Diagnosis of Pneumonia

Identifying Types of Pneumonia

Although often independently classified, the diagnosis of pneumonia acquired in the community and pneumonia acquired in the hospital share many common features. Clinical suspicion should rest on a constellation of clinical findings, which are often quite sensitive but nonspecific. In community-acquired pneumonia (CAP), patients often present with a variety of symptoms or a pneumonia syndrome, while in the hospital, alteration of laboratory parameters (fever, white blood cell (WBC) count, blood urea nitrogen [BUN], sputum quality and quantity) will often prompt further evaluation. The essential goals are to answer the following clinical questions:

• Does the patient have pneumonia or is the clinical picture due to some other problem?
• What is the microbial etiology of pneumonia?
• Where should the patient be treated?
• What is the best therapeutic approach to the patient’s problem?

Certain combinations of symptoms and signs should prompt the clinician to suspect CAP (Figure 3-1). The most commonly seen symptoms in the younger, previously healthy hosts include:

• Fatigue
• Cough
• Sputum production
• Fever >38°C (100.4°F)
• Chills
• Anorexia
• Dyspnea.

Less commonly seen are:

• Abdominal pain
• Pleuritic chest pain
• Hemoptysis.

Characteristic laboratory alterations are either elevated or decreased (WBC count >12,000 cells/mL or <4,000 cells/mL). The inflammatory marker procalcitonin is elevated in most forms of bacterial pneumonia and may help in diagnosis. In elderly patients, these findings are often attenuated and altered mental status or failure to thrive in the absence of fever or cough may be the only harbingers of pneumonia.

Enlarge 

Typical Pneumonia

Typical pneumonia has been recognized for centuries as acutely producing:

• Fever (60% to 80%)
• Chills and rigors (15% to 50%)
• Chest pain (40%)
• Purulent sputum associated with intractable cough.

This syndromic grouping is observed at initial presentation in 80% of patients, regardless of age group.
S pneumoniae is the most commonly identified organism (>60%) causing this syndrome.

Concurrent and sequential mixed infections involving viral, bacterial and multiple bacteria are likely to occur. Estimates of frequency of mixed polymicrobial infections range from 5% to 40% of cases.

Atypical Pneumonia

The original descriptions of an atypical pneumonia syndrome (Table 3-1) recognized that in somewhat younger patients, presentation is:

• More often subacute
• Associated with nonpulmonary symptoms, such as:
• Myalgia
• Joint pains
• Anorexia.

Atypical pneumonias were initially associated exclusively with M pneumoniae infection, but it is now recognized that a variety of other pathogens (e.g., C pneumoniae, Legionella pneumophila, Coxiella burnetii [Q fever], various respiratory viruses) may be causative. Patients usually seek medical attention within the first week of illness in typical pneumonia and after a longer duration of symptoms in atypical pneumonia. This syndromic approach has been questioned by more recent research, which demonstrated significant overlap in the clinical findings among pathogens traditionally associated with either syndrome; however, pattern recognition can suggest Legionella or Mycoplasma infection. Legionella is a common cause of severe CAP.

Additionally, in patients over 75 years of age, symptoms may be incompletely expressed. Fever and cough may be absent at presentation in 15-40% of these patients. Tachypnea, tachycardia, or altered mentation resulting in delirium or stupor may be the only significant findings (50-75% of patients). Occasionally, in the very old, pneumonia may be confused with primary neurologic disease, such as stroke or intracranial bleeding. In the hospitalized patient, similar symptomatology may be the presenting picture of pneumonia.

Health care–associated pneumonia (HCAP), which includes hospital-acquired pneumonia (HAP) and ventilator-associated pneumonia (VAP), also poses significant diagnostic challenges to the clinician. Since patients are frequently already quite ill, the ability to express symptoms and signs of classic infection will be diminished. The most commonly described presentation of HCAP includes:

• Fever
• Leukocytosis
• Purulent secretions (expectorated or sampled via the endotracheal tube in ventilated patients; findings which are sensitive but not specific).

