ESSENTIALS OF DIAGNOSIS
Fever or hypothermia, tachypnea, cough with or without sputum, dyspnea, chest discomfort, sweats or rigors (or both).
Bronchial breath sounds or inspiratory crackles on chest auscultation.
Parenchymal opacity on chest radiograph.
Occurs outside of the hospital or within 48 hours of hospital admission in a patient not residing in a long-term care facility.
Community-acquired pneumonia (CAP) is a common disorder, with approximately 4–5 million cases diagnosed each year in the United States, 25% of which require hospitalization. It is the deadliest infectious disease in the United States and the eighth leading cause of death. Mortality in milder cases treated as outpatients is less than 1%. Among patients hospitalized for CAP, in-hospital mortality is approximately 10–12% and 1-year mortality (in those over age 65) is greater than 40%. Risk factors for the development of CAP include advanced age; alcoholism; tobacco use; comorbid medical conditions, especially asthma or COPD; and immunosuppression.
The patient’s history, physical examination, and imaging studies are essential to establishing a diagnosis of CAP. None of these efforts identifies a specific microbiologic cause, however. Sputum examination may be helpful in selected patients but 40% of patients cannot produce an evaluable sputum sample and Gram stain and culture lack sensitivity for the most common causes of pneumonia. Since patient outcomes improve when the initial antibiotic choice is appropriate for the infecting organism, the American Thoracic Society and the Infectious Diseases Society of America recommend empiric treatment based on epidemiologic data (Table 9–9). Such treatment improves initial antibiotic coverage, reduces unnecessary hospitalization, and appears to improve 30-day survival.
Table 9–9.Recommended empiric antibiotics for community-acquired pneumonia. ||Download (.pdf) Table 9–9. Recommended empiric antibiotics for community-acquired pneumonia.
For previously healthy patients who have not taken antibiotics within the past 3 months:
A macrolide (clarithromycin, 500 mg orally twice a day; or azithromycin, 500 mg orally as a first dose and then 250 mg orally daily for 4 days, or 500 mg orally daily for 3 days), or
Doxycycline, 100 mg orally twice a day.
For patients with comorbid medical conditions such as chronic heart, lung, liver, or kidney disease; diabetes mellitus; alcoholism; malignancy; asplenia; immunosuppressant conditions or use of immunosuppressive drugs; or use of antibiotics within the previous 3 months (in which case an alternative from a different antibiotic class should be selected):
A respiratory fluoroquinolone (moxifloxacin, 400 mg orally daily; gemifloxacin, 320 mg orally daily; levofloxacin, 750 mg orally daily) or
A macrolide (as above) plus a beta-lactam (amoxicillin, 1 g orally three times a day; amoxicillin-clavulanate, 2 g orally twice a day are preferred to cefpodoxime, 200 mg orally twice a day; cefuroxime, 500 mg orally twice a day).
In regions with a high rate (> 25%) of infection with high level (MIC ≥ 16 mcg/mL) macrolide-resistant Streptococcus pneumoniae, consider use of alternative agents listed above in (2) for patients with comorbidities.
Inpatient management not requiring intensive care
A respiratory fluoroquinolone. See above for oral therapy. For intravenous therapy, moxifloxacin, 400 mg daily; levofloxacin, 750 mg daily; ciprofloxacin, 400 mg every 8–12 hours, or
A macrolide plus a beta-lactam. See above for oral therapy. For intravenous therapy, ampicillin, 1–2 g every 4–6 hours; cefotaxime, 1–2 g every 4–12 hours; ceftriaxone, 1–2 g every 12–24 hours.
Inpatient management requiring intensive care (all agents administered intravenously, except as noted)
Azithromycin (500 mg orally as a first dose and then 250 mg orally daily for 4 days, or 500 mg orally daily for 3 days) or a respiratory fluoroquinolone plus an anti-pneumococcal beta-lactam (cefotaxime, ceftriaxone, or ampicillin-sulbactam, 1.5–3 g every 6 hours).
For patients allergic to beta-lactam antibiotics, a fluoroquinolone plus aztreonam (1–2 g every 6–12 hours).
For routine patients at risk for Pseudomonas infection:
An anti-pneumococcal, anti-pseudomonal beta-lactam (piperacillin-tazobactam, 3.375–4.5 g every 6 hours; cefepime, 1–2 g twice a day; imipenem, 0.5–1 g every 6–8 hours; meropenem, 1 g every 8 hours) plus azithromycin (500 mg orally as a first dose and then 250 mg orally daily for 4 days, or 500 mg orally daily for 3 days) or a respiratory fluoroquinolone.
