Classification

Pneumonia can be classified based on the clinical setting in which it occurs (Table 22-1). This classification is useful because it predicts the likely microbial causes and determines empiric antimicrobial chemotherapy while a definitive microbiologic diagnosis is awaited. (The term empiric therapy refers to treatment that is initiated based on the most likely cause of infection when the specific causative organism is still unknown.)

Community-acquired pneumonia can be divided into two types—acute and chronic—based on its clinical presentation. Acute pneumonia generally appears as an illness of relatively sudden onset over a few hours to several days. The clinical presentation may be typical or atypical, depending on the pathogen. The onset of chronic pneumonia is more insidious than acute pneumonia, often with gradually escalating symptoms over days, weeks, or months.

Pneumonia acquired in health care settings is often caused by different microorganisms than community-acquired pneumonia. Previously termed nosocomial pneumonia, this clinical entity has been further classified as health care–associated pneumonia (HCAP), hospital-acquired pneumonia (HAP), and ventilator-associated pneumonia (VAP).³ HCAP is defined as pneumonia occurring in any patient hospitalized for 2 or more days in the past 90 days in an acute care setting or who in the past 30 days has resided in a long-term care or nursing facility; attended a hospital or hemodialysis clinic; or received intravenous antibiotics, chemotherapy, or wound care. HAP is defined as lower respiratory tract infection that develops in hospitalized patients more than 48 hours after admission and excludes community-acquired infections that are incubating at the time of admission. VAP is defined as lower respiratory tract infection that develops more than 48 to 72 hours after endotracheal intubation.

HAP is a common clinical problem and represents the second most common nosocomial infection in the United States, accounting for 15% to 18% of all such infections.4,5 Current estimates suggest that more than 250,000 individuals develop this complication each year. HAP increases hospital length of stay 7 to 9 days at an average incremental per-patient cost of $40,000. In selected patient populations, such as patients in the intensive care unit (ICU) and bone marrow transplant recipients, the crude mortality rate from HAP may approach 30% to 70%, with attributable mortality of 33% to 50%. Certain microorganisms, such as Pseudomonas aeruginosa and Acinetobacter species, are associated with higher rates of mortality.⁶

Pathogenesis

Six pathogenetic mechanisms may contribute to the development of pneumonia (Table 22-2). To minimize nosocomial spread, knowledge of these mechanisms is important to the understanding of the various disease processes and to the formulation of effective infection control strategies within the hospital. The fact that tuberculosis is acquired by inhalation of infectious particles is the basis for a policy whereby patients with suspected or proven tuberculosis who are coughing are placed in respiratory isolation, minimizing the risk of disease transmission within the hospital setting.

Aspiration of oropharyngeal secretions is the second mechanism that may contribute to the development of lower respiratory tract infection. Healthy individuals may aspirate periodically, especially at night during sleep, and a small volume of oropharyngeal secretions, which are colonized with potential pathogens such as Streptococcus pneumoniae and Haemophilus influenzae, may contribute to the development of community-acquired pneumonia. Certain patient populations are at risk of large volume aspiration, such as patients with impaired gag reflexes from narcotic use, alcohol intoxication, or prior stroke. Aspiration also may occur as a result of seizure disorder, cardiac arrest, or syncope.

Aspiration seems to be the major mechanism responsible for the development of some types of mixed aerobic and anaerobic, gram-negative, and staphylococcal HAP. In intubated patients, chronic aspiration of colonized secretions through a tracheal cuff has been linked to the subsequent occurrence of pneumonia,⁴ which has led to the development of novel strategies to prevent HAP, such as continuous suctioning of subglottic secretions in mechanically ventilated patients and elevation of the head of the bed.^7,8

Direct inoculation of microorganisms into the lower airway is a less common cause of lower respiratory tract infection that may contribute to the development of nosocomial pneumonia in mechanically ventilated patients who undergo frequent suctioning of lower airway secretions. In this instance, passage of a suction catheter through the oropharynx may result in inoculation of colonizing organisms into the trachea.