It is essential to recognize the importance of signs and symptoms individually or as syndromic groupings. A confirmatory evaluation process should be triggered, including (Table 3-2):

• Chest radiograph
• WBC count and differential
• Chemistry profile (renal liver function tests)
• Tests of oxygenation (saturation or arterial blood gas)
• Molecular (PCR-based) testing
• Sputum examination (Gram’s stain, culture)
• Blood cultures
• Pleural fluid evaluation
• Urine studies for Legionella antigen and pneumococcus antigen
• Serology
• Invasive procedures.

The choice among these procedures aids in the determination of whether to hospitalize and where.

Special studies may be necessary when agents associated with bioterrorism (anthrax, plague, tularemia) are suspected.

Enlarge 

Enlarge 

Chest Imaging

The diagnosis of pneumonia may be suspected on clinical grounds but must be confirmed by chest imaging. Because many pulmonary and nonpulmonary conditions mimic the clinical findings of pneumonia, imaging permits confirmation of clinical suspicion (i.e., that it is not an upper respiratory tract infection). Additionally, imaging adds information that enables more accurate prognostication or leads to an alternative differential diagnosis. Imaging modalities include x-ray, computed tomography (CT) scanning and ultrasonography; the relative advantages and drawbacks of each are shown in Table 3-3. It has been estimated that while pneumonia or lower respiratory tract infection may be present with a normal chest x-ray (CXR), this is unusual (<5%). In fact, <2% of pneumonia suspects who initially had a negative chest radiograph had abnormalities on subsequent CXR. CT scanning will often reveal previously undetected infiltrates when pneumonia is suspected (25%). The suspicion of a false-negative chest radiograph should be heightened in the setting of dehydration. Case reports and animal models have suggested that dehydration will mask radiographic findings in pneumonia, which later blossoms as the patient is volume resuscitated.

Among the entities most commonly confused with pneumonia are the following:

• In the outpatient setting:
  • Upper respiratory tract infection
  • Lung cancer (<10%)
  • Acute exacerbations of chronic bronchitis
  • HF
• In hospitalized patients:
  • HF (pulmonary edema)
  • PE
  • Atelectasis
  • ARDS.

In the absence of vital sign abnormalities (tachypnea, tachycardia, or reduced BP), it is unlikely that the chest radiograph will be abnormal. Rarely, systemic vasculitides, such as Wegener’s granulomatosis (WG) or Churg-Strauss syndrome, may confuse clinicians because of systemic symptoms and radiographic shadowing. In fact, many patients presenting with WG are initially thought to have severe CAP. When the chest radiograph is thought to be falsely negative or when the clinical picture is uncertain, CT of the chest may be extremely useful. While the chest roentgenogram should be routinely performed when pneumonia is suspected, the indications for initial CT of the chest are much less clear and should be reserved for complex or confusing clinical presentations or for pneumonia that either does not resolve or progresses while the patient is on appropriate antibiotic treatment. Ultrasonography is a radiation-free alternative that, when performed by experienced personnel, has high sensitivity (94%) and specificity (96%) and may be useful for certain populations (e.g., children and pregnant individuals); however, it is dependent on operator skill and is not widely used.

While the chest radiograph adds to the specificity of pneumonia diagnosis, there is little evidence to support its role in differentiating among the microbial causes of lower respiratory tract infection. Numerous studies have demonstrated significant overlap between radiographic patterns of typical and atypical pneumonia. In fact, a minority of pneumococcal infections result in a lobar radiographic pattern. Both Legionella and Mycoplasma pneumonia, the classic atypical pneumonias, may present with lobar infiltration. In several surveys, radiographic patterns were often no better than a coin toss in identifying specific etiologies, and interobserver reproducibility among radiologists is low. Radiologic examination may be of greater use in the diagnosis of tuberculosis (TB). In VAP, alveolar infiltrates with air bronchograms or progress are highly sensitive for the diagnosis of pneumonia (>85%); specificity in this situation is unknown.

