Pneumonia

From WiserWiki

[edit] Pneumonia

Randolph J. Lipchik

More than 2 million cases of community-acquired pneumonia (CAP) occur each year in the United States, resulting in approximately 10 million physician visits, more than 500,000 hospitalizations, and approximately 50,000 deaths. Over time the number of microorganisms identified as pathogens has increased, along with new broad-spectrum antibiotics available for treatment. At the same time, common pathogens have become increasingly resistant to frequently used antibiotics, complicating the management of CAP and prompting the development of management guidelines.

[edit] EPIDEMIOLOGY

The actual incidence of pneumonia in ambulatory patients is difficult to estimate because the etiologic agent is rarely identified except in clinical trials, and CAP is not currently considered a reportable disease. Each year in the United States there are 2 to 3 million cases of CAP. The incidence of hospitalization is estimated at 260 cases per 100,000 population but is about fourfold higher in those over age 65. CAP results in about 500,000 hospitalizations annually, with approximately 45,000 deaths; pneumonia is the sixth most common cause of death in the United States. Between 1979 and 1994, pneumonia and influenza–related death rates have increased because of the increasing number of patients over 65 and patients with underlying illnesses. Studies of patients with CAP report mortality rates of 5.1% to 36.5%, averaging about 14%.^[1] An analysis of 1993 hospital discharge data from Washington, Illinois, and Florida revealed death rates of 7.0%, 8.1%, and 9.7%, respectively.^[2] Risk factors for mortality include age, alcoholism, bacteremia, and multilobar involvement on radiographs. Contributing factors include underlying malignancy, immunosuppression, neurologic disease, congestive heart failure, and diabetes. Aspiration, postobstructive, gram-negative, and Staphylococcus aureus forms of pneumonia are also associated with higher mortality risk.

[edit] Pathophysiology

Traditionally thought to be responsible for 60% to 70% of pneumonias, the prevalence of Streptococcus pneumoniae has decreased with the identification of other agents (Table 73-1). The likelihood of each of these agents causing disease in a given patient is not certain, although certain host factors and geographic location may predispose to certain infections (Table 73-2). Travel to southwestern United States, including California and Texas, and contiguous areas of Mexico suggests Coccidioides immitis. Histoplasma capsulatum is endemic in states bordering the Mississippi and Ohio rivers. Blastomyces dermatitidis is endemic in southeastern United States but also in Wisconsin, Minnesota, and neighboring Canadian provinces. Exposure to birds necessitates the addition of psittacosis to the differential diagnosis, and exposure to parturient cats, cattle, or sheep suggests Q fever (Coxiella burnetii).

Table 73-1 Causes of Community-acquired Bacterial Pneumonia

✢High prevalence in specific geographic locations.

Table 73-2 Association of Host Factors with Particular Pathogens

The respiratory tract is a unique system in that it is open to the external environment and therefore continuously exposed to microorganisms, particulate matter, and fumes. In addition to all the organisms that are coughed or sneezed into the environment by others, humans regularly aspirate nasopharyngeal flora during sleep. Multiple defense mechanisms counteract these continuous exposures, including mechanical, anatomic, and immunologic barriers. The cough reflex, the mucociliary transport mechanism, and secretory immunoglobulins remove and neutralize microbes in the upper and central airways. In the alveoli the alveolar macrophages, immunoglobulins, and complement combine to clear organisms from the distal lung. Alterations in mental status may reduce the cough reflex; mucous production and ciliary function can be overcome by viral illness or tobacco smoke; and the immune response can be blunted by many illnesses or medications. Loss of these defenses in the setting of a large inoculum or particularly virulent organism can produce significant infection. Whether or not colonization of the upper airway is necessary before the development of pneumonia is unclear. In the outpatient population the carrier rate for S. pneumoniae is quite high but the incidence of pneumonia quite low. In hospitalized patients, however, colonization by gram-negative organisms probably occurs before development of pneumonia. In a few patients, pneumonia may result from hematogenously spread infection.

