Pneumonia in Adults

This information explains the advice about pneumonia in adults that is set out in NICE guideline CG191. In September 2019 we withdrew some of the advice on antibiotics because it was replaced by advice in the NICE guidelines on pneumonia (community-acquired): antimicrobial prescribing and pneumonia (hospital-acquired): antimicrobial prescribing. Does this information apply to me? Yes, if you have symptoms of pneumonia or have been diagnosed with pneumonia. This information may also be useful if you are a family member or carer of a person who has or may have pneumonia. It does not cover pneumonia that develops in people who: have a tube placed in their airway to help them breathe (called intubation) or are in an intensive care unit or

The incidence of pneumonia is increasing. The exact annual incidence of pneumonia is difficult to determine because it is not a notifiable disease. The incidence varies with geography, population demographics, comorbid conditions, and the severity of the flu season. There are approximately 4 million cases of communityacquired pneumonia and at least 300,000 cases of hospital-acquired pneumonia reported annually. Pneumonia accounts for the largest mortality of any common infectious disease. Pneumonia ranks in the top seven among the major leading causes of death in adults in the USA. It is the fourth leading cause of death in the US elderly population. A large variety of organisms cause pneumonia (Table 15-1). The important complications of pneumonia include abscess formation, pleural effusion and empyema formation, bacteremia and sepsis, septic shock, acute respiratory distress syndrome (ARDS), meningitis, endocarditis, and multiple organ dysfunction or failure.
The rapid initiation of broad-spectrum empirical antibiotic therapy correlates with long-term morbidity and mortality from pneumonia. The earlier the specific infectious agent is appropriately treated the better the outcome is likely to be.
The Centers for Disease Control (CDC) defines pneumonia to be present when the chest radiograph reveals a new or progressive infiltrate, pleural effusion, or cavitation, and any one of the following: (1) a change in the character, quantity, or consistency of the sputum, (2) the pathogen is isolated from the lower respiratory tract, by cultures, or by lung biopsy, (3) a virus or viral antigen is isolated in respiratory secretions, (4) a diagnostic serum IgM antibody titer or a fourfold rise in IgG antibody titers can be demonstrated in paired serum samples, (5) histologic evidence of pneumonia can be demonstrated in tissue biopsy. Note that these general criteria do not include sputum Gram stain or culture.
Pneumonia treatment is simplified by numerous guidelines which are described below. The standard of medical care for the management of pneumonia has evolved considerably over the last decade based on large multicenter outcome studies. The present evidence-based

Treatment of Serious Infection
Serious infection should be treated as a medical emergency and antibiotics must be initiated as soon as possible; there should be no delay in obtaining cultures or awaiting culture and sensitivity data.
approach is based on best-practice data regarding diagnosis, risk stratification, and antibiotic management protocols. However, evolving patterns of antibiotic resistance, emerging pathogens, and new antibiotics require continual reassessment of existing guidelines. There is also sufficient variability between hospitals to require individualization of therapy based on local or institutional antibiogram data.
There is a clear genetic component to the risk of pneumonia and sepsis. Genetic polymorphism for molecules important in antigen recognition and binding such as mannose-binding lectin, CD-14, and toll-like receptors as well as for inflammatory mediators such as tumor necrosis factor, the interleukin-1 family, interleukin-10, and angiotensin converting enzyme are likely to have pharmacologic significance.
There are a great many noninfectious diseases that cause febrile pneumonitis and mimic pneumonia. These include eosinophilic pneumonia, interstitial lung disease associated with connective tissue disorders or vasculitis or pulmonary airway disease, neoplasms, sarcoidosis, ARDS, exposure to inhaled gases or toxins, radiation pneumonitis, cytotoxin therapy, and pulmonary infarction.
Noncytotoxic pharmacologic agents which can mimic a pneumonia include antimicrobial agents, phenytoin, amiodarone, and narcotics. Aspiration pneumonitis, Mendelson's syndrome, occurs following aspiration of acidic gastric content and may present the lobar infiltrate progressing to a generalized pneumonitis (Figure 15-1). Antibiotic therapy is not indicated in the initial treatment of aspiration pneumonitis.
The general typical presentation of a pneumonia is that of a systemic inflammatory response syndrome (SIRS) with evidence of a widened arteriolar-alveolar oxygen diffusion gradient. History and physical examination and microbiologic or serologic studies help confirm the diagnosis. consolidation of the right upper and left lung fields. The accentuated interstitial markings with early alveolar changes reflect evolving local inflammation. Notably, the distribution of early aspiration pneumonitis findings depends on the patient's position at the time of aspiration and need not be localized to lower lung fields.

RISK FACTORS FOR AND MECHANISMS OF PNEUMONIA DEVELOPMENT
The risk factors that determine the virulence and severity of pneumonia are related to both pathogen and host characteristics. Characteristics of the pathogen include the organism, its resistance characteristics, its virulence, and the dose of the inoculum. Host factors are commonly divided into alterations of mucosal integrity or impaired immunity. Mucosal integrity can be compromised by epithelial injury ("viral priming"), mechanical injury such as endotracheal tubes or repetitive mucosal trauma resulting from vigorous tracheal suctioning, chemical injury from aspiration of gastric contents, tobacco abuse, and malnutrition. Impaired host immunity may be due to systemic immunosuppression due to infection, malignancy, malnutrition, liver or renal failure, diabetes mellitus; or the exposure to immunosuppressant agents such as steroids, chemotherapeutic agents, toxins, or pharmaceuticals with immunosuppressant side effects. Notably, blood transfusions have been demonstrated to suppress systemic immunity and to increase the risk of nosocomial infections including pneumonia.
The lung is the body organ with the largest epithelial surface area in continuous contact with the external environment and therefore it is vulnerable. An area of approximately 70 m 2 participates in gas exchange and is continually exposed to airborne inorganic and organic particles, as well as microorganisms. Pulmonary infection can be caused by inhalation of aerosolized inoculate, aspiration of secretions or particulate matter, hematogenous spread, or direct inoculation. The most common mechanism of pneumonia development is by contamination of the naso-oropharynx or upper airway, and subsequent aspiration of infected secretions into the distal respiratory tree. In otherwise healthy patients, pharyngeal secretions have a bacterial concentration of approximately 10 10 organisms/ml. Therefore, aspiration of exceedingly small volumes (<0.1 ml) introduces a high-titer bacterial inoculum into the respiratory tract. In patients with incapacitating or coexisting disease, decreased salivary flow, decreased cough reflexes, or poor oral hygiene (such as occurs in intubated and mechanically ventilated patients) the bacterial, viral, and fungal load of pharyngeal secretions is both greater and more virulent, and therefore a significant risk factor to the development of pneumonia.
Patients with chronic or acute sinusitis and postnasal drip are at risk for the aspiration of contaminated nasopharyngeal secretions. Rarely, in patients with pharyngeal pouches or esophageal (Zenker's) diverticuli, sequestered secretions and food particles may be aspirated during sleep.
One of the causes of contamination of the lung with organisms is the ventilator. The ventilator circuit regularly develops stagnant water as condensate which becomes rapidly colonized with pathogenic bacteria. If this water is inadvertently washed into the endotracheal tube, pneumonia results. Patients who undergo elective surgical tracheostomy must have their oropharynx well suctioned and preferably also decontaminated prior to deflation of the endotracheal tube cuff. A large number of ventilator-associated pneumonias inadvertently result from aspiration of pooled oropharyngeal secretions during the tracheostomy procedure.
Epithelial damage predisposes to successful pathogen colonization and the development of an infective nidus. Such epithelial injury can be caused by chronic or acute lung diseases such as environmental toxins, tobacco abuse, genetic predisposition such as alpha-1 antitrypsin deficiency, or chronic aspiration syndromes. Notably, mucociliary clearance is diminished in smokers and prior infections with Mycoplasma pneumoniae, viruses, and Haemophilus influenza have been shown to both destroy cilia and impair ciliary function. Additionally, the SIRS alters epithelial structure and function.
The presence of the endotracheal tube has been demonstrated in itself to inhibit synchronized ciliary motility. Therefore mucociliary clearance is impaired in intubated patients. Cytologic studies demonstrate structural changes in epithelial glycoprotein binding characteristics, mucus production, ciliary function, surfactant protein composition and concentration, and extracellular enzymatic profiles during severe systemic illness and which are thought to provide a critical initial link to bacterial adhesion, colonization, and infection. Fibronectin is an epithelial coating found on normal mucosa which prevents Gram-negative pathogens from adhering to respiratory epithelium. Loss of fibronectin facilitates bacterial pneumonia development after viral infections, for example. Hematogenous seeding of the lung from other infected sites is another possibility for the development of pneumonia. Infected prostheses, intravascular devices, CLINICAL CAVEAT Aspiration • Aspiration of oropharyngeal secretions is especially common in patients with depressed levels of consciousness due to intoxication, sedation, or central nervous system dysfunction.
• Disorders of deglutition due to neuromuscular disorders, gastroesophageal reflux, esophageal dysmotility, or prolonged nasogastric intubation also predispose to the aspiration of pooled secretions. or endocarditis and even possibly bacterial translocation from the lumen of ischemic abdominal viscus are potential sources of secondary or metastatic infection. Tumors compromise tissue integrity and suppress immunity. Tumors both predispose to colonization and local invasion but also are an important cause of postobstructive inspissation of mucus and secretions.
There is also a hypothesis whereby bacteria are thought to translocate from the gut lumen into the surrounding venous and lymphatic plexuses and induce generalized inflammation or bacteremia and this has been termed the "gut motor hypothesis."

Radiologic Diagnosis of Pneumonia
The chest radiograph is the mainstay of radiologic diagnosis and follow-up of pneumonia. The radiologic picture may lag behind the clinical condition. Interobserver variability remains a limitation to radiologic diagnosis. The radiologic appearance of hospital-and ventilator-acquired pneumonias is difficult to distinguish from atelectasis. However, serial radiographs may reveal atelectasis to resolve more rapidly. It is widely believed but unproven that a patient's hydration status affects the radiologic appearance of an underlying pneumonia. Therefore, the contention that a chest radiograph of a dehydrated patient is less likely to reveal a pneumonia-related infiltrate is an unreliable although attractive hypothesis.
Radiologic infiltrates may be more difficult to identify in patients with hyperexpanded lung fields (COPD), fibrosis and scars, atelectasis, and pulmonary edema. The radiologic appearance of pneumonia has traditionally resulted in a general classification based on appearance and distribution. Alveolar consolidation may occur as a result of inflammatory edema in alveolar segments. When the consolidation involves the entire lobe, it is called lobar pneumonia. Characteristics of lobar pneumonia include air bronchograms, volume loss, and mediastinal shift toward the side of consolidation. Consolidated pneumonias frequently occur with S. pneumoniae and Klebsiella pneumoniae (Figure 15-2). Air bronchograms demonstrate an air-soft tissue interface; they are especially helpful in the indirect delineation of consolidation posterior to the heart in the left hemithorax.
Bronchopneumonia, or lobular pneumonia, involves an inflammatory response primarily at the site of the bronchi and surrounding parenchyma tissue. Bronchopneumonia is usually seen as segmental or patchy infiltrates consistent with partial segmental consolidation. Consolidation is more likely to be locally segmental and involvement of the parenchyma more patchy with bronchopneumonia than with lobar pneumonia.
Bronchopneumonia is especially common when a pneumonic process is superimposed on underlying chronic bronchitis or bronchiectasis. Bronchopneumonias are commonly seen with H. influenzae, S. aureus, and C. pneumoniae (Figure 15-3).
Interstitial pneumonias, also known as peribronchovascular infiltrates, typically appear as reticular or reticulonodular patterns reflecting inflammation primarily localized to the interstitial tissue. Interstitial prominence resembles the early stages of pulmonary edema and is also associated with noninfectious causes of interstitial lung disease which must be included in the differential diagnosis. This interstitial picture is especially common with M. pneumoniae, P. carinii, other atypical organisms, and viruses. The radiologic appearance of Mycoplasma pneumonia is a unilateral or bilateral lobar or segmental infiltrate that shows a patchy or confluent air space disease. More severe cases reveal a diffuse bilateral reticulonodular pattern in both lung fields ( Figure 15-4).
Nodular infiltrates are typically well-defined focal lesions greater than 1 cm 2 on chest radiographs. Nodular infiltrates may be abscesses, fungal or tuberculous granulomas, malignancy, Wegener's granulomatosis, or the necrotic vascular lesions that accompany severe P. aeruginosa infection. Nodular infiltrates may suggest histoplasmosis or tuberculosis. Cavitation is often associated with staphylococcal pneumonia or infection with Aspergillus. There is no good correlation between the radiologic appearance and the causative pathogen.
Pneumonia may be accompanied by pleural fluid collections known as parapneumonic infiltrates (Figure 15-5).
These may resolve or progress to organized empyemas which may be generalized or loculated.

Nonradiologic Diagnosis of Pneumonia
Direct microscopic examination of the Gram-stained sputum has major diagnostic value. The Gram stain is inexpensive, easily available, and has the advantage of immediacy. The sputum Gram stain can help guide antibiotic therapy based on the prevalence of Gram-positive or Gram-negative organisms. The false negative rate for sputum cultures obtained by expectoration is greater than 50%. The quality of sputum is determined by the presence of epithelial cells which indicate oropharyngeal contamination. The presence of >25 squamous epithelial cells per low-power field (×100) is indicative of specimen contamination with oropharyngeal secretions. The presence of one or more macrophages indicates that the sputum sample has been obtained from the lower respiratory tract. The presence of elastin fibers in a sputum sample smear prepared with 40% potassium hydroxide indicates that the specimen has been obtained from the lower respiratory tract and also indicates the presence of a necrotizing pulmonary process. The presence of ≥25 neutrophils per low-power field indicates infection such as pneumonia or tracheobronchitis. Sputum is plated on  infiltrates with prominent interstitial markings. The classic radiographic interstitial of "atypical pneumonia" process reveals accentuated reticular markings within the lung parenchyma as well as septal (Kerly A and B) lines. Air bronchograms are clearly visible bilaterally indicating lung consolidation which accentuates the air-tissue interface. There is no hilar prominence. There is a characteristic fluid meniscus on the right and there is evidence of some aeration within the right lower lobe. The underlying pneumonia is not apparent in this view but could be defined in a lateral decubitus view after the surrounding fluid layers away from the pulmonary tissue. culture media to determine innoculum density (colony forming units or cfu), definitive identification of colony types, and antibiotic susceptibility profiles.
Fever in the critically ill patient is a nonspecific sign of inflammation and does not always reflect the presence of underlying infection. Also, elevated liver function tests may occur with any infection and reflect a nonspecific finding but is often associated with Legionella, tuberculosis, Mycoplasma, Q fever, tularemia, and psittacosis.
Serologic testing is one of the most important diagnostic modalities to identify organisms that are not easily cultured (Table 15-2).

Community-Acquired Pneumonia
Community-acquired pneumonia (CAP) accounts for 10% of all medical intensive care unit (ICU) admissions. Approximately 50% of ICU patients with CAP will require mechanical ventilation. The incidence of CAP is approximately 12 cases per 1000 adults per year in the USA, accounting for almost one million hospital admissions per year. The cost of treating CAP exceeds $10 billion per year in the USA. Although approximately three-quarters of patients with CAP are successfully managed as outpatients, those that require hospitalization account disproportionately for pneumonia-related health care costs. Since the cost of inpatient management of pneumonia exceeds the cost of outpatient care by at least a factor of 15, it is estimated that 90% of the total costs of treating CAP are hospital-related. The mean duration of hospitalization for CAP is 9 days, which includes both routine hospitalization as well as intensive care stay. The average mortality from CAP averages 40%. When CAP is complicated by ARDS, which occurs in approximately 5% of patients with CAP, the mortality rate approaches 70%.
Risk factors for CAP include chronic respiratory or cardiac disease, institutionalization, advanced age, and alcoholism. Additionally, dysphagia and sedation increase the risk for both aspiration and nonaspiration pneumonia in the elderly population. Risk factors for mortality from CAP include multi-lobar infiltrates, positive blood cultures, rapidly progressive infiltrates, polymicrobial infection, hypoalbuminemia, renal insufficiency, respiratory failure, serious comorbidities or advanced age, immunosuppression, altered mental status or coma, and prolonged mechanical ventilation.
Prevention may be an important strategy to reduce the impact of CAP, especially in vulnerable populations such as the elderly. Tobacco and alcohol abuse cessation programs may improve pulmonary function, mucus clearance, and decrease the risk of aspiration. Nutritional status is an under-recognized risk factor for pneumonia development. Additionally, the use of a polyvalent pneumococcal vaccine and annual influenza vaccinations many controversially decrease hospitalization, complications, and death.

Typical Community-Acquired Pneumonias
Streptococcal pneumonia is the single most important cause of CAP but may also be responsible for a significant number of hospital-acquired pneumonias. S. pneumoniae case mortality approaches 40%. Among patients admitted to the ICU with pneumococcal pneumonia with bacteremia, the case mortality rate approaches 80%. Penicillin-resistant S. pneumoniae (PRSP) is the result of configurational changes in penicillin-binding protein structure; the strain was identified in 1967 and its rate of isolation now approaches 40%. S. pneumoniae has also evolved resistance to macrolides, clindamycin, and streptogramins.
Some patients with prior medical problems are more at risk for specific organisms. H. influenzae and K. pneumoniae are especially common isolates from pneumonias in alcoholic patients. Patients with chronic

Sputum Cultures
Sputum cultures that are obtained after the initiation of empiric antibiotic therapy will yield a lower incidence of true positives but a greater incidence of false negatives. Empiric initial antibiotic therapy for CAP must provide effect of coverage for S. pneumoniae; agents that also cover Legionella and Mycoplasma should be considered. Antibiotics of choice are ceftriaxone or cefotaxime with the addition of either a macrolide (e.g., azithromycin) or a fluoroquinolone.

CRITICAL CARE: THE REQUISITES IN ANESTHESIOLOGY
Fine et al. have used both clinical investigation and meta-analysis to identify and quantify the effect of risk factors on mortality. The Fine criteria are described in Table 15-3. The Fine Prediction Rule is an algorithmic application of the Fine criteria to determine the optimal point of care (Table 15-4).
Blood cultures have low yield and are not recommended in the workup of uncomplicated CAP. On the other hand, in severe CAP, especially pneumococcal in origin, bacteremia is an important predictor of mortality.
The Infectious Disease Society of America (IDSA) and the American Thoracic Society (ATS) have promulgated guidelines for the treatment of CAP (Table 15-5).

Atypical Community-Acquired Pneumonias
The term "atypical pneumonia" was originally used to describe a clinical syndrome of pneumonia which differed clinically and radiologically from typical pneumococcal pneumonia. Since atypical pathogens were more difficult to identify on Gram stain and cultures, and the term "atypical pneumonia" was coined. The atypical pneumonias can cause serious clinical syndromes with high morbidity and mortality and frequently require ICU care. Presently, the diagnostic tests for atypical pathogens include polymerase chain reaction, complement fixation, microimmunofluorescence, and enzyme immunoassay (for C. pneumoniae and M. pneumoniae); others include the serum IgM antibody ("cold agglutinin") for Mycoplasma, and direct fluorescent antibody (DFA) and indirect immunofluorescence antibody test (for Legionella spp.). The most established of these techniques is complement fixation which is relatively uncomplicated but has low sensitivity. The fluorescence techniques determine antibody titer in diluted serum samples but are subjective and costly. Enzyme-linked immunosorbent assays (ELISA) to not require dilution and because the results are measured in optical density (OD) units there is less subjectivity. Urinary antigen testing for Legionella pneumophilia serogroup 1 is a nonserologic test which becomes positive only after a few days of illness and may require a second test during the convalescent phase of illness to confirm the diagnosis, if such confirmation is necessary.
The type of environment may play an important role in the organisms to which the patient is exposed. Patients who are immunocompromised, who live in shelters, or who are incarcerated are at especially high risk for infection with Mycobacterium tuberculosis. Hunters and

CLINICAL CAVEAT Presentation of Atypical Pneumonia
The hallmark presentation of atypical pneumonia is a nonproductive cough.