For patients who are at risk for Pseudomonas infection AND are critically ill, at increased risk for drug resistance, or if the unit incidence of monotherapy-resistant Pseudomonas is > 10%:
An anti-pneumococcal, anti-pseudomonal beta-lactam (piperacillin-tazobactam, 3.375–4.5 g every 6 hours; cefepime, 1–2 g twice a day; imipenem, 0.5–1 g every 6–8 hours; meropenem, 1 g every 8 hours) plus ciprofloxacin (400 mg every 8–12 hours) or levofloxacin, or
The above beta-lactam plus an aminoglycoside (gentamicin, tobramycin, amikacin, all weight-based dosing administered daily adjusted to appropriate trough levels) plus azithromycin or a respiratory fluoroquinolone.
For patients at risk for methicillin-resistant Staphylococcus aureus infection, add vancomycin (interval dosing based on kidney function to achieve serum trough concentration 15–20 mcg/mL) or linezolid (600 mg twice a day).
Decisions regarding hospitalization and ICU care should be based on prognostic criteria.
Definition & Pathogenesis
CAP is diagnosed outside of the hospital in ambulatory patients who are not residents of nursing homes or other long-term care facilities. It may also be diagnosed in a previously ambulatory patient within 48 hours after admission to the hospital.
Pulmonary defense mechanisms (cough reflex, mucociliary clearance system, immune responses) normally prevent the development of lower respiratory tract infections following aspiration of oropharyngeal secretions containing bacteria or inhalation of infected aerosols. CAP occurs when there is a defect in one or more of these normal defense mechanisms or when a large infectious inoculum or a virulent pathogen overwhelms the immune response.
Prospective studies fail to identify the cause of CAP in 30–60% of cases; two or more causes are identified in up to 26% of cases. Bacteria are more commonly identified than viruses. The most common bacterial pathogen identified in most studies of CAP is S pneumoniae, accounting for approximately two-thirds of bacterial isolates. Other common bacterial pathogens include H influenzae, Mycoplasma pneumoniae, C pneumoniae, S aureus, Neisseria meningitidis, M catarrhalis, Klebsiella pneumoniae, other gram-negative rods, and Legionella species. Common viral causes of CAP include influenza virus, respiratory syncytial virus, adenovirus, and parainfluenza virus. A detailed assessment of epidemiologic risk factors may aid in diagnosing pneumonias due to the following uncommon causes: Chlamydophila psittaci (psittacosis), Coxiella burnetii (Q fever), Francisella tularensis (tularemia), endemic fungi (Blastomyces, Coccidioides, Histoplasma), and sin nombre virus (hantavirus pulmonary syndrome).
Most patients with CAP experience an acute or subacute onset of fever, cough with or without sputum production, and dyspnea. Other common symptoms include sweats, chills, rigors, chest discomfort, pleurisy, hemoptysis, fatigue, myalgias, anorexia, headache, and abdominal pain.
Common physical findings include fever or hypothermia, tachypnea, tachycardia, and arterial oxygen desaturation. Many patients appear acutely ill. Chest examination often reveals inspiratory crackles and bronchial breath sounds(AUDIO 9–8). Dullness to percussion may be observed if lobar consolidation or a parapneumonic pleural effusion is present. The clinical evaluation is less than 50% sensitive compared to chest imaging for the diagnosis of CAP (see Imaging section below). In most patients, therefore, a chest radiograph is essential to the evaluation of suspected CAP.
Diagnostic testing for a specific infectious cause of CAP is not generally indicated in ambulatory patients treated as outpatients because empiric antibiotic therapy is almost always effective in this population. In ambulatory outpatients whose presentation (travel history, exposure) suggests an etiology not covered by standard therapy (eg, Coccidioides) or public health concerns (eg, Mycobacterium tuberculosis, influenza), diagnostic testing is appropriate. Diagnostic testing is recommended in hospitalized CAP patients for multiple reasons: the likelihood of an infectious cause unresponsive to standard therapy is higher in more severe illness, the inpatient setting allows narrowing of antibiotic coverage as specific diagnostic information is available, and the yield of testing is improved in more acutely ill patients.
Diagnostic tests are used to guide initial antibiotic therapy, permit adjustment of empirically chosen therapy to a specific infectious cause or resistance pattern, and facilitate epidemiologic analysis. Three widely available, rapid point-of-care diagnostic tests may guide initial therapy: the sputum Gram stain, urinary antigen tests for S pneumoniae and Legionella species, and rapid antigen detection tests for influenza. Sputum Gram stain is neither sensitive nor specific for S pneumoniae, the most common cause of CAP. The usefulness of a sputum Gram stain lies in broadening initial coverage in patients to be hospitalized for CAP, most commonly to cover S aureus (including community-acquired methicillin-resistant strains, CA-MRSA) or gram-negative rods. Urinary antigen assays for Legionella pneumophilia and S pneumoniae are at least as sensitive and specific as sputum Gram stain and culture. Results are available immediately and are not affected by early initiation of antibiotic therapy. Positive tests may allow narrowing of initial antibiotic coverage. Urinary antigen assay for S pneumoniae should be ordered for patients with leukopenia, asplenia, active alcohol use, chronic severe liver disease, pleural effusion, and those requiring ICU admission. Urinary antigen assay for L pneumophilia should be ordered for patients with active alcohol use, travel within previous 2 weeks, pleural effusion, and those requiring ICU admission. Rapid influenza testing has intermediate sensitivity but high specificity. Positive tests may reduce unnecessary antibacterial use and direct isolation of hospitalized patients.
Rapid turnaround multiplex-polymerase chain reaction (PCR) amplification is clinically available. Different commercial products can identify multiple strains of bacteria and viruses, in addition to genes that encode for antibiotic resistance, with results available in 60–90 minutes. Early experience with multiplex-PCR shows improved overall diagnostic yield, particularly for viral infections, and a higher incidence of bacterial/viral coinfection than previously recognized. Given the lack of effective treatment for most respiratory viral infections, the value of multiplex-PCR may be to avoid antibacterial therapy in viral infections, and early adjustment of empiric antibiotic therapy according to resistance patterns. Limitations of multiplex-PCR include cost and lack of availability, in addition to the challenge of interpreting potentially false-positive results from a highly sensitive test.
Additional microbiologic testing including pre-antibiotic sputum and blood cultures (at least two sets with needle sticks at separate sites) has been standard practice for patients with CAP who require hospitalization. The yield of blood and sputum cultures is low, however; false-positive results are common, and the impact of culture results on patient outcomes is small. As a result, targeted testing based on specific indications is recommended. Culture results are not available prior to initiation of antibiotic therapy. Their role is to allow narrowing of initial empiric antibiotic coverage, adjustment of coverage based on specific antibiotic resistance patterns, to identify unsuspected pathogens not covered by initial therapy, and to provide information for epidemiologic analysis.
Apart from microbiologic testing, hospitalized patients should undergo complete blood count with differential and a chemistry panel (including serum glucose, electrolytes, urea nitrogen, creatinine, bilirubin, and liver enzymes). Hypoxemic patients should have arterial blood gases sampled. Test results help assess severity of illness and guide evaluation and management. HIV testing should be considered in all adult patients, and performed in those with risk factors.
A pulmonary opacity on chest radiography or CT scan is required to establish a diagnosis of CAP. Chest CT scan is more sensitive and specific than chest radiography and may be indicated in selected cases. Radiographic findings range from patchy airspace opacities to lobar consolidation with air bronchograms to diffuse alveolar or interstitial opacities. Additional findings can include pleural effusions and cavitation. Chest imaging cannot identify a specific microbiologic cause of CAP, however. No pattern of radiographic abnormalities is pathognomonic of any infectious cause.
Chest imaging may help assess severity and response to therapy over time. Progression of pulmonary opacities during antibiotic therapy or lack of radiographic improvement over time are poor prognostic signs and also raise concerns about secondary or alternative pulmonary processes. Clearing of pulmonary opacities in patients with CAP can take 6 weeks or longer. Clearance is usually quickest in younger patients, nonsmokers, and those with only single-lobe involvement.
Patients with CAP who have significant pleural fluid collections may require diagnostic thoracentesis (glucose, lactate dehydrogenase [LD], and total protein levels; leukocyte count with differential; pH determination) with pleural fluid Gram stain and culture. Positive pleural cultures indicate the need for tube thoracostomy drainage.
Patients with cavitary opacities should have sputum fungal and mycobacterial cultures.
Sputum induction and fiberoptic bronchoscopy to obtain samples of lower respiratory secretions are indicated in patients who cannot provide expectorated sputum samples or who may have P jirovecii or M tuberculosis pneumonia.