Annual vaccination against influenza should reduce its incidence and that of secondary bacterial pneumonias. Vaccination against pneumococcal infection is recommended for patients 65 years and older and for younger persons at increased risk, that is, with anatomic or functional asplenia (including sickle cell disease), cardiovascular disease, pulmonary disease, diabetes mellitus, alcoholism, cirrhosis, and cerebrospinal leaks. The current vaccine is a 23-valent preparation that provides coverage against approximately 90% of the most frequently reported capsular types. Routine revaccination of adults is not currently recommended unless the patient is at high risk for pneumococcal infection (asplenic) and originally received the 14-valent vaccine. Revaccination should also be considered for persons 65 or older who received the vaccine 5 or more years earlier and were under age 65 at the time of primary vaccination. Although evidence is lacking, single revaccination should also be considered in immunocompromised patients if 5 or more years have elapsed since initial vaccination.

[edit] PATIENT EVALUATION

[edit] History

Because of the multiple potential pathogens that cause pneumonia, the history becomes especially important in the evaluation of a patient with pneumonia. Presence or absence of fever, dry or productive cough, acute or gradual onset, and presence of chest pain and dyspnea may help distinguish upper from lower respiratory infection and a ``typical from an ``atypical pneumonia. In contrast to typical (pneumococcal) pneumonia, atypical pneumonia is characterized by lack of sputum production, lack of chest pain, and radiographic infiltrates that are not evident on physical examination. Agents causing atypical pneumonia are Mycoplasma pneumoniae, Chlamydia pneumoniae, viruses, and Coxiella burnetii. Concomitant medical conditions, recent travel, and animal exposure help to direct evaluation and therapy.

[edit] Physical Examination

Attention to all aspects of the physical examination is crucial to determining severity of illness, hospitalization, and possibly treatment. Fever is nonspecific, but a pulse-temperature disparity (normal pulse in the setting of high fever) suggests pneumonia from Mycoplasma, Legionella, Chlamydia, or virus. Tachypnea and cyanosis indicate significant respiratory compromise and thus careful consideration before choosing outpatient vs. inpatient therapy. Examination of the thorax may be unremarkable, reveal evidence of consolidation (dullness to percussion, increased tactile fremitus, egophony), suggest interstitial infiltrates (crackles), or present evidence of a pleural effusion (dullness to percussion, decreased tactile fremitus, decreased breath sounds). Extrapulmonary findings should not be overlooked and can offer clues to the underlying pathogen (Table 73-3). Neurologic disease, altered level of consciousness, and recent seizures suggest aspiration pneumonia. Periodontal disease makes an anaerobic infection more likely.

Table 73-3 Extrapulmonary Findings and Causes of Pneumonia

[edit] DIAGNOSIS

[edit] Diagnostic Procedures

The chest radiograph is the gold standard for determining the presence or absence of pneumonia. For many years, the radiographic pattern was thought to be useful in determining the etiology of the pneumonia, but with more pathogens and more elderly and immunocompromised patients, this has become less reliable. The presence of an abscess, central mass, or pleural effusion, however, is very helpful in management decision making. Because of the lack of specificity of the chest radiograph, supplemental studies are necessary to determine an etiology. Examination of a sputum sample provides data that are available at the presentation and may help guide therapy. Unfortunately, most patients cannot produce a sample, and if they do, it is often contaminated by oral flora. Nonetheless, a sputum smear with fewer than 10 squamous cells and greater than 25 neutrophils per high-power field should be representative of the secretions in the lung. Sputum culture requires 24 to 48 hours and may not provide diagnostic information. For example, the sensitivity for pneumococcus is only about 50%.

Several invasive methods can be employed when sputum is unobtainable, but the risks must be considered compared with the use of empiric antibiotics, a safe and usually successful treatment. Transtracheal aspiration consists of passing a catheter via a large-bore needle through the cricothyroid membrane to allow aspiration of material distal to the oropharynx, thus avoiding contamination by oral flora. Its sensitivity has been questioned, the potential for complications is real, and transtracheal aspiration is rarely employed. Bronchoscopy can be employed to obtain distal samples, but passage through the upper airway makes contamination difficult to avoid. Quantitative cultures obtained during bronchoscopy with a protected brush or by bronchoalveolar lavage may distinguish infection from colonization or contamination. If the patient has already received antibiotics, however, the results have poor predictive value. This approach does have a useful role in evaluating the immunocompromised patient and cases of nonresolving pneumonias, but routine use is not justified because standard antibiotics are effective and low risk.

In addition to sputum cultures, blood cultures should be obtained, although on average only 10% to 15% of patients hospitalized for pneumonia have bacteremia. If a pleural effusion is present, a thoracentesis is performed to obtain material for Gram's stain and culture. Infections with Mycoplasma, Legionella, Chlamydia, Coxiella, and some viruses can be proved with serologic assays, but because convalescent titers must be drawn at least 3 weeks after onset of illness, empiric therapy is still necessary. Cold agglutinins may rise after 7 to 14 days of infection in Mycoplasma infections but are nonspecific, and similar increases can also be seen with influenza. Legionella can be detected in sputum, pleural fluid, or tissue by direct immunofluorescent staining with a specificity of 90% to 100% but a sensitivity of only 25% to 50%. Detection of Legionella antigen in the urine is more useful, with sensitivity as high as 86%. Blood chemistries and leukocyte count with differential are quite nonspecific. An elevated leukocyte count with a left shift is common in any infection, and a low leukocyte count does not rule out a bacterial infection. Although hyponatremia, hypophosphatemia, and liver function abnormalities are associated with Legionella infection, these abnormalities can also be seen with other severe infections.

[edit] Differential Diagnosis

Radiographic pulmonary infiltrates, cough, dyspnea, and fever are the presenting features of not only pneumonia, but many noninfectious conditions as well, including pulmonary embolism, congestive heart failure (CHF), malignancy, vasculitis, collagen vascular disease, hypersensitivity, and idiopathic processes (Table 73-4). Often the presentation and clinical clues help distinguish these from a community-acquired infection, but failure to respond to standard empiric therapy or unexpected progression should alert the physician to reconsider the initial diagnosis. Pulmonary embolism usually presents with sudden onset of dyspnea and often in the setting of immobilization, the perioperative period, CHF, or malignancy. Chest pain and hemoptysis, which signify pulmonary infarction, occur in a minority of patients. The chest radiograph is usually normal in patients without infarction but may show evidence of atelectasis or a small pleural effusion. CHF is usually distinguishable from pneumonia because of elevated jugular venous pressure, no fever, and no productive cough. In patients with chronic obstructive pulmonary disease (COPD), pulmonary edema may produce infiltrates that are asymmetric or focal. Resolution within hours to a day with diuresis will confirm the presence of pulmonary edema.

Table 73-4 Differential Diagnosis of Pneumonia and Selected Noninfectious Pulmonary Diseases

Malignancy may present as a segmental consolidation, atelectasis, or diffuse interstitial pattern but presents clinically with a subacute development of cough or dyspnea. Fever, if present, is usually low grade, and the cough is frequently dry. Radiographic abnormalities are often present without significant symptoms. The abrupt onset of dyspnea, hypoxemia, radiographic infiltrates, variable degrees of hemoptysis, and a falling hematocrit are seen in alveolar hemorrhage. This can occur in the setting of rapidly progressive glomerulonephritis or systemic necrotizing vasculitis (e.g., Wegener's granulomatosis, microscopic polyarteritis, systemic lupus erythematosus). The rapid progression of renal abnormalities and extrapulmonary evidence of vasculitis should suggest a noninfectious process. The presentation of hypersensitivity pneumonitis (e.g., farmer's lung, humidifier lung) includes fever, dry cough, and pulmonary infiltrates within 4 to 6 hours of exposure to inhaled organic antigens (e.g., thermophilic actinomycetes). Eliciting the history of inhalational exposure is the only way to distinguish this illness from CAP.