The radiographic pattern is useful in establishing prognosis and in assisting the physician in deciding the location where the patient will be best treated. Lower respiratory tract infection, which is either progressive or multilobar, identifies patients who are at higher risk for a complicated course or death. Bilateral infiltrates and pleural effusions in conjunction with parenchymal evidence of pneumonia are also independent risk factors for poor outcome. These patients usually need hospitalization, and if other high-risk clinical features are present, intensive care (Table 3-4).

Enlarge 

Enlarge 

Blood Tests: Blood Counts and Chemistry

Obtaining a WBC count is recommended when patients are being considered for inpatient therapy. While an elevated leukocyte count with left shift has often been associated with bacterial infection, it has low sensitivity and specificity. Measurement of the WBC count in outpatients should be reserved for cases in which this information will aid in the decision regarding hospitalization. Extremely high or low WBCs and bandemia are associated with poor outcome. Chemistry profiles recommended for hospitalized patients or those with complicated pneumonias in an outpatient setting include the following:

• Renal profile
• Liver function tests
• Electrolytes.

The role of such testing is to detect potential complications of pneumonia, including renal failure, hyponatremia and hepatitis. Elevated BUN has been identified as an independent risk factor for poor outcome.

Procalcitonin, a polypeptide produced in the normal physiological condition by the medullary C-glands of the thyroid, is elevated with bacterial but not viral infection. This characteristic has raised the possibility that it can be used a biomarker to distinguish bacterial from viral pneumonia, and therefore to start or stop antibiotic treatment. Several meta-analyses have shown that procalcitonin use can reduce antibiotic use without impacting outcomes, but more recent studies have questioned significant benefit. Thus, while procalcitonin may be informative in the diagnostic and treatment selection process, by itself it cannot be relied on to guide these decisions. However, procalcitonin is a widely-used indicator of when to stop antimicrobial treatment (see Switch Therapy, Step-Down Therapy and Prevention).

Oxygenation Status

Arterial blood gas testing has for the most part been supplanted by the use of pulse oximetry. While routinely recommended in hospitalized patients, pulse oximetry may aid in the decision to hospitalize. If the arterial PO₂ is <60 mm Hg or saturation <85% on room air, patients should receive oxygen supplementation and be observed as inpatients.

Molecular Tests: Multiplex PCR

Obtaining a microbiological diagnosis is often difficult; a definite identification is obtained in only ~40% of hospitalized CAP cases. However, the increased proliferation and availability of multiplex polymerase chain reaction (mPCR) platforms since the 2010s have made pathogen identification much easier. A key impact of this development has been a reduction in the use of broad-spectrum antibiotics, which is expected to improve antibiotic resistance. Pharyngeal swabs and lower respiratory tract samples are both appropriate for mPCR testing.

Many mPCR platforms are now FDA-approved for pneumonia pathogen testing; such commercially available platforms can detect a broad array of organisms, including Gram-positive and Gram-negative cocci, Gram-negative bacilli, atypical bacteria (C pneumoniae, L pneumophila and M pneumoniae), and many of the most common respiratory viruses (e.g., adenoviruses, coronaviruses, influenza viruses, respiratory syncytial virus [RSV] and others). Furthermore, many panels also test for the presence of antibiotic resistance genes, making them highly informative for therapeutic decision-making. Finally, mPCR platforms have a turnaround time much faster than traditional, culture-based pathogen detection (e.g., 5h for Unyvero HPN and 60-90 minutes for FilmArray).

Sputum Examination

When sputum is expectorated from deep within the lower respiratory tract (i.e., when there is a lack of significant contamination from the upper airway tract), it provides important material for diagnostic evaluation when there is an absence of prior antibiotic therapy (Table 3-5). The former is usually determined by the cytologic criteria of <10 squamous epithelial cells (SECs) and >25 polymorphonuclear (PMN) neutrophils per low-power field. When performed by an experienced examiner, Gram’s staining of the sputum may demonstrate characteristic lancet-shaped diplococci of pneumococcus. Gram’s stain has a sensitivity >50% and a specificity >80% for this diagnosis. Correlation with blood cultures has been high when the specimen is of high quality. However, even studies that demonstrated strong correlation with blood culture isolates found only 2550% of specimens met acceptable cytologic criteria. Perhaps this reflects the increasing frequency with which patients are treated with empiric antibiotics prior to evaluation.

In addition, atypical pathogens, mycobacteria, tuberculosis (TB), mycobacteria (MAI), fungi and viral pathogens will not be detected by traditional Gram’s staining; therefore, the absence of organisms in the presence of significant WBCs on sputum examination should raise clinical suspicions. These organisms make up an increasing proportion of the pathogens encountered in CAP. Given the increasing availability of molecular testing the role of Gram staining in initial diagnostic evaluation remains uncertain and requires an updated cost-benefit analysis. Nevertheless, this test remains a relatively simple and inexpensive means to establish a specific diagnosis, and performance should be based on the patient’s ability to expectorate and the expertise of the examiner.

Sputum culture is rarely recommended in outpatients. Problems with diagnostic accuracy include contamination with upper airway flora that overgrow the true pathogen, resulting in false-positives and false-negatives. Additionally, problems are caused by prior use of antibiotics and inability of the patient to produce sputum derived from the lower airway. In bacteremic S pneumoniae infections, sputum cultures are positive less than half the time.

Enlarge 

Blood and Pleural Fluid

Cultures obtained from sources deep within the body which are normally sterile provide reliable and specific information regarding etiologic diagnosis. Blood cultures are reported to be positive infrequently (5% to 20%) in patients admitted with CAP. They are generally higher in patients with pneumococcal pneumonia (10%), which indicates a worse prognosis. Blood cultures should be obtained prior to starting antibiotics. Although two sets of blood cultures in patients hospitalized with pneumonia are often recommended, the cost-effectiveness of this practice has not been determined and clinical decision-making is rarely influenced by blood culture results. Guideline-directed antimicrobial therapy is designed to cover various differential etiologies since treatment should be initiated before obtaining results and since blood culture yield is generally low. The utility of blood cultures for surveillance purposes is important; however, obtaining cultures should never be sacrificed to timely initiation of antibiotics, especially when pneumonia is life threatening. Most guidelines suggest that blood cultures are valuable in only critically ill patients. (For a complete discussion of blood cultures, see Community-Acquired Pneumonia).

In addition to blood cultures, significant pleural effusions should be drained and examined by Gram’s stain and culture. Determination of pH and leukocyte count of pleural fluid may aid in the decision regarding the type of drainage:

• Thoracentesis
• Chest tube
• Surgical.

Serology

Routine serologic testing is not advised as the cost-benefit ratio is likely to be high. Paired samples for Legionella spp, Mycoplasma, or Chlamydia spp require several weeks to increase (acute to convalescent levels) long after treatment decisions will have been made. For example, immunoglobulin M (IgM) antibodies to C pneumoniae take approximately 3 weeks to develop, while IgG may require 2 months. In Legionella, a titer >1:256 has been considered presumptive evidence for acute infection but is present in <15% of documented infections. Serologic studies for viruses are also thought to be of little practical use. Cold agglutinin titers >1:64 have a sensitivity of 30% to 60% for Mycoplasma and require 1 week for positivity. Thus serologic investigation aids in epidemiologic surveillance but adds little to clinical decision making.

Rapid diagnostic testing has increasingly relied on antigen detection. These techniques vary widely and include:

• Counterimmunoelectrophoresis
• Latex agglutination
• Immunofluorescence
• Enzyme-linked immunofluorescence.

Urinary antigen detection for L pneumophila (serogroup 1) and S pneumoniae has excellent sensitivity and specificity but may require 5 days to become positive. These tests are suggested for moderate to severe CAP. DNA amplification techniques offer promise for detection of a variety of respiratory pathogens, including pneumococcus, Mycoplasma, Legionella, Chlamydia, MAI and respiratory viruses. Although highly sensitive, false-positive results remain problematic at the present time.

Bronchoscopy and Related Procedures

To avoid the problems inherent with contamination of sputum specimens by upper airway flora, techniques have been developed to bypass this anatomic region and gain access to the lower respiratory tract. Transtracheal aspiration was previously used to achieve this goal but is now infrequently employed due to poor specificity and high complication rates. Bronchoscopy also can be used to obtain specimens from the lower airway but is also plagued by the problem of upper airway contamination except when TB, fungal, or Pneumocystis infection is suspected. The use of either protected specimen brush (PSBs) or protected bronchoalveolar lavage (PBAL) allows the sample to be obtained directly from the lower airway. These techniques are employed in critically ill patients. In VAP, recovery of bacteria in high concentrations (>1,000 colonies/mL for PSB or >10,000/mL for PBAL) allows differentiation of noninfectious from infectious causes of pulmonary infiltrates.

Bronchoscopy is not recommended for routine use because of increased expense and potential complications. The use of endotracheal aspirates performed by nonphysician personnel has achieved similar degrees of accuracy at lower cost. Failure to grow bacteria in an endotracheal aspiration has a high (>90%) negative predictive value. However, the yield of quantitative culture is significantly compromised when patients have received prior antibiotics. The use of fine-needle aspiration for the diagnosis of pneumonia has been demonstrated to have diagnostic yields of 40-80% but carries a significant risk of pneumothorax and hemorrhage. To date, there are no studies that indicate outcome (morbidity, mortality, or quality of life) is improved when invasive techniques are employed compared with empiric therapy. Bronchoscopy is of greatest value in pneumonias that progress or fail to resolve after initial empiric therapy (Table 3-6).

Enlarge 

Hospitalization

The majority of pneumonia patients can be treated as outpatients (>95%). Hospitalization is mandatory in patients who are at high risk for morbidity and mortality. Additionally, social factors, including adequacy of home care, mental competence and likelihood of compliance with therapeutic regimens, must also enter into the decision about the site of treatment. Among the most important determinants of poor outcome are:

• Age
• Altered mental status
• Abnormal hemodynamic status
• High-risk pathogens (e.g., Pseudomonas, MRSA)
• Cancer.

Many different strategies have been developed to assess severity and mortality risk among pneumonia patients in order to help decide site of care and appropriate therapeutics. The Pneumonia Patient Outcomes Research Team developed a stratification system, the Pneumonia Severity Index (PSI; see Figure 4-1 in Community-Acquired Pneumonia), which was validated in approximately 40,000 patients. It relies upon a composite scoring algorithm and is meant to be calculated upon the patient’s admission to the hospital. Although somewhat cumbersome to score, when used appropriately, it can be a very useful and accurate decision support tool. It also has value to guide thinking in broader terms about where patients are best cared for by aggregating the constituent demographic factors and signs and symptoms. It is also used to stratify patients in clinical trials. Patients were divided into categories based on:

• Age
• Comorbid illness
• Physical findings
• Laboratory findings.

Those falling into category 1 or 2 (mortality <1%) were safely treated as outpatients, while category 3 (mortality 3%) required short periods of inpatient observation. Those in category 4 (mortality 8%) and category 5 (mortality 30%) required inpatient management, the latter in intensive care. This algorithm has been endorsed by the guidelines of the Infectious Disease Society of America (IDSA) and the American College of Emergency Medicine. Implementation of these guidelines has significantly increased the recognition of low-risk patients who can be safely treated at home.

Another severity scoring strategy is called CURB-65 (confusion, urea, respiratory rate, blood pressure, age >65). Its components are among the best-validated criteria for severity and are as follows:

• Respiratory rate >30
• Diastolic BP <60 mm Hg
• BUN >20
• Confusion
• Age >65 years.

Enlarge