Contiguous spread of microorganisms to the lungs or pleural space from adjacent areas of infection, such as subdiaphragmatic or liver abscesses, is an infrequent cause of pneumonia. This pathogenetic mechanism may occur in patients with pyogenic or amebic liver abscesses involving the dome of the liver in whom rupture of the abscess through the diaphragm leads to the development of pulmonary infection or empyema.

The spread of infection through the bloodstream from a remote site is called hematogenous dissemination. Hematogenous dissemination is an uncommon cause of pneumonia, which may occur in patients with right-sided bacterial endocarditis in whom fragments of an infected heart valve break off and produce either pneumonia or septic pulmonary infarcts after embolization through the pulmonary arteries to the lungs. Certain parasitic pneumonias, including strongyloidiasis, ascariasis, and hookworm, arise through hematogenous dissemination. In such cases, migrating parasite larvae travel to the lungs through the bloodstream from remote sites of infection, such as the skin or the gastrointestinal tract.

Pneumonia also may develop when a latent infection, acquired earlier in life, is reactivated. This reactivation may occur for no apparent reason, as in the case of reactivation pulmonary tuberculosis, but most often it is attributable to the development of cellular immunodeficiency. Pneumocystis jiroveci (previously called Pneumocystis carinii) pneumonia is a prime example of lower respiratory tract infection arising as a result of this mechanism. In developed countries, most healthy individuals have acquired P. jiroveci by age 3 years and show serologic evidence of prior infection. The organism remains dormant in the lung but may reactivate later in life and produce pneumonia in individuals with compromised cell-mediated immunity, such as patients with human immunodeficiency virus (HIV) infection or recipients of long-term immunosuppressive therapy. Cytomegalovirus pneumonia is another example of a latent infection that can reactivate during chronic immunosuppression, especially in solid organ and bone marrow transplant recipients. Immunosuppressive drugs used to modify inflammatory diseases, such as tumor necrosis factor inhibitors, have been associated with pulmonary and extrapulmonary tuberculosis.⁹

Microbiology

The microbiology of community-acquired and nosocomial pneumonia has been studied extensively. Knowledge of which organisms are most commonly associated with pneumonia in different settings is essential because the microbial differential diagnosis guides the diagnostic evaluation and the selection of empiric antimicrobial therapy.

In most studies, S. pneumoniae, also called pneumococcus, has been the most commonly identified cause of community-acquired pneumonia, accounting for 20% to 75% of cases (Table 22-3). Various other organisms have been implicated with varying frequencies. H. influenzae, Staphylococcus aureus, and gram-negative bacilli each account for 3% to 10% of isolates in many reports.¹⁰ Legionella species, Chlamydophila pneumoniae, and Mycoplasma pneumoniae together account for 10% to 20% of cases. These latter organisms, called atypical pathogens, vary in frequency in more recent reports, depending on the age of the patient population, the season of the year, and geographic locale. Legionellosis and C. pneumoniae, in particular, seem to exhibit significant geographic variation in incidence.

Many studies examining the epidemiology and microbiology of community-acquired pneumonia are potentially biased because they focus on patients requiring hospitalization. In patients with less severe illnesses not requiring hospitalization, more recent studies suggest that organisms such as M. pneumoniae and C. pneumoniae account for 38% of cases and may be more common than typical bacterial pathogens such as pneumococcus and H. influenzae.¹¹ In patients who are ill enough to require admission to the ICU, Legionella species, gram-negative bacilli, and pneumococcus are disproportionately more common.¹ A virulent strain of methicillin-resistant S. aureus (MRSA) has emerged as a cause of severe necrotizing community-acquired pneumonia.¹²

In urban settings that have a high incidence of endemic HIV infection, P. jiroveci may be a more common cause of community-acquired pneumonia and, according to one report, may account for 13% of cases.¹³ Viruses such as influenza, respiratory syncytial virus, and adenovirus are occasional causes of community-acquired pneumonia, especially in patients with milder illnesses not requiring hospitalization and encountered in the late fall and winter months. A worldwide pandemic of H1N1 influenza during 2009-2010¹⁴ and ongoing sporadic cases of transmission of H5N1 influenza from birds to humans have led to heightened international awareness of influenza epidemiology, pathogenesis, and prevention.

Mixed aerobic and anaerobic aspiration pneumonia may account for 10% of cases. This pneumonia is an important consideration for nursing home residents and for individuals with impaired gag reflexes or recent loss of consciousness.

The outbreak in 2000-2001 of inhalation anthrax in the United States adds another microbial differential diagnostic consideration in patients with fulminant community-acquired lower respiratory tract infection.¹⁵ To date, inhalation anthrax remains a rare disease. However, it must be considered in selected clinical and epidemiologic settings (see later). A new human pathogen, severe acute respiratory syndrome–associated coronavirus, emerged and spread worldwide in 2002-2003. No cases have been identified since 2004, but this virus should also be considered in the appropriate clinical and epidemiologic setting.¹⁶

In most published series, no microbiologic diagnosis is established in 50% of patients. This situation is attributable to many factors, including the following:

The common microbial agents producing HCAP, HAP, and VAP are summarized in Table 22-1 and include gram-negative bacilli, S. aureus, Legionella species, and, rarely, viruses such as influenza or respiratory syncytial virus. The last-mentioned viruses are considerations only during the winter months, when they are endemic in the community and may be brought into the hospital by health care workers, visitors, or patients with incubating or active infections.

The relative frequencies and antimicrobial susceptibilities of these respective bacteria may vary considerably from one institution to another. Knowledge of which nosocomial isolates are most common within one’s own institution and community, along with their drug-sensitivity profiles, has important implications with regard to selection of antibiotic therapy, formulation of infection control policies, investigation of potential outbreaks, and selection of antimicrobial agents for the hospital formulary. For example, patients developing severe VAP in ICUs with a high prevalence of carbapenem resistance among gram-negative organisms such as Klebsiella pneumoniae and Acinetobacter baumannii may warrant empiric antimicrobial therapy for these organisms pending culture information. Similarly, nosocomial legionellosis is so uncommon in some institutions that empiric therapy in critically ill patients with nosocomial lower respiratory tract infection does not require coverage of this pathogen. However, in other institutions, nosocomial legionellosis occurs more frequently, and patients with HAP may require empiric treatment for this organism.

Nosocomial pathogens capable of producing HAP can be transmitted directly from one patient to another, as in the case of tuberculosis. However, transmission from health care workers (including respiratory therapists [RTs]), contaminated equipment, or fomites (objects capable of transmitting infection through physical contact with them) is more common, especially for gram-negative bacilli, S. aureus, and viruses. The RT has an important role to play in preventing the transmission and development of nosocomial pneumonia (see further discussion later).

Clinical Manifestations

Patients with community-acquired pneumonia typically have fever and respiratory symptoms, such as cough, sputum production, pleuritic chest pain, and dyspnea. Not all of these symptoms are present all the time, especially in elderly patients in whom the presentation may be subtle. Other problems, such as hoarseness, sore throat, headache, and diarrhea, may accompany certain pathogens. Fever, cough, and sputum production may occur in other illnesses such as acute bronchitis or flare-ups of chronic bronchitis.

In the past, clinicians often distinguished between typical and atypical clinical syndromes as a means of predicting the most likely microbial causes. A typical presentation consisted of the sudden onset of high fever, shaking, chills, and cough with purulent sputum. Such a presentation was considered more common with bacterial pathogens such as pneumococcus and H. influenzae. An atypical presentation was an illness characterized by the gradual onset of fever, headache, constitutional symptoms, diarrhea, and cough, often with minimal sputum production. Coughing was often a relatively minor symptom at the outset, and the illness was initially dominated by nonrespiratory symptoms. Such a presentation was thought to be more common with pathogens such as M. pneumoniae, C. pneumoniae, Legionella species, and viruses. More recent studies have shown that these distinctions are not ironclad and that considerable overlap exists in the clinical presentations of pneumonia with typical and atypical pathogens.¹⁷ The occurrence of concomitant diarrhea, previously considered indicative of legionellosis, is now known to be common in pneumococcal and mycoplasmal pneumonia.