Several medications have been associated with acute onset of a pneumonia-like illness. Nitrofurantoin can cause fever, dry cough, dyspnea, and infiltrates; a detailed history and peripheral blood eosinophilia are clues to the diagnosis. Amiodarone can also produce pulmonary infiltrates; the diagnosis is often one of exclusion because no distinguishing diagnostic features are specific for this process.

An idiopathic process such as bronchiolitis obliterans with organizing pneumonia (BOOP) may be initially confused with infectious pneumonia because of cough, dyspnea, alveolar infiltrates, and constitutional symptoms. Failure to respond to antibiotics may lead to open lung biopsy, the definitive way to confirm the diagnosis.

[edit] MANAGEMENT

Management of pneumonia must include both antibiotic therapy directed against the causative organism and supportive measures. The latter include rest, adequate hydration to correct for fever-induced fluid loss and poor intake, supplemental oxygen for saturation less than 90%, and analgesia for chest pain. Chest percussion and postural drainage may be useful in selected patients with bronchiectasis or those too weak to generate an adequate cough. Routine use of this time-consuming and labor-intensive modality, however, is not beneficial in uncomplicated pneumonias.

Antibiotic therapy should be administered as quickly as possible once the diagnosis has been confirmed radiographically. Ideally the choice of antibiotic should be guided by a Gram's stain of sputum. Many polymorphonuclear neutrophils (PMNs), no epithelial cells, and a predominant organism allow more specific therapy; most often, however, the choice is empiric, and all the clinical data must be considered before deciding on a regimen. In 1993 the American Thoracic Society published guidelines for the initial management of patients with CAP.^[3] In 1998 the Infectious Diseases Society of America published their guidelines for CAP in immunocompetent adults, which is evidence based where possible.^[4] Both these documents are guidelines and require prospective validation.

Unless the presentation or history suggests a particular pathogen, the macrolides, newer fluoroquinolones, and doxycycline are the drugs of first choice because they are effective for the agents causing the vast majority of CAPs (i.e., S. pneumoniae, M. pneumoniae). For sicker patients with comorbidities, initial therapy may need to be broader. Therapy can be altered or narrowed if sputum or blood cultures yield an organism (Tables 73-5 and 73-6).

Table 73-5 Empiric Antibiotic Selection for Patients with Community-acquired Pneumonia

Table 73-6 Antibiotic Choices for Specific Pathogens

[edit] Penicillin-resistant Pneumococci

The empiric use of penicillin for CAP is no longer acceptable given the rise of penicillin-resistant pneumococci. The incidence of penicillin resistance is approximately 23% in selected centers in western Europe and the United States^[5] but as high as 50% to 60% in South Africa, Spain, eastern Europe, and Korea. The resistance is caused by alteration in penicillin-binding proteins, not the production of a β-lactamase. Of even more concern is the parallel resistance seen to other antibiotics, such as macrolides, trimethoprim-sulfamethoxazole (TMP-SMX), and tetracycline. About 30% of penicillin-resistant pneumococci are resistant to erythromycin, which predicts resistance to azithromycin and clarithromycin. Primary care physicians must be aware of the microbiology in their institutions and geographic locales to provide effective empiric antibiotic therapy.

[edit] Elderly Patients

The diagnosis of pneumonia and therapy in the geriatric population deserves special attention, particularly if the patient is a nursing home resident. Classic symptoms of cough, sputum production, chest pain, and fever occur much less often in weak, debilitated patients. Coughing requires adequate muscle strength, and pleuritic pain and fever result from a vigorous inflammatory response; these may be lacking in elderly patients with poor nutrition or poor general health. Confusion and mental status changes may be the predominant clinical findings. Although S. pneumoniae is the most common pathogen, other agents such as H. influenzae (which produces β-lactamase in greater than 15% of cases) and M. catarrhalis must be considered, particularly in patients with COPD. Gram-negative bacilli are also more common, particularly in the chronically institutionalized patient. M. pneumoniae is an uncommon pathogen in the older patient. Legionella incidence is variable, being more prevalent in certain U.S. regions and during epidemic outbreaks.

