Pediatric Pneumonia Workup

Identifying the causative infectious agent is the most valuable step in managing a complicated case of pneumonia. Unfortunately, an etiologic agent can be difficult to identify as various organisms cause pneumonia. Bacterial (including mycoplasmal, chlamydial, and acid fast), viral, and fungal infections are relatively common and have similar presentations, complicating clinical diagnosis. Furthermore, basic laboratory testing and radiologic testing are often not helpful in determining the etiology of pneumonia, and the treatments widely vary.

In patients with complicated pneumonia who have not had a treatment response or who require hospital admission, several diagnostic studies aimed at identifying the infectious culprit are warranted, including cultures, serology, a complete blood cell (CBC) count with the differential, and acute-phase reactant levels (erythrocyte sedimentation rate [ESR], C-reactive protein [CRP]).

Numerous factors may interfere with the ability to grow a likely pathogen in the microbiology laboratory, including (but not limited to) the following: pretreatment with antibiotics that limit in vitro but not in vivo growth, sputum contaminants that overgrow the pathogen, and pathogens that do not replicate in currently available culture systems. Techniques that may help overcome some of these limitations include antigen detection, nucleic acid probes, polymerase chain reaction (PCR)-based assays, or serologic tests.

Although once widely used, tests such as latex agglutination for detection of group B streptococcal antigen in urine, serum, or other fluids have fallen into disfavor because of poor predictive value; however, new generations of nonculture-based technologies continue to undergo development and may be more accurate and widely available in the future.

Testing should include a CBC count with differential and evaluation of acute-phase reactants (ESR, CRP, or both) and sedimentation rate. The total white blood cell (WBC) count and differential may aid in determining if an infection is bacterial or viral, and, together with clinical symptoms, chest radiography, and ESR can be useful in monitoring the course of pneumonia. In cases of pneumococcal pneumonia, the WBC count is often elevated.

Before widespread pneumococcal immunization, Bachur et al observed that approximately 25% of febrile children with a WBC count of more than 20,000/µL, but without lower respiratory tract findings on examination, had radiographic pneumonia (termed occult pneumonia).^[25] Although blood testing was obtained less frequently in the post-Prevnar era, recent studies by the same group demonstrated that leukocytosis was still associated with occult pneumonia.^{[26, 27]}

Sputum is rarely produced in children younger than 10 years, and samples are always contaminated by oral flora. In the cooperative older child with a productive cough, a sputum Gram stain may be obtained (see the image below); however, very few children are able to cooperate with such a test. An adequate sputum culture should contain more than 25 PMN cells per field and fewer than 10 squamous cells per field.

In situations in which a microbiologic diagnosis is essential, endotracheal cultures and/or bronchoalveolar lavage culture can be sent for the isolation of offending pathogens. This is most important in patients with enigmatic and/or severe pneumonia, and it should be considered a priority in patients with compromised immune systems. Routine cultures for respiratory pathogens should be requested, along with special stains for PCP and special stains and cultures for Legionella, fungi, and acid-fast organisms. Viral cultures are also routinely requested.

Although blood cultures are technically easy to obtain and relatively noninvasive and nontraumatic, the results are rarely positive in the presence of pneumonia and even less so in cases of pretreated pneumonia. In a study of 168 patients with known pneumonia, Wubbel et al found only sterile blood cultures.^[28]

In general, blood culture results are positive in 10-15% of patients with pneumococcal pneumonia. The percentage is even less in patients with Staphylococcus infection. However, a blood culture is still recommended in complicated cases of pneumonia. It may be the only way to identify the pathogen and its antimicrobial susceptibility patterns.

In neonates, blood culture with at least 1 mL of blood from an appropriately cleaned and prepared peripheral venous or arterial site is essential, because many neonatal pneumonias are hematogenous in origin and others serve as a focus for secondary seeding of the bloodstream. Older patients should have larger-volume blood culture samples obtained based on their age and the ease with which blood can be obtained. Blood culture samples obtained through freshly placed indwelling vascular catheters may be helpful, but they generally are discouraged because of the possibility of contamination with skin flora. Multiple cultures of blood from different sites and/or those drawn at different times may increase the culture yield.

Because of the relatively low yield of cultures, more efforts are under way to develop quick and accurate serologic tests for common lung pathogens, such as M pneumoniae, Chlamydophila species, and Legionella.

In a Finnish study, of 278 patients diagnosed with community-acquired pneumonia, a total of 24 (9%) confirmed diagnoses of Mycoplasma infection were made, all of which had positive results with IgM-capture test with convalescent-phase serum.^[29] Acute and convalescent serum samples were collected and tested using enzyme immunoassay for M pneumoniae IgM and IgG antibodies. Nasopharyngeal aspirates were tested using PCR and cultured with a Pneumofast kit.

Positive results were confirmed with Southern hybridization of PCR products and an IgM test with solid-phase antigen. Using an IgM-capture test in acute-phase serum, 79% of results were positive, 79% were positive using IgG serology, 50% positive using PCR, and 47% positive using culture.^[29]

The authors of this study concluded that IgM serologic studies for Mycoplasma infection were not only quick but also sensitive and were the most valuable tools for diagnosis of M pneumoniae infection in any age group.^[29] IgM serology is much more sensitive than cold agglutinin assessments, which are more commonly used to aid in the diagnosis of Mycoplasma infection and which demonstrate positive results in only 50% of cases.

Chlamydophila and Legionella species can be grown by experienced microbiologists; however, serologic testing is also routinely performed to support or establish the diagnosis. Similarly, lung infections caused by dimorphic fungi (eg, histoplasmosis) are more commonly diagnosed serologically.

The use of markers of inflammation to support a diagnosis of suspected infection, including pneumonia, remains controversial because results are nonspecific. Various indices derived from differential leukocyte counts have been used most widely for this purpose, although noninfectious causes of such abnormal results are numerous. Many reports have been published regarding infants with proven infection who initially had neutrophil indices within reference ranges.

Quantitative measurements of CRP, procalcitonin, cytokines (eg, interleukin [IL]-6), inter-alpha inhibitor proteins (IaIp),^[30] and batteries of acute-phase reactants have been touted to be more specific but are limited by suboptimal positive predictive value.

Lag time from infection to abnormal values are noted for inflammatory markers; thus, serial measurements are often necessary and do offer a high negative predictive value. However, although these tests may be useful in assessing the resolution of an inflammatory process, including infection, they are not sufficiently precise to establish a diagnosis without additional supporting information. Decisions about antimicrobial therapy should not be based on inflammatory markers alone.

Relatively rapid testing (1-2 d) of viral infections through multiplex PCR is available in many hospitals. PCR is more sensitive than antigen assays, and for some viruses (eg, hMPV), this study may be the only test available. PCR also shows promise of being useful in diagnosing streptococcal pneumonia. In one series of 63 children with empyema, a pathogen could be detected in 53 (84%) by PCR compared to only 24 (35%) by culture of blood and/or pleural fluid. The most frequently detected pathogen was pneumococcus, found in 45 patients.^[31]

PCR is noninvasive, an advantage over lung aspirate or bronchoalveolar lavage (BAL) cultures. Similarly, C pneumoniae infection is diagnosed more readily with PCR than with culture; however, positive test results must correlate with acute symptoms to have any validity, because 2-5% of the population may be asymptomatically infected with C pneumoniae.

PCR testing for TB is also widely available and is helpful in early identification of TB from other mycobacteria in acid-fast cultures.

These tests are used in diagnosing TB. Mantoux skin test (intradermal [ID] inoculation of 5 tuberculin units [TU] of purified protein derivative [PPD]) results should be read 48-72 hours after placement.

Even if the child has received the BCG vaccine, Mantoux test results should be interpreted using the criteria outlined as follows. In children older than 4 years without any risk factors, test results are positive if the induration (not the area of erythema, which may be larger) is 15 mm or larger. Among children younger than 4 years, those who have an increased environmental exposure to TB or other medical risk factors (eg, lymphoma, diabetes mellitus, malnutrition, renal failure), results are positive if the induration is 10 mm or larger. In immunosuppressed children or those in close contact with others who have known or suspected cases of TB, test results are positive if the induration is 5 mm or larger. Chest radiography helps to confirm the diagnosis of a child with positive Mantoux test results (see Chest Radiography).

In a child with suspected pulmonary TB, the cough may be scarce or nonproductive. Therefore, the best test for diagnosis is an early-morning gastric aspirate sent for acid-fast bacilli (AFB) stain, culture, and, if available, PCR. Gastric aspirates should be obtained by first placing a nasogastric (NG) tube the night before sample collection; a sample is aspirated first thing the following morning, before ambulation and feeding. This should be repeated on 3 consecutive mornings.

In the young child or school-aged child with pneumonia, particularly the patient with a gradual onset of symptoms and a prodrome consisting of headache and abdominal symptoms, a bedside cold agglutinins test may help confirm the clinical suspicion of mycoplasmal infection.

This test is easily performed by placing a small amount of blood in a specimen tube containing anticoagulant and inserting this into a cup filled with ice water. After a few minutes in the cold water, the tube is held up to the light, tilted slightly, and slowly rotated. Small clumps of RBCs coating the tube are indicative of a positive test result. Unfortunately, this test is positive in only half the cases of mycoplasmal infection, and it is not very specific.

Although antigen detection assays for S pneumoniae lack a high specificity in children, Neuman and Harper observed that 76% of febrile children with a lobar infiltrate on chest radiograph had a positive rapid urine antigen assay.^[32]

Although antiviral therapies are not often used, performing a nasal wash or nasopharyngeal swab for RSV and influenza enzyme-linked immunoassay (ELISA) and viral culture can help to establish a rapid diagnosis, which may be helpful in excluding other causes. Viral cultures can be obtained in 1-2 days using newer cell culture techniques and may permit discontinuation of unnecessary antibiotics. In addition, correct diagnosis allows for appropriate placement of patients in the hospital. For example, if necessary, 2 infants with RSV infection may share a room, whereas such patients would normally need isolation and may unnecessarily tie up a bed.

Chest radiography is indicated primarily in children with complications such as pleural effusions and in those in whom antibiotic treatment fails to elicit a response. Computed tomography (CT) scanning of the chest and ultrasonography are indicated in children with complications such as pleural effusions and in those in whom antibiotic treatment fails to elicit a response. For more information, see Imaging in Pediatric Pneumonia.

Flexible fiberoptic bronchoscopy is occasionally useful to obtain lower airway secretions for culture or cytology. This procedure is most useful in immunocompromised patients who are believed to be infected with unusual organisms (Pneumocystis, other fungi) or in patients who are severely ill.

The technique of direct rigid bronchoscopy may be used in larger infants; fiberoptic technique is occasionally possible in smaller infants or infants in whom the site is not easily reached using the rigid technique. Both this technique and protected brush tracheal aspirate sampling may not be well tolerated in infants with significant lung disease and poor gas exchange who are very dependent on continuous positive pressure ventilation.

Careful consideration of the diagnostic possibilities is necessary to send the samples for the appropriate tests. Contamination of the bronchoscopic aspirate with upper airway secretions is common; quantitative cultures can help distinguish contamination from infection. Culture and Gram stain of an endotracheal aspirate obtained by aseptic technique as soon as possible after intubation may be useful.

Under typical circumstances, airway commensals take as long as 8 hours to migrate down the trachea. At least one study demonstrated that culture of endotracheal aspirates obtained within 8 hours of birth correlates very well with blood culture results and probably reflects aspirated infected fluid.^[33] The longer the tube has been in place, the greater the likelihood that recovered organisms represent colonizing organisms rather than invasive pathogens; nonetheless, recovery of a single recognized pathogen in large quantities may be helpful in the selection of antibiotic therapy, especially if culture results from normally sterile sites are negative.

The absence of significant inflammatory cells in an endotracheal aspirate or other respiratory specimen suggests that organisms recovered from that site are unlikely to be truly invasive (unless the infant is markedly leukopenic). Thus, the organism represents colonization of the respiratory tract and not infection.

Quantitative culture techniques, such as bronchoscopic alveolar lavage have been assessed in non-neonatal populations and reportedly offer a specificity of more than 80%, depending on the threshold selected (values from >100-100,000 colony-forming units [CFU]/mL have been used).^{[34, 35]} Data from studies of neonates with suspected congenital pneumonia are lacking.

Sites distant from the larger bronchi often cannot be sampled. Specimens may have an increased risk of contamination with oral or airway commensals compared with bronchoscopic sampling but are thought to be more accurate than a conventional endotracheal aspirate.

Nondirected specimens have been obtained through endotracheal tubes 3 mm or greater in internal diameter and intuitively appear to offer decreased probability of contamination. Data from neonates are sparse.^[36] Unlike bronchoscopically obtained specimens, ensuring sampling from a particular involved site is more difficult.

Lung aspiration is underused and is a significantly more efficient method of obtaining a culture. If a prominent infiltrate can be adequately localized in multiple planes, direct aspiration of the infected lung may be performed for culture or biopsy. Lung CT scanning may facilitate such localization.

Lung aspiration is associated with a greater risk of postprocedural air leak and usually requires a larger-bore needle than is used to obtain pleural fluid. Because the risk associated with this procedure is high, lung aspiration is usually reserved for patients who are ill enough to require hospitalization, have not improved with previous empiric treatment, or are immunocompromised and an exact etiology is needed. A lung aspirate should not be performed in patients who are on ventilators, who have a bleeding diathesis, or who are suspected of having an infection with Pneumocystis.

With advances in surgical techniques and increased experience, many clinicians prefer to seek open surgical biopsy or thoracoscopic sampling in such circumstances, especially because success and specimen size are greater and the ability to deal directly with any complication is enhanced.

A study demonstrated positive blood cultures implicated an organism in 18% of patients compared with 52% with positive lung aspirates.^[37] The investigators compared the incidence of positive culture results obtained with blood culture with positive culture results obtained with lung aspiration in 100 children aged 3-58 months with pneumonia. The organisms obtained in the blood and lung aspirate differed in 4 of 8 children in whom both culture results were positive, suggesting that a blood culture may not always accurately reveal the lung pathogen.^[37]

Other studies have demonstrated lung aspirate results to be positive in 50-60% of patients with known pneumonia. In these studies, 1.5-9% of patients had a pneumothorax and 0.7-3% had transient small hemoptysis complicating their lung aspirations.

Although used much less frequently than in previous decades, diagnostic lung puncture may still be useful in circumstances in which pleural and subpