patients who did not.11 The initial selection of an antimicrobial agent is almost always made on an empiric basis and is based on factors such as severity of infection, patient-specific risk factors, and total number of days in hospital before onset.3 All empiric treatment regimens should include coverage for a group of core organisms that includes aerobic gram negative bacilli (Enterobacter spp, Escherichia coli, Klebsiella spp, Proteus spp, Serratia marcescens, and Hemophilus influenzae) and gram-positive organisms such as Streptococcus pneumoniae and Staphylococcus aureus. In patients with mild or moderate infections and no specific risk factors for resistant or unusual pathogens, monotherapy with a second-generation cephalosporin such as cefuroxime; a nonpseudomonal third-generation cephalosporin such as ceftriaxone; or a beta-lactam/beta-lactamase inhibitor such as ampicillin/sulbactam, ticarcillin/clavulanate, or piperacillin/tazobactam may be appropriate. For patients in this low-risk category who have an allergy to penicillin, it is appropriate to initially use a fluoroquinolone or clindamycin and aztreonam.1,2 While it is acceptable to use a fluoroquinolone in the empiric regimen of patients with penicillin allergies, one recent paper studied the use of penicillin skin testing in these patients and found that most patients with a history of penicillin allergy can safely be treated with penicillin antibiotics, so penicillin skin testing may be a means by which the use of fluoroquinolones can be decreased.12 Patients with mild or moderate infections with specific risk factors should have broadened empiric coverage. In patients with witnessed aspiration or in those who have had recent thoracoabdomial surgery, it may be advisable to cover anaerobes by adding clindamycin, although use of a beta-lactam alone may be sufficient. Patients with coma, head trauma, recent infections with influenza virus, diabetes, or chronic renal failure, or who are injection drug users, are at risk for infections by Staphylococcus aureus and may require the addition of vancomycin to cover methicillin-resistant strains until sensitivities are known. Patients who have received long courses of steroids should have a macrolide as a part of their initial therapy, given their increased risk for Legionella spp.2 Regardless of the severity of the infection, patients who have received antibiotics previous to developing pneumonia, patients with structural lung disease, patients who have received steroids, and patients with a prolonged ICU course (more than 5 days) should receive a combination of antibiotics to cover not only core pathogens but also infection from Acinetobacter spp or Pseudomonas aeruginosa. Combination therapy should be employed in these cases because of the high rate of acquired resistance among these organisms. Appropriate combinations for this group of patients include an aminoglycoside or ciprofloxacin in addition to a beta-lactam with antipseudomonal coverage. Additionally, vancomycin should be considered if the patient has risk factors that suggest methicillin-resistant Staphylococcus aureus could be a pathogen.1,2 A pneumonia is defined as severe if there is need for admission to an ICU, radiographic evidence of rapid progression, need for mechanical ventilation or high levels of inspired oxygen, or evidence of sepsis. Patients with severe pneumonias who have been hospitalized for less than 5 days but who have no specific risk factors should receive empiric coverage aimed at the core group of organisms only; however, monotherapy is probably not appropriate in these patients and combination therapy should be used. Any patient with a severe pneumonia and any risk factor, including a hospitalization of more than 5 days, should again receive a combination of antibiotics that cover infections by Acinetobacter spp and P aeruginosa.1,2 There is no consensus regarding the duration of antibiotic treatment for all patients with HAP, although if the initial clinical suspicion was low, antibiotics may be safely discontinued after 72 hours if the clinical picture has not changed significantly.13 Recommendations from the American Thoracic Society suggested that duration of treatment should be guided by severity, time to clinical response, and the pathogenic organism1; however, a recent panel of experts suggested that "the main factor for deciding the duration of therapy should be the time to clinical response and not the pathogen involved" and that patients should be treated for at least 72 hours after a clinical response is achieved.3 Clinical response to antimicrobial therapy is not likely in the first 48 to 72 hours, so the empiric antibiotic regiment should not be changed during this time unless as directed by the results of microbiologic investigation.1 In patients who fail to respond after this initial period, recommendations are that antibiotic coverage should be broadened, noninfectious causes considered, and invasive diagnostic testing performed. Appropriate diagnostic testing may include bronchoscopy with PSB and BAL, radiographic tests to evaluate for the possibility of pleural effusions or abscesses that limit response, or CTs of the sinuses to evaluate for sinusitis, as this may be a cause of persistent symptoms. After excluding all other etiologies in the nonresponding patient, it may be advisable to perform open-lung biopsy for diagnostic purposes, even though this technique has not been shown to improve outcomes.2 |
|||||||||||||||||||||||||||||||||||||
| PREVENTION | |||||||||||||||||||||||||||||||||||||
Prevention is important because even when appropriate diagnostic and therapeutic procedures are used, HAP is associated with a high rate of mortality. Prevention strategies should focus on general measures for infection control, measures directed at specific patient risk factors, and measures to limit the use of antibiotics in an attempt to decrease the prevalence of resistant organisms. Although most clinicians think that strategies for prevention are important, it must be noted that very few techniques have been shown in well-designed experimental or epidemiologic studies to be of definite value, so most of the specific recommendations for prevention of HAP are made on the basis of expert opinion rather than hard data. In fact, the Centers for Disease Control and Prevention recently published a set of 74 recommendations for preventing bacterial nosocomial pneumonia, and only 15 of these recommendations were "strongly supported by well-designed experimental or epidemiologic studies"14 and 14 of those 15 recommendations dealt with general issues such as surveillance, education, hand washing, sterilization, proper use of gloves, value of vaccination, and sanitation. The recommendation that prophylactic antibiotics not be routinely administered was the only specific recommendation supported by well-designed studies; all other specific recommendations, such as prevention of aspiration and prevention of colonization, were based on less stringent evidence or on expert opinions.14 General measures for infection control such as proper hand washing, use of gloves, and measures to reduce contamination from respiratory therapy equipment are important factors in preventing HAP. Although the benefits of these practices are universally recognized, they are often not performed, as evidenced by an observational study performed in an ICU setting where only 10% of health care workers washed their hands before having direct patient contact and only 32% washed their hands after patient contact.15 Compliance with these general measures has repeatedly been shown to correlate with favorable outcomes, and stricter adherence to these guidelines is needed. Patient-specific measures should be considered and, if other circumstances allow, followed in every hospitalized patient. Efforts should be made to reduce immunosuppression, reduce sedation, avoid transportation of patients out of the ICU, and provide adequate nutrition, as these measures have been shown to reduce the incidence of HAP. It has been shown that pneumonia is more common in patients with sinusitis.2 Because the presence of nasally placed tubes is associated with sinusitis, it is recommended that endotracheal or gastric tubes be placed orally if possible.2 Another intervention with proven benefit in mechanically ventilated patients is the use of subglottic secretion drainage, a method by which opropharyngeal secretions are continuously suctioned in an effort to prevent pooling and thus aspiration.1 Finally, body positioning seems to have an impact on the development of HAP, as evidenced by a study that was stopped early after an interim analysis found a significant difference in the development of HAP between patients who were in a supine position and patients in a semirecumbent position, probably due to the protective effects against aspiration of refluxed gastric contents by those in the semirecumbent position.16 Other techniques, such as administration of prophylactic antibiotics or nonabsorbable antibiotics for the purpose of gastrointestinal decontamination, have been attempted in the past with varying degrees of reported success. The recent literature suggests that these methods should not be used because they have not consistently been shown to influence mortality or length of stay, and the use of antibiotics in this manner may lead to the development of resistant organisms.3 On the other hand, the results regarding the role of histamine blockers in the development of HAP are conflicting, but their use should not be specifically avoided in patients for this reason.1 The notion that antibiotics select out resistant and/or virulent organisms was suggested in a study which found that in mechanically ventilated patients who had previously received antibiotics, 65% of pneumonias were due to Acinetobacter or Pseudomonas, but in antibiotic-naïve patients only 19% of infections were due to these pathogens.17 In another recent study, 11 of 14 patients who had previously received ceftazadime developed HAP caused by resistant strains of Acinetobacter but only 11 of 29 control patients developed similar infections.15 These observations have resulted in efforts to decrease the use of antibiotics, which in turn decreases the incidence of antibiotic-resistant pathogens, and potentially improves patient outcomes.18,19 |
|||||||||||||||||||||||||||||||||||||
| REFERENCES | |||||||||||||||||||||||||||||||||||||
|