[edit] Immunocompromised Patients

Altered immunity can result from HIV infection, leukemia or lymphoma, chemotherapy-induced granulocytopenia, or treatment for a variety of illnesses with long-term steroids and cytotoxic agents. The evaluation and treatment of a pneumonia in these patients should be in hospital. Because the morbidity and mortality of infections are much higher than in the general population, prompt empiric therapy and diagnostic procedures are essential. Initial empiric antibiotic therapy must cover multiple potential pathogens (Table 73-7). Because of the possibility of unusual and multiple pathogens, invasive diagnostic procedures such as bronchoscopy with bronchoalveolar lavage are considered in the first 24 to 48 hours in addition to cultures of sputum, blood, and other fluids. Although pulmonary infiltrates may also represent drug or radiation toxicity or an underlying leukemia or lymphoma, these can be considered only after an infectious etiology has been ruled out.

Table 73-7 Pneumonia in Immunocompromised Patients

Once therapy has been initiated, fever, respiratory and cardiovascular status, and general features such as energy and appetite should be monitored. Most patients receiving appropriate antibiotics will improve within 48 to 72 hours. Fever that continues 24 hours into therapy is not necessarily a failure of antibiotics. A gradual decrease in the maximum daily temperature is the usual response to therapy. Persistent fever with worsening clinical status may indicate a suppurative complication (e.g., empyema), an inappropriate choice of antibiotics, the wrong diagnosis, or drug fever.

Few data address duration of therapy for pneumonia in general. S. pneumoniae should be treated until the patient has been afebrile for at least 72 hours. Mycoplasma and Chlamydia pneumonia should be treated for 14 days, and confirmed cases of Legionella require 14 to 21 days. For hospitalized patients, controversy surrounds when to switch from intravenous to oral antibiotics. In general, once a patient has become afebrile, is clinically improving, can tolerate oral intake, and has a functioning gastrointestinal tract, oral antibiotics can be considered. Radiographic infiltrates may completely clear only after many weeks, particularly with pneumococcal pneumonia. Slow radiographic resolution does not mean a failure of therapy in the face of clinical response, and frequent chest radiographs are not necessary. Consultation by a pulmonary or infectious diseases specialist should be considered for immunocompromised patients, those who fail to respond in a typical manner, and those with suppurative complications or respiratory compromise.

[edit] Indications for Hospitalization

The need for hospitalization must be carefully considered because inadequate therapy can lead to increased morbidity and mortality (Box 73-1). Because rates of hospitalization vary greatly from region to region, data from more than 38,000 patients were used to develop a prediction rule to identify patients with CAP at low risk for mortality.^[6] This was then validated prospectively on approximately 2200 patients in the pneumonia patient outcomes (PORT) cohort study. An algorithm stratifies patients into different risk groups (Fig. 73-1 and Tables 73-8 and 73-9). Outpatient therapy for patients in risk classes I and II, brief inpatient observation for risk class III, and standard inpatient care for risk classes IV and V would theoretically have reduced the number of patients receiving inpatient therapy by more than 30%. When this prediction rule was later used in an emergency department, the percentage of patients initially treated as outpatients increased from 42% to 57%, although this was offset somewhat by a 9% increase in the number of outpatients subsequently admitted to hospital.^[7] Use of such a prediction rule requires further study. Clinical judgment must still ultimately guide such decisions in individual cases.

Figure 73-1 Assessing risk in community-acquired pneumonia (CAP) and identifying patients in risk class I. (From Fine MJ, Auble TE, Yealy DM, et al: N Engl J Med 336:243, 1997.)

Table 73-8 Point Scoring System to Identify Risk in Patients with Pneumonia

Table 73-9 Hospitalization and Mortality According to Pneumonia Risk Class

[edit] EVIDENCE-BASED MEDICINE

The primary source for this chapter was MEDLINE. Electronic searches dating back to 1994 focused on systematic reviews, meta-analyses, and randomized trials.

[edit] REFERENCES