Viral Pneumonia
Viral pneumonias typically occur in the setting of an outbreak and occur predominantly in the winter and early spring. Patients at risk for viral pneumonias include those with underlying heart disease, chronic pulmonary disease, and pregnancy. Pulmonary infiltrates can occur in 20% of young adults with varicella but a frank pneumonia presentation is rare and is more suggestive of a bacterial superinfection (Figure 15-6). Herpetic pneumonias occur with high frequency in postoperative cardiac and transplant patients. The true incidence is probably under-recognized. The route of infection is probably the aspiration of saliva contaminated by the virus.   The clinical picture is usually limited to hypoxia and dyspnea. The virus may be isolated in culture. Treatment is with antiviral agents such as acyclovir. Pulmonary Hantavirus syndrome with the Sin Nombre virus was identified in 1993 in the American Southwest and is acquired through exposure to the deer mouse, its excrement, or contaminated dust; however, the disease has been reported throughout the USA. The syndrome is characterized by a nonspecific 3-6-day prodrome followed by hypoxemia and a sepsis syndrome. Although the clinical picture suggests a pneumonia, the microscopic picture is more consistent with pulmonary capillary leak syndrome such as occurs in ARDS. The case fatality rate for Hantavirus pulmonary syndrome is 50% or more.
Over the last two years there has been a growing concern over large-scale epidemics of viral pneumonias. The one with the greatest impact, severe acute respiratory syndrome (SARS), is a coronavirus infection that causes pneumonia and severe ARDS. It was first recognized in Southeast Asia in November 2002 from where it spread to other countries. Repeat epidemic episodes are likely. The SARS mortality rate worldwide is approximately 10.5%; the ICU admission rate ranged from 20% to 38%; over 60% of ICU patients require mechanical ventilatory support. The mortality rate of SARS patients admitted to the ICU ranges from 5% to 67%.
Emerging diseases are likely to become more of a global threat as viruses increasingly jump species and mutate rapidly.

Fungal Pneumonia
Fungal pneumonias represent a subset of CAPs caused by endemic fungi such as Histoplasma capsulatum, Blastomyces dermatitidis, and Coccidioides immitis. However, fungal pneumonias also represent opportunistic infections in immunocompromised populations and are likely very important but largely under-diagnosed pathogens in nosocomial and especially ventilatorassociated pneumonias. Histoplasma is found in soil contaminated with bird and bat excreta. Histoplasmosis is the most common systemic mycosis in the USA and is characteristically found in the Ohio and Mississippi river valleys. Although an estimated 500,000 people develop histoplasmosis in the USA annually, less than 1% develop a clinically symptomatic disease. In immunocompetent individuals with intact T-cell function, infection is usually subclinical and self-limited. In those patients who develop symptomatic histoplasmosis, the presentation is usually that of a flu-like syndrome with a nonproductive cough and pleuritic chest pain. Less than 5% of infected individuals develop rheumatologic syndromes such as pericarditis and arthralgias or inflammatory granulomatous mediastinitis. In patients with impaired immunologic function or patients who have received a high inoculum exposure, such as construction or agricultural workers, there may be an acute overwhelming pulmonary histoplasmosis characterized by profound hypoxemia and diffuse bilateral pulmonary infiltrates. Rare complications of histoplasmosis include chronic upper lobe cavitary pneumonic disease, fibrosing mediastinitis, and progressive disseminated histoplasmosis. Definitive diagnosis usually requires immunodiagnostic testing which is not definitive, or periodic acid Schiff staining. Treatment is either with itraconazole or amphotericin B; the concomitant systemic corticosteroid therapy should be considered (Figure 15-7).
Blastomyces is also found in moist soil rich with decaying organic material. Blastomycosis is endemic along the Mississippi and Ohio river basins as well as the Great Lakes. Approximately 75% of infected patients have isolated lung involvement and the remaining 25% develop a disseminated lymphohematogenous disease involving primarily the skin, bone, and genitourinary system. Pulmonary blastomycosis develops approximately six weeks after exposure and presents as a flu-like syndrome with a productive mucopurulent sputum. Chest radiographs reveal nonspecific diffuse reticulonodular infiltrates. Chronic common blastomycosis is similar to tuberculosis and is accompanied by night sweats, weight loss, and productive cough, and may progress to cavitary disease. Blastomycosis can be diagnosed by PAS or silver staining of sputum. Ketoconazole, itraconazole, or fluconazole are the mainstay therapies (Figure 15-8).
Coccidioides is found in semiarid desert soil. Coccidiomycosis is endemic in the southwestern USA and approximately 100,000 new cases occur annually in the Sonoran Desert and the Central (San Joaquin) Valley of California. Approximately 40% of infected individuals develop clinically evident disease in one to three weeks following exposure. The disease is typically a self-limited flu-like pneumonitis characterized by nonproductive CLINICAL CAVEAT SARS • The most common clinical symptoms and signs of SARS are fever, cough, dyspnea, myalgias, and malaise.

•
The chest radiograph pattern ranges from focal infiltrates to diffuse airspace disease. The risk of mortality increases with advanced age, comorbidities, a high lactic dehydrogenase, or a high neutrophil count at admission. Pneumonia in Adults cough and pleuritic chest pain. Rheumatologic complications occur in about 20% of patients. Radiographic findings are typically patchy or nodular infiltrates which may coalesce to mimic those of other CAPs. Some 20% of patients develop pleural effusions and hilar adenopathy. About 5% of patients develop chronic pulmonary cavities or nodules which are considered a hallmark of coccidiomycosis. Immunocompromised patients are at especially high risk for disseminated disease which is manifested by plaques and papules, pustules, and chronic granulomatous meningitis. Serologic testing to IgM using precipitin or similar testing, complement fixation of IgG, or skin tests are commonly employed diagnostic modalities. Treatment is with ketoconazole, fluconazole, or amphotericin B (Figure 15-9).

Nosocomial Pneumonia
The definition of a nosocomial pneumonia is one that starts at least 48 hours after hospital admission. Nosocomial pneumonia or hospital-acquired pneumonia   is the main cause of nosocomial mortality; 60% of all nosocomial infections that result in death are due to nosocomial pneumonias.
In patients who are hospitalized for stroke, the development of pneumonia increases the 30-day risk of death threefold. The development of a nosocomial pneumonia increases the length of hospital stay by an average of 6-7 days; and the cost of treating a hospital-acquired pneumonia ranges from $5000 for patients treated on wards to $20,000-$40,000 per case of ventilatorassociated pneumonia treated in ICUs.
Drug-resistant pathogens increasingly account for both community-and hospital-acquired pneumonia. The need for infectious disease (ID) services consultation is subjective and provider-specific. In institutions with restricted antibiotic formularies it may be necessary to obtain ID approval for the use of specific antibiotics; such approval is not synonymous with a formal ID consultation. An ID consultation might be considered in high-risk circumstances such as immunosuppressed patients, patients with complicated histories of frequent infections or the risk of pathogen resistance is high, and in situations where patients' conditions worsen despite aggressive antibiotic therapy.
Nosocomial pneumonias are often divided into hospitalacquired pneumonias (HAPs) which are nosocomial pneumonias that develop in patients who are not mechanically ventilated either on the wards or in the ICU; and ventilator-associated pneumonias (VAPs) which occur in those patients who are receiving mechanical ventilation. VAP usually refers to patients in the ICU; but VAPs may occur also in those patients who are receiving chronic mechanical ventilation outside the ICU.