Despite the limitations in predicting with certainty the microbial diagnosis based on the clinical presentation, clinicians use certain historical clues and physical findings at the bedside to determine the likely cause of pneumonia in patients presenting from the community. In patients presenting with high fever, teeth-chattering chills, pleuritic pain, and a cough producing rust-colored sputum, pneumococcal pneumonia is the most likely diagnosis. Patients with pneumonia accompanied by foul-smelling breath, an absent gag reflex, or recent loss of consciousness are most likely to have a mixed aerobic and anaerobic infection as a consequence of aspiration. Community-acquired pneumonia accompanied by hoarseness suggests that the culprit is C. pneumoniae. Pneumonia in a patient with a history of splenectomy suggests infection with an encapsulated pathogen such as pneumococcus or H. influenzae. Epidemics of pneumonia occurring within households or closed communities, such as dormitories or military barracks, suggest pathogens such as M. pneumoniae or C. pneumoniae. Pneumonia accompanied by splenomegaly prompts consideration of psittacosis or Q fever. Bullous myringitis and erythema multiforme are associated with Mycoplasma infection. Relative bradycardia (defined as a heart rate <100 beats/min) in the presence of fever and the absence of preexisting cardiac conduction system disease or beta-blocker therapy, may suggest infection with an atypical pathogen. Pneumonia accompanied by conjunctivitis suggests adenovirus infection.

The clinical presentation of community-acquired pneumonia in elderly patients warrants special mention because it may be subtle. Older individuals with pneumonia may not have a fever or cough and may simply present with shortness of breath, confusion, worsening congestive heart failure (CHF), or failure to thrive.

As noted previously, inhalation anthrax is a rare disease, but it warrants mention because of the small epidemic believed to have been an act of bioterrorism.¹⁵ This outbreak affected mainly postal workers who were exposed to mail containing anthrax spores. Most patients presented with a febrile flulike illness of several days’ duration accompanied by dry cough and shortness of breath. Some patients in whom the diagnosis was not quickly considered went on to develop septic shock, meningitis, and disseminated intravascular coagulation over several days, culminating in death.

Because of a lack of prior host immunity or unique viral virulence factors, patients infected with pandemic influenza strains may have unusually severe presentations. During the 2009-2010 pandemic of H1N1 influenza, clinical presentations varied from mild upper respiratory syndromes to fulminant pneumonias with acute respiratory distress syndrome (ARDS) and shock.¹⁴ Severe acute respiratory syndrome manifests with high fever and myalgia for 3 to 7 days followed by nonproductive cough and progressive hypoxemia with progression to mechanical ventilation in 20%.¹⁶

HCAP, HAP, and VAP usually manifest with new onset of fever in hospitalized or institutionalized patients. Nonintubated patients may have a recent history of vomiting, seizure, or syncope, during which aspiration of oropharyngeal or gastric secretions may have occurred. In intubated patients, VAP traditionally manifests with new onset of fever, purulent endotracheal secretions, and a new pulmonary infiltrate. The diagnosis of HCAP, HAP, or VAP can be extremely difficult to make in patients with preexisting abnormalities on the chest radiograph, such as CHF or ARDS. In mechanically ventilated patients, purulent tracheobronchitis may be accompanied by fever, and in patients with preexisting abnormalities on chest x-ray, the distinction between bronchitis and pneumonia can be especially difficult.

Chest Radiograph

In patients with a compatible clinical syndrome, the diagnosis of community-acquired pneumonia is established by the presence of a new pulmonary infiltrate on the chest radiograph. Not all healthy outpatients with suspected pneumonia require a chest radiograph, and physicians may elect to forego radiography and treat empirically for community-acquired pneumonia in individuals with mild illnesses who are at low risk for morbidity or mortality.