Procalcitonin is a calcitonin precursor released in response to bacterial toxins and inhibited by viral infections. This divergent response to bacterial and viral infections offers laboratory support for a clinical judgment of a viral process in patients with lower respiratory symptoms. Multiple clinical trials have shown that procalcitonin measurement allows clinicians to reduce both initial administration of antibiotics and the duration of antibiotic therapy in CAP without compromising patient outcomes.
The differential diagnosis of lower respiratory tract infection is extensive and includes upper respiratory tract infections, reactive airway diseases, heart failure, cryptogenic organizing pneumonitis, lung cancer, pulmonary vasculitis, pulmonary thromboembolic disease, and atelectasis.
Two general principles guide antibiotic therapy once the diagnosis of CAP is established: prompt initiation of a medication to which the etiologic pathogen is susceptible.
In patients who require specific diagnostic evaluation, sputum and blood culture specimens should be obtained prior to initiation of antibiotics. Since early administration of antibiotics to acutely ill patients is associated with improved outcomes; obtaining other diagnostic specimens or test results should not delay the initial dose of antibiotics.
Optimal antibiotic therapy would be pathogen directed, but a definitive microbiologic diagnosis is rarely available on presentation. A syndromic approach to therapy, based on clinical presentation and chest imaging, does not reliably predict the microbiology of CAP. Therefore, initial antibiotic choices are typically empiric, based on acuity (treatment as an outpatient, inpatient, or in the ICU), patient risk factors for specific pathogens, and local antibiotic resistance patterns (Table 9–9).
Since S pneumoniae remains a common cause of CAP in all patient groups, local prevalence of drug-resistant S pneumoniae significantly affects initial antibiotic choice. Prior treatment with one antibiotic in a pharmacologic class (eg, beta-lactam, macrolide, fluoroquinolone) predisposes the emergence of drug-resistant S pneumoniae, with resistance developing against that class of antibiotics to which the pathogen was previously exposed. Definitions of resistance have shifted based on observations of continued clinical efficacy at achievable serum levels. In CAP, for parenteral penicillin G or oral amoxicillin, susceptible strains have a minimum inhibitory concentration (MIC) 2 mcg/mL or less; intermediate resistance is defined as an MIC between 2 mcg/mL and 4 mcg/mL because treatment failures are uncommon with MIC 4 mcg/mL or less. Macrolide resistance has increased; approximately one-third of S pneumoniae isolates now show in vitro resistance to macrolides. Treatment failures have been reported but remain rare compared to the number of patients treated. Current in vivo efficacy appears to justify maintaining macrolides as first-line therapy except in areas where there is a high prevalence of resistant strains. S pneumoniae resistant to fluoroquinolones is rare in the United States (1% to levofloxacin, 2% to ciprofloxacin) but is increasing.
Community-acquired methicillin-resistant S aureus (CA-MRSA) is genetically and phenotypically different from hospital-acquired MRSA strains. Most produce Panton-Valentine leukocidin, a cytotoxin associated with tissue necrosis. CA-MRSA is a rare cause of necrotizing pneumonia, empyema, respiratory failure, and shock; it appears to be associated with prior influenza infection. Linezolid may be preferred to vancomycin in treatment of CA-MRSA pulmonary infectionbecause linezolid may also act to reduce Panton-Valentine leukocidin toxin production. For expanded discussions of specific antibiotics, see Chapters 30-02 and e1-02.
A. Treatment of Outpatients
See Table 9–9 for specific medication dosages. The most common etiologies of CAP in outpatients who do not require hospitalization are S pneumoniae; M pneumoniae; C pneumoniae; and respiratory viruses, including influenza. For previously healthy patients with no recent (90 days) use of antibiotics, the recommended treatment is a macrolide (clarithromycin or azithromycin) or doxycycline.
In patients at risk for drug resistance (antibiotic therapy within the past 90 days, age greater than 65 years, comorbid illness, immunosuppression, exposure to a child in daycare), the recommended treatment is a macrolide plus a beta-lactam (high-dose amoxicillin and amoxicillin-clavulanate are preferred to cefpodoxime and cefuroxime) or a respiratory fluoroquinolone (moxifloxacin, gemifloxacin, or levofloxacin).
In regions where there is a high incidence of macrolide-resistant S pneumoniae, initial therapy in patients with no comorbidities may include the combination of a beta-lactam added to a macrolide or a respiratory fluoroquinolone.
There are limited data to guide recommendations for duration of treatment. The decision should be influenced by severity of illness, etiology, response to therapy, comorbid medical problems, and complications. Infectious Diseases Society of America/American Thoracic Society guidelines recommend administering a minimum of 5 days of therapy and continuing antibiotics until the patient is afebrile for 48–72 hours. There appears to be no advantage to routinely extending antibiotic therapy beyond 3 days following clinical improvement with defervescence.
B. Treatment of Hospitalized and ICU Patients
The most common etiologies of CAP in patients who require hospitalization but not intensive care are S pneumoniae, M pneumoniae, C pneumoniae, H influenzae, Legionella species, and respiratory viruses. Some patients have aspiration as an immediate precipitant to the CAP without a specific bacterial etiology. First-line therapy in hospitalized patients is the combination of a macrolide (clarithromycin or azithromycin) plus a beta-lactam (cefotaxime, ceftriaxone, or ampicillin) or a respiratory fluoroquinolone (eg, moxifloxacin, gemifloxacin, or levofloxacin) (see Table 9–9).
Almost all patients admitted to a hospital for treatment of CAP receive intravenous antibiotics. However, no studies in hospitalized patients demonstrated superior outcomes with intravenous antibiotics compared with oral antibiotics, provided patients were able to tolerate oral therapy and the medication was well absorbed. Duration of inpatient antibiotic treatment is the same as for outpatients.
The most common etiologies of CAP in patients who require admission to intensive care are S pneumoniae, Legionella species, H influenzae, Enterobacteriaceae species, S aureus, Pseudomonas species, and respiratory viruses. First-line therapy in ICU patients with CAP is an anti-pneumococcal beta-lactam (cefotaxime, ceftriaxone, or ampicillin-sulbactam) combined with either azithromycin or a respiratory fluoroquinolone (moxifloxacin, gemifloxacin, or levofloxacin). In patients with specific risk factors for Pseudomonas infection, combine an anti-pneumococcal, anti-pseudomonal beta-lactam (piperacillin-tazobactam, cefepime, imipenem, meropenem) with either azithromycin or a respiratory fluoroquinolone (moxifloxacin, gemifloxacin, or levofloxacin). In critically ill patients, in those at increased risk for drug resistance, or if the unit incidence of monotherapy-resistant Pseudomonas is greater than 10%, then use two agents with anti-pseudomonal efficacy: either ciprofloxacin or levofloxacin plus the above anti-pneumococcal, anti-pseudomonal beta-lactam or an anti-pneumococcal, anti-pseudomonal beta-lactam plus an aminoglycoside (gentamicin, tobramycin, amikacin) plus either azithromycin or a respiratory fluoroquinolone. Patients with specific risk factors for methicillin-resistant S aureus (MRSA) and those with severe disease (respiratory failure requiring mechanical ventilation or septic shock) also should be treated with vancomycin or linezolid.
Pneumococcal vaccines have the potential to prevent or lessen the severity of pneumococcal infections in immunocompetent patients. Two pneumococcal vaccines for adults are available and approved for use in the United States: one containing capsular polysaccharide antigens of 23 common strains of S pneumoniae in use for many years (Pneumovax 23) and a conjugate vaccine containing 13 common strains approved for adult use in 2011 (Prevnar-13). Current recommendations are for sequential administration of the two vaccines in those aged 65 years or older and in immunocompromised persons, starting with Prevnar-13. Adults with chronic illness that increases the risk of CAP (see Chapter 30-14) should receive the 23-valent vaccine regardless of age. Immunocompromised patients and those at highest risk for fatal pneumococcal infections should receive a single revaccination of the 23-valent vaccine 5 years after the first vaccination regardless of age. Immunocompetent persons 65 years of age or older should receive a second dose of the 23-valent vaccine if the patient first received the vaccine 6 or more years previously and was under 65 years old at the time of first vaccination.
The seasonal influenza vaccine is effective in preventing severe disease due to influenza virus with a resulting positive impact on both primary influenza pneumonia and secondary bacterial pneumonias. The seasonal influenza vaccine is administered annually to persons at risk for complications of influenza infection (aged 65 years or older, residents of long-term care facilities, patients with pulmonary or cardiovascular disorders, patients recently hospitalized with chronic metabolic disorders) as well as health care workers and others who are able to transmit influenza to high-risk patients.
Hospitalized patients who would benefit from pneumococcal and influenza vaccines should be vaccinated during hospitalization. The vaccines can be given simultaneously, and may be administered as soon as the patient has stabilized.