Hospital-Acquired Pneumonia
HAP has an incidence of 0.5-1.0% of hospital admissions. Pneumonia is the second most common cause of, and accounts for up to 20% of, bacteremia and sepsis in hospitalized patients. Risk factors include advanced age, malnutrition, obesity, impaired immunity, depressed level of consciousness, and prior surgery. HAP due to anaerobes should be considered primarily in the setting of recent abdominal surgery, anaerobic infection, or suspected aspiration. Patients who are admitted with pneumonia from the community but come from an institutionalized setting or who have been recently hospitalized are treated as though they have HAP.
"Severe HAP" is defined based on admission to the ICU, respiratory failure based on the need for an FIO 2 > 35% to maintain SpO 2 > 90% or the need for mechanical ventilation, rapid radiologic progression of lung infiltrate, or evidence of severe sepsis with hypotension and/or endorgan dysfunction.
The core pathogens for HAP are S. aureus, S. pneumoniae, and the enteric Gram-negative bacilli (EGNB). The treatment of early-onset HAP must provide coverage against the core pathogens. The recommended regimens for the treatment of HAP include ceftriaxone, cefuroxime, or cefotaxime ± a macrolide. However, antimicrobial resistance is developing rapidly. Monotherapy with a fourth-generation cephalosporin may be cost-effective therapy for HAP. HAP S. aureus pneumonia due to methicillin-/oxacillin-resistant S. aureus (MRSA/ORSA) strains will likely require vancomycin or linezolid (Zyvox) for effective therapy.
Legionella or opportunistic Aspergillus should be suspected as a HAP in the setting of the administration of highdose steroids or the new onset or worsening of pulmonary symptoms in hospitalized patients. Endemic Aspergillus infections have been recently related to hospital construction and renovation where immunocompromised or critically ill patients have been exposed to construction dust.

Ventilator-Associated Pneumonia
A nosocomial pneumonia that develops at least 48 hours after the initiation of mechanical ventilation is known as a VAP. Nosocomial pneumonia accrues an approximate 25% of adult ICU patients, at an incidence of 21 times greater than that of non-ICU patients. Approximately 17% of nosocomial pneumonias occur in the 1% of hospitalized patients who received mechanical ventilation. VAP has an incidence of 3-20% in patients receiving short-term mechanical ventilation. The cumulative incidence of VAP increases with the number of ventilator days. Statistically, the incremental risk of developing VAP during mechanical ventilation ranges from 1% to 3% per day.

CLINICAL CAVEAT Nosocomial Pneumonia
• Clinical diagnosis is a relatively poor indicator of nosocomial pneumonia and chest radiography is also nonspecific.
• Empiric therapy for nosocomial pneumonia must provide adequate coverage against the core organisms responsible for the majority of nosocomial pneumonias in any given institution.
• Antibiograms that compile statistics regarding antibiotic efficacy profiles for pathogens in any given institution must also be consulted.
• Strategies to decrease the incidence of nosocomial pneumonia include the use of routine infectioncontrol practices, vigorous handwashing and decontamination of equipment by hospital staff, nutritional assessment and support, and the isolation of patients with resistant respiratory pathogens.

Pneumonia in Adults
Important risk factors for VAP include emergency surgery, intubation, recumbent positioning, oral pharyngeal secretions, pre-existing cardiopulmonary or neurologic disease, severity of illness, and tracheostomy. The duration of mechanical ventilation and the administration of systemic glucocorticoids independently predict risk of VAP. Patients who require emergency intubation, emergent re-intubation, or receive a tracheostomy are at higher risk of developing VAP. Occult VAP is a common cause of failure to wean from mechanical ventilation. Endotracheal tubes predispose to pneumonia formation because they bypass the normal filtering mechanisms in the upper airways and provide a direct conduit for the aspiration of pathogens into the lower respiratory tree, they have been shown to impair mucociliary clearance and epithelial ciliary function, they rapidly develop a contaminated biofilm which is then mechanically introduced distally routine suctioning, and the endotracheal tube cuff interferes with deglutition and also promotes the pooling of oral pharyngeal secretions above the endotracheal tube cuff. The high-volume low-pressure endotracheal tube cuffs used in clinical practice do not completely occlude the upper airway and therefore do not prevent either gross aspiration syndromes or microaspiration of oral pharyngeal secretions.
VAP, similar to any other nosocomial infection, must be managed especially aggressively in those patients who have implanted foreign bodies such as vascular graft material, arthroplasties, and pacemakers and defibrillators. The virulence of nosocomial pathogens and the potential for immunologic compromise in inpatients makes hematogenous seeding of foreign bodies especially likely and thereafter very difficult to treat.
VAP is also further characterized based on the time of onset of pneumonia. Early-onset VAP occurs within the first 96 hours of initiation of mechanical ventilation and is most commonly attributable to antibiotic-sensitive organisms such as S. aureus, H. influenzae, S. pneumoniae, and EGNB (Table 15-6). Late-onset VAP is pneumonia which is diagnosed 96 hours or more after the initiation of mechanical ventilation and is typically caused by MRSA, ORSA, P. aeruginosa, Enterobacter spp., or Acinetobacter spp. These organisms are more likely to be resistant and to contribute to morbidity and mortality. Therefore, the key core pathogens in VAP are generally P. aeruginosa, S. pneumoniae, Acinetobacter spp., S. aureus, and EGNB.
VAP is polymicrobial approximately 50% of the time. Empiric antibiotic therapy for VAP must account for both aggressive combination therapy based on an antibiogram specific to the ICU, as well as criteria for the duration of therapy. As microbiological data become available, usually within 48 hours, the empiric therapy should be tapered or modified accordingly. A key dilemma in the diagnosis of VAP is the differentiation between endotracheal tube colonization, airway colonization, bronchitis, and true pneumonia. Colonization can be defined as the persistence of microorganisms at an anatomic site without evidence of host response or local invasion. The differentiation between pneumonia and colonization is especially difficult when the diagnosis is based on nonquantitative cultures obtained by simple endotracheal tube suctioning. A variety of microorganisms normally colonize the upper respiratory tract and these include Viridans streptococci, Streptococcus pyogenes, Neisseria spp., Moraxella, Corynebacterium, Lactobacillus, and Candida spp.
S. aureus is an extremely common cause of communityand hospital-acquired pneumonia. However, in hospitals its endemic rate of colonization is so high as to make it an almost ubiquitous hospital pathogen. MRSA and ORSA account for 70-80% of hospital isolates of S. aureus. Resistance is a function of both an altered penicillin and binding protein which confers beta-lactam resistance, as well as target modification which confers macrolides and possibly fluoroquinolone resistance. Vancomycin has become the mainstay therapy for S. aureus, but the emergence of vancomycin-resistant S. aureus (VRSA) and vancomycin (glycopeptide) intermediate-sensitive S. aureus (VISA or GISA) increasingly requires complex therapy with agents such as Quinupristin/dafopristin (Synercid) or Linezolid.
Gram-negative pathogens account for approximately 60% of clinical laboratory isolates in patients with nosocomial pneumonia. The enterobacteriaceae species that are most commonly seen include K. pneumoniae, Proteus spp., Serratia spp., Citrobacter spp., and Escherichia spp. These agents have been resistant to beta-lactams through the production of bata-lactamases; however, extended-spectrum beta-lactamases (ESBL) and beta-lactamases resistant to beta-lactamase inhibitors (inhibitor-resistant TEM-derived beta lactamases) have become significant.  faecium and E. faecalis are becoming serious threats. However, the former is vastly more common than the latter at the present time. Linezolid and Synercid are important antibiotics against this class of organism.