A normal chest radiograph does not exclude the diagnosis of pneumonia. The chest radiograph may be normal in patients with early infection, dehydration, or P. jiroveci infection. The pattern of radiographic abnormality is not diagnostic of the causative agent, although specific radiographic findings should suggest specific microbial differential diagnoses (Table 22-4).

Consolidation involving an entire lobe is called lobar consolidation, whereas bronchopneumonia refers to the presence of a patchy infiltrate surrounding one or more bronchi, without opacification of an entire lobe. Both radiographic patterns suggest the presence of a bacterial pathogen. Pleural effusions are common in patients with bacterial pneumonia and uncommon in patients with viral, P. jiroveci, C. pneumoniae, or fungal pneumonia. Pleural effusions are seen in approximately 10% of patients with M. pneumoniae and Legionella pneumophila pneumonia, and they occur occasionally in patients with reactivation pulmonary tuberculosis. Interstitial infiltrates, especially if diffuse, suggest viral disease, P. jiroveci, or miliary tuberculosis in patients with community-acquired pneumonia. Cavitary infiltrates are seen in reactivation pulmonary tuberculosis; fungal pneumonias, such as histoplasmosis and blastomycosis; nocardiosis; pyogenic lung abscess; and, rarely, P. jiroveci pneumonia. Patients with severe staphylococcal or gram-negative pneumonias may develop small cavities called pneumatoceles. Legionellosis should be seriously considered in sicker patients with pneumonia of a single lobe, which quickly spreads to involve multiple lobes over 24 to 48 hours. Inhalation anthrax manifests with widening of the mediastinal silhouette resulting from mediastinal lymphadenopathy. Parenchymal pulmonary infiltrates are typically absent, and pleural effusions are common.

The chest radiograph may be helpful in diagnosing HCAP or HAP in nonintubated patients with a suspected aspiration event and a previously normal chest film. In such cases, development of a new infiltrate may confirm the clinical suspicion of aspiration pneumonia. The chest radiograph is often less helpful in the diagnosis of VAP because mechanically ventilated patients often have other reasons for radiographic abnormalities, such as ARDS, CHF, pulmonary thromboembolism, alveolar hemorrhage, or atelectasis. In these patients, the accurate diagnosis of a new nosocomial lower respiratory tract infection can be difficult. Clinical diagnosis, defined as the presence of fever, purulent respiratory secretions, new leukocytosis, and a new pulmonary infiltrate, is sensitive but not specific for the diagnosis of VAP. Other strategies to diagnose VAP more accurately have been investigated.

Risk Factors for Mortality and Assessing the Need for Hospitalization

Many cases of community-acquired pneumonia can be managed successfully on an outpatient basis. The challenge for the clinician is to identify individuals at higher risk of morbidity and mortality for whom hospitalization is indicated. Over the past 20 years, numerous studies have analyzed risk factors for mortality in patients with community-acquired pneumonia.17–19 Risk factors predictive of a high risk of death are summarized in Box 22-1.

Fine and associates¹⁹ performed a meta-analysis of 127 cohorts of patients with community-acquired pneumonia. The study examined risk factors for fatal outcome. The overall mortality for the 33,148 patients in these cohorts was 13.7%. Eleven prognostic variables were significantly associated with mortality, including male sex, absence of pleuritic chest pain, hypothermia, systolic hypotension, tachypnea, diabetes mellitus, cancer, neurologic disease, bacteremia, leukopenia, and multilobar infiltrates on chest radiograph. Mortality varied according to the infecting agent and was highest for P. aeruginosa (61.1%), Klebsiella species (35.7%), Escherichia coli (35.3%), and S. aureus (31.8%). Mortality rates for more common pathogens were lower but still substantial and included Legionella species (14.7%), S. pneumoniae (12.3%), C. pneumoniae (9.8%), and M. pneumoniae (1.4%).