Once a diagnosis of CAP is made, the first management decision is to determine the site of care: Is it safe to treat the patient at home or does he or she require hospital or intensive care admission? There are two widely used clinical prediction rules available to guide admission and triage decisions, the Pneumonia Severity Index (PSI) and the CURB-65.
A. Hospital Admission Decision
The PSI is a validated prediction model that uses 20 items from demographics, medical history, physical examination, laboratory results, and imaging to stratify patients into five risk groups. The PSI is weighted toward discrimination at low predicted mortality. In conjunction with clinical judgment, it facilitates safe decisions to treat CAP in the outpatient setting. An online PSI risk calculator is available at https://www.thecalculator.co/health/Pneumonia-Severity-Index-(PSI)-Calculator-977.html. The CURB-65 assesses five simple, independent predictors of increased mortality (confusion, uremia, respiratory rate, blood pressure, and age greater than 65) to calculate a 30-day predicted mortality (https://www.mdcalc.com/curb-65-score-pneumonia-severity). Compared with the PSI, the simpler CURB-65 is less discriminating at low mortality but excellent at identifying patients with high mortality who may benefit from ICU-level care. A modified version (CRB-65) dispenses with serum blood urea nitrogen and eliminates the need for laboratory testing. Both have the advantage of simplicity: Patients with zero CRB-65 predictors have a low predicted mortality (less than 1%) and usually do not need hospitalization; hospitalization should be considered for those with one or two predictors, since they have an increased risk of death; and urgent hospitalization (with consideration of ICU admission) is required for those with three or four predictors.
B. Intensive Care Unit Admission Decision
Expert opinion has defined major and minor criteria to identify patients at high risk for death. Major criteria are septic shock with need for vasopressor support and respiratory failure with need for mechanical ventilation. Minor criteria are respiratory rate 30 breaths or more per minute, hypoxemia (defined as PaO2/FIO2 250 or less), hypothermia (core temperature less than 36.0°C), hypotension requiring aggressive fluid resuscitation, confusion/disorientation, multilobar pulmonary opacities, leukopenia due to infection with WBC less than 4000/mcL (less than 4.0 × 109/L), thrombocytopenia with platelet count less than 100,000/mcL (less than 100 × 109/L), uremia with blood urea nitrogen 20 mg/dL or more (7.1 mmol/L or more), metabolic acidosis, or elevated lactate level. Either one major criterion or three or more minor criteria of illness severity generally require ICU-level care.
In addition to pneumonia-specific issues, good clinical practice always makes an admission decision in light of the whole patient. Additional factors suggesting need for inpatient hospitalization include the following:
Exacerbations of underlying disease (eg, heart failure) that would benefit from hospitalization.
Other medical or psychosocial needs (such as cognitive dysfunction, psychiatric disease, homelessness, drug abuse, lack of outpatient resources, or poor overall functional status).
Failure of outpatient therapy, including inability to maintain oral intake and medications.
et al; CDC EPIC Study Team. Community-acquired pneumonia requiring hospitalization among U.S. adults. N Engl J Med. 2015 Jul 30;373(5):415–27.
et al. Antibiotic therapy for adults hospitalized with community-acquired pneumonia: a systematic review. JAMA. 2016 Feb 9;315(6):593–602.
et al. Procalcitonin for selecting the antibiotic regimen in outpatients with low-risk community-acquired pneumonia using a rapid point-of-care testing: a single-arm clinical trial. PLoS One. 2017 Apr 20;12(4):e0175634.
et al. Pneumonia severity assessment tools for predicting mortality in patients with healthcare-associated pneumonia: a systematic review and meta-analysis. Respiration. 2017;93(6):441–50.
et al; CAP-START Study Group. Antibiotic treatment strategies for community-acquired pneumonia in adults. N Engl J Med. 2015 Apr 2;372(14):1312–23.
et al. Community-acquired pneumonia. Lancet. 2015 Sep 12;386(9998):1097–108.
et al. Corticosteroids for pneumonia. Cochrane Database Syst Rev. 2017 Dec 13;12:CD007720.
et al. Duration of antibiotic treatment in community-acquired pneumonia: a multicenter randomized clinical trial. JAMA Intern Med. 2016 Sep 1;176(9):1257–65.
et al. Immunogenicity of the 13-valent pneumococcal conjugate vaccine in older adults with and without comorbidities in the Community-Acquired Pneumonia Immunization Trial in Adults (CAPiTA). Clin Infect Dis. 2017 Sep 1;65(5):787–95.
et al. Advances in the causes and management of community acquired pneumonia in adults. BMJ. 2017 Jul 10;358:j2471.