Vancomycin resistance in Enterococcus and
Staphylococcus is a reflection of its overuse. It is estimated that 60-65% of vancomycin use is inappropriate. Pseudomonas aeruginosa is an extremely common pathogen for HAP and VAP with a higher mortality rate than most other Gram-negative bacilli. Plasmid-mediated resistance is rapid. Pseudomonas pneumonia increases ICU mortality by a factor of 2.6-6.4. In patients who have P. aeruginosa or acinetobacter infections, crude ICU mortality exceeds 70%.
Fungal pneumonias are both under-diagnosed and under-treated in critically ill patients. Mycotic infection is known to occur in the presence of adjunctive antibiotic therapy, immunosuppression, and severe illness. Unfortunately, fungal infection is more likely be a postmortem diagnosis than a clinical one. A key limitation is the inability to culture rapidly specimens containing fungal pathogens and perform sensitivity testing on isolates. The development of antifungal agents with decreased systemic toxicity has facilitated the treatment of mycoses in critically ill patients. Specific criteria for initiation and discontinuation of systemic fungal agents remain controversial.
There is no gold standard method for the diagnosis of VAP. Studies have repeatedly been unable to demonstrate with statistical significance the clear value of any technique over any other. However, a number of diagnostic modalities are used in practice: endotracheal tube aspirates (ETA); bronchoscope-directed protected brushings (BDPB); protected brush specimen (PBS); bronchoscopic bronchoalveolar lavage (BBAL); and blind BAL or QTL which are quantitative tracheal lavage techniques. Endotracheal tube aspirates are the most commonly performed method of culturing respiratory secretions. Since endotracheal tube aspirates culture both contaminants and colonization as well as infectious pneumonia, the use of semiquantitative cultures and the determination of sputum neutrophil count can greatly increase their sensitivity. Bronchoscopy is an important adjunct to the diagnosis of VAP, giving direct visualization of the airways which can reveal inflammation, mucus, and mucus plugging. A protected brush specimen can be obtained either blindly or with a bronchoscope. The protected brush is designed to reduce false positive cultures which are due to contamination from upper airway secretions and may represent only tracheobronchitis rather than true pneumonia. The protected brush is housed within a closed cannula. When the cannula is passed into the distal airways, the brush is advanced through the occluding plug of the cannula and is exposed directly to distal respiratory secretions. Structures are retracted into the cannula and removed for plating and culture. Quantitative cultures revealing colony growth ≥10 3 colony forming units (cfu)/ml are considered positive and correspond to a bacterial density of 10 5 -10 6 cfu/ml in undiluted respiratory secretions.
Bronchoalveolar lavage is performed through a bronchoscope by washing a specific lung segment with isotonic saline and collecting the effluent directly through the bronchoscope suction for culture. When the BAL is performed using a wedged catheter tip, which isolates the lung segment during lavage, the technique is known as a protected BAL. A minimum lavage volume of approximately 120 ml is recommended for adequate sampling; when smaller volumes of saline (10-50 ml) are used, the technique is called a mini-BAL. A threshold of 10 4 cfu/ml is considered positive for pneumonia when the BAL is used but this corresponds to the same bacterial titer in undiluted respiratory secretions as a protected brush. The sensitivity and specificity of BAL is considered to range from 70% to 100%. Open lung biopsy is the most definitive diagnostic procedure for histopathologic diagnosis of pneumonia in immunocompromised hosts.

Treatment of Ventilator-Associated Pneumonia
The risk of undertreatment is increased morbidity and mortality. The risk of overtreatment includes the development of resistant strains of bacteria, the significant cost of continued surveillance cultures, medications, consultations, drug level monitoring, and the increased potential for medication errors.

•
The single best predictor of mortality in patients with VAP is the resolution of pulmonary shunting as measured by the pAO 2 / FIO 2 ratio.
• Successful prevention of VAP is a clearly more costeffective strategy then surveillance and treatment.

•
The avoidance of patient cross-contamination by vigorous handwashing is probably the most important precaution that health care workers can take to prevent spread of pneumonia and other infections in hospitals. Universal precautions should be considered universally.
• The use of endotracheal tubes designed to permit continuous removal of pooled subglottic secretions has been proven to decrease the incidence of VAP.

᭤ ᭤
Pleural or parapneumonic effusions are associated with up to 50% of bacterial pneumonias. These effusions can be free or loculated and may either resolve or develop into empyema. These effusions should be drained immediately if there is radiologic evidence of an air-fluid level. Parapneumonic effusions should be evaluated by paracentesis for Gram stain and culture, pH, LDH, protein, and glucose in order to differentiate between transudative and exudative fluid. Grossly purulent fluid should be drained.
The clinical pulmonary infections score (CPIS) has utility in both detecting the onset of VAP and also determining the sufficiency and adequacy of treatment. The CPIS also correlates with 28-day mortality. The diagnosis of pneumonia is generally based upon variations of the CPIS originally developed by Pugin et al. in 1990 (Table 15-7). The CPIS has a specificity (85-95%) similar to that of bronchoscopic diagnosis of pneumonia. The key parameters include temperature, quantity of secretions, leukocyte count, chest radiographic findings, hypoxemia, and BAL Gram stain and culture. A CPIS of >6 indicates infectious pneumonia with a sensitivity of 93% and specificity of 100%. Therefore, a CPIS of >6 almost excludes acute lung injury, pulmonary edema, or atelectasis as the etiology of a pulmonary infiltrate.

Pneumonia in Patients with Compromised Immune Status
The causes of pneumonia in the immunocompromised host include all pathogens listed above and extend to opportunistic pathogens. Other causes of pulmonary infection in immunocompromised patients include Mycobacterium tuberculosis and M. avium-intracellulare, Cryptococcus, Toxoplasma, and cytomegalovirus. Many of these infections also involve other organs such as the brain and the eye (Figures 15-10 and 15-11). Pneumonia in immunocompromised patients presents atypically. Dyspnea, hypoxemia, nonproductive cough, and a generalized radiographic infiltrate frequently constitute the presenting picture. Such patients may be hypothermic rather than febrile, and may have a leukopenia rather than leukocytosis. The responsible pathogen may be isolated in only 40% of immunocompromised patients. The early evolution of systemic sepsis in this population makes rapid diagnosis and intervention extremely important.
Pneumocystis carinii (PCP) is a protozoan pneumonia which is common in HIV-infected patients. The pneumonia

KEY POINTS-Cont'd
• Patient positioning in semi-recumbent rather than supine position decreases the incidence of VAP. The elevation of the head of the patient's bed can decrease the risk of aspiration.
• More controversial strategies include stress ulcer prophylaxis, selective digestive decontamination, aggressive oral and dental care, and frequent changes of ventilator circuits and filters.
• Potential future developments include the use of bioactive and drug-impregnated biomaterials for endotracheal tube construction.

᭤ ᭤
can be rapidly progressive and lethal. The classic "ground glass" appearance on chest radiographs is nonspecific and may represent early ARDS. The initial antibiotic choice for PCP is trimethoprim-sulfamethoxazole (TMP-SMX) and co-administration of steroids must be considered. The initial response to therapy is frequently a clinical deterioration which may require intubation and mechanical ventilation. Treatment with TMP-SMX is considered to have failed if a favorable response is not apparent within three to five days. The second-line agent is intravenous pentamidine. Pentamidine is associated with significant side effects which include QTc prolongation, torsades de points, renal insufficiency and pancreatitis, neutropenia, as well as hyper-and hypoglycemia ( Figures  15-12 and 15-13).

Fluid Resuscitation
In all patients with severe infective processes, supportive therapy is necessary to prevent secondary injury and the development of complications. Fluid resuscitation is a basic tenet of managing infection. Many patients are unable to "drink plenty of fluids" and therefore require intravenous hydration. Both systemic infection and   systemic inflammatory response produce early peripheral vasodilatation and capillary leak which result in diminished intravascular fluid volume. Since increased stroke volume is usually not possible in such patients, the decreased intravascular volume is initially compensated for by tachycardia to maintain cardiac output and oxygen delivery to tissues. In patients who are unable to mount a compensatory tachycardia (diabetes, myocardial pump dysfunction, beta-blocker therapy, etc.) there are early signs of diminished peripheral perfusion and shock. In the setting of prolonged under-resuscitation, the inflammatory cascade is activated, regional hypoperfusion results, and acidosis and elevated serum lactate, acute renal failure due to acute tubular necrosis from regional underperfusion results, and the patient becomes progressively more difficult to resuscitate later.
Vasopressors should not be used as the mainstay of therapy for sepsis. Although vasopressors may be used to temporize while fluid resuscitation is ongoing, the substitution of vasopressors for fluid results in progressive tissue ischemia and organ failure. The development of progressive hypoxemia with intravascular volume therapy does not necessarily mean intravascular fluid overload -it is also consistent with a worsening of the pneumonia or the interval development of ARDS.
There are no data to support the use of colloid over crystalloid for intravascular volume resuscitation. However, emerging data may suggest that judicious use of colloids may help maintain intravascular oncotic pressure and decrease edema formation. Intravascular volume replacement therapy should be titrated to urine output, signs of improvement in organ perfusion (mental status, cyanosis resolution, etc.), serum lactate and acid-base balance, central venous filling or other preload measure correlated with cardiac output, or echocardiographic determination of end systolic volume and cardiac performance. A history of poor cardiac function, congestive heart failure, or pulmonary edema are not contraindications to intravascular volume replacement.

Nutritional Support
Patients with severe infection are highly catabolic and nutritional support should be instituted immediately, if possible. Many patients with pneumonia are already nutritionally depleted on initial presentation, and baseline nutritional status (pre-albumin, transferrin) should be documented. Daily calorie counts and serial nutritional indices may be required to ensure adequacy of nutritional support. Unless there is a definite contraindication to enteric feeding, it is the route of choice for nutritional support. Nutritional depletion has been shown to increase bacterial adherence to the airways, reduce alveolar macrophage function, impair neutrophil and macrophage recruitment, decrease levels of circulating complement factors, and reduce levels of airway IgA.

Antibiotic Principles
The ability of an antibiotic to exert its pharmacologic effect is related to penetration into infected tissue and achievement of adequate tissue concentrations. Inappropriate treatment of pneumonia is most often due to either nonsusceptibility or resistance of the pathogen to the chosen antibiotic regimen. Inappropriate initial antibiotic selection can range from 27% to 73% and has a dramatic impact on outcome measured by length of stay, multiple organ dysfunction syndrome, and mortality. Antibiotics that have good penetration into respiratory secretions that is not dependent on localized inflammation include the quinolones, azithromycin, clarithromycin, tetracycline, clindamycin, and TMP-SMX. Antibiotics that have impaired penetration into respiratory secretions and/or are dependent on inflammation for concentration within lung tissue include aminoglycosides and the beta-lactams (penicillins, cephalosporins, and carbapenems). Aminoglycosides have poor pulmonary penetration, and especially do not penetrate well into respiratory secretions. Therefore aminoglycosides should Upright anteroposterior (AP) view. Widespread bilateral interstitial and alveolar infiltrates resembling ARDS. Interstitial markings result in a "honeycomb" appearance. An endotracheal tube is present. not be used as monotherapy for Gram-negative pneumonias. Azithromycin has replaced erythromycin in clinical practice because erythromycin has side effects of both QTc prolongation as well as enteric prokinesis which causes cramping discomfort. Fluoroquinolones are also associated with QTc prolongation, especially at higher doses.
Pathogen susceptibilities are often reported in terms of the minimum inhibitory concentration (MIC); the agent of choice will therefore depend on the MIC, guidelines and restrictions, and cost. Optimal pharmacologic action of antibiotics is that for which every concentrationdependent mechanism of action occurs when C max /MIC is greater than 10.
In any patient who has failed to respond clinically after 48-72 hours of antimicrobial therapy, the therapeutic strategy must be reevaluated. Generally, this will require further workup including revaluation of culture and sensitivity data, consideration of superinfection, reconsideration of drug dosing or route of administration, and consideration of an alternative diagnosis. Evaluation of the patient's immunologic status may also be necessary.
Other routes of antibiotic therapy may be important in the care of ICU patients. Inhaled antibiotic therapy is probably underutilized. Inhaled antibiotics should be considered for severe HAP and in the treatment of refractory pneumonia in populations such as patients with cystic fibrosis.
Mechanisms of antimicrobial resistance include targetmediated resistance, enzymatic inactivation, cell membrane permeability alterations, and active expulsion of antibiotics from bacterial cells. Target-mediated resistance is a result of alterations in the number or affinity of antimicrobial binding sites within bacteria. Enzymatic inactivation results from the production of bacterial enzymes often carried by plasmids, which degrade antibiotic agents. Beta-lactamase production is the most common form of enzymatic resistance. Cell membrane permeability changes cause decreased antimicrobial uptake into bacterial cells. The promotion of active reflux of antibiotic out of bacterial cells is an energydependent mechanism which is carried on by plasmids and the creation of molecules known as transposons.
The rotation of first-line antibiotics of choice has been demonstrated to decrease the rate of evolution of pathogen resistance. Therefore, the choice of antibiotic should also be made in accordance with the preferred antibiotic in any particular cycle at any particular hospital.

Nonantibiotic Adjunctive Pharmacologic Therapy
In rare cases the use of immunomodulatory agents such as granulocyte colony-stimulating factor (GCSF) may need to be considered to enhance host polymorphonuclear leukocyte (PMN) response.
Bronchodilators are a very important adjunctive therapy in those patients with reactive airway disease who develop a superimposed respiratory tract infection. Since airway infection can predispose some patients to bronchospasm and thereby interfere with their ability to cough and clear secretions, bronchodilator therapy is beneficial in this population. Patients who use bronchodilator therapy regularly prior to admission may require an increase in the dose or frequency. Nonsteroidal bronchodilators can be either beta-2 agonists such as albuterol or anticholinergic such as Atrovent. The beta-2 sympathomimetics can still precipitate tachycardia. Lev-albuterol is a selective isomer with significantly less cardiac activity and possibly also a longer effective half-life.
Steroids are important bronchodilators in patients with underlying chronic obstructive pulmonary disease (COPD) or severe reactive airway disease. Methylprednisolone is the intravenous steroid bronchodilator of choice; prednisone is administered enterally. Patients for whom steroids have been chronically prescribed prior to admission require supplemental "stress doses" to avoid adrenal insufficiency in the setting of infectious stress. Mucolytic agents and expectorants may have a role when mucus clearance is difficult in the setting of dehydration or inspissated mucus is difficult for the patient to clear. N-acetylcysteine, recombinant DNase, and guaifenesin are prototypic mucolytic agents. Nacetylcysteine can precipitate bronchospasm and should be administered concomitantly with bronchodilators.
Humidification of inspired gases is important when the endotracheal tube bypasses the normal humidification and filtering systems of the upper airways. Humidification is important to avoid dry mucosal surfaces in patients with inflamed upper airways and humidification can improve expectoration of lower airway secretions. On the other hand, humidification systems must be carefully monitored for contamination with hypophilic organisms such as Pseudomonas and Legionella.
Stress ulcer prophylaxis must be considered in any hospitalized or critically ill patient exposed to metabolic stress; stress ulcer prophylaxis has been shown to decrease the incidence of gastrointestinal bleeding as a standard of care. The hypothesis that increasing gastric pH with stress ulcer prophylaxis increases incidence of nosocomial pneumonia due to aspiration of colonized gastric contents is no longer favored. Commonly used agents include proton pump inhibitors such as pantoprazole or others; H 2 blockers such as famotidine; or barrier agents such as sucralfate.
The use of oral care agents such as topical Mycostatin liquid may be beneficial because it focuses attention on oral care and the potential for bacterial and fungal overgrowth in the oral pharynx of integrated patients who have decreased salivary production and clearance. Also, oral care agents may have an antiseptic effect which decreases bacterial colonization. Patients who develop perioral or intraoral herpetic vesicles should have topical antiviral agents such as acyclovir applied.
Although pulmonary arteriolar vasodilators such as prostacyclin (PGI 2 ) and nitric oxide have been used in the management of severe cases of pneumonia, the use of these costly agents remains controversial and there is variability in their effectiveness. Hypoxic pulmonary vasoconstriction (HPV) is an adaptive response whereby a reflex vasoconstriction in the arteriolar supply to inflamed or diseased alveoli which have low alveolar oxygen concentrations results in minimization of pulmonary shunt fraction. Therefore, pulmonary arteriolar vasodilators which inhibit the HPV response are likely to increase blood supply to diseased hypoxic alveoli and actually increase shunt fraction and systemic hypoxemia.

GENERAL PRINCIPLES OF CHEST RADIOGRAPH INTERPRETATION
The portable anterior-posterior (AP) chest radiograph is the most commonly obtained radiologic study in the ICU. Although the critical care practitioner cannot be expected to render a complete final radiologic reading, it is incumbent on intensivists to be able reliably and independently to render preliminary interpretations which address specific questions regarding patients' clinical conditions. The chest radiograph is also key in the evaluation of a patient's response to therapy. The critical care notes should reflect independent readings as well as later confirmations or variations seen by radiologists. (i) Endotracheal tube: should be midway between the clavicles and the carina within the trachea. It is important to note that the endotracheal tube will move approximately 2.5 cm in or out of the trachea with extension and flexion of the neck, respectively. (ii) Nasogastric and feeding tubes: must be followed within the esophagus into the stomach or small bowel. Tubes that enter the trachea will often pass through one mainstem bronchus. (iii) Tube thoracostomy: mention is usually made of the position of the tube, the intra-or extrathoracic placement of a last collecting hole of the chest tube, and resolution of the liquid or air collection which the tube was intended to treat. (iv) Other hardware: tracheostomy tubes, pacemakers, implantable cardiac defibrillation wires, intra-aortic balloon pump tips, prosthetic cardiac valves, evidence of vascular prostheses and stents, sternal wires, surgical clips, orthopedic prostheses, and airway stents. The mention of other foreign bodies, especially in trauma cases, is particularly important. (c) Bony structures: (i) Fractures and dislocations of the clavicles, sternum, scapula, ribs, and vertebrae.
(ii) Hyperostosis, osteopenia, or potential bony involvement of surrounding tumor or infection. (d) Cardiovascular silhouette: (i) The AP portable radiograph is an unreliable indicator of cardiomegaly.
(iii) The thoracic aorta. Special mention should be made of calcifications within the thoracic aorta and dilatation of the thoracic aorta. (iv) The pulmonary vasculature. Hilar accentuation. Pulmonary arterioles seen end-on are usually apparent and present as small opaque circles in the hilar regions and are normal findings. Evidence of vascular prominence, interstitial prominence, or alveolar flooding consistent with stages of congestive heart failure and pulmonary edema, respectively, must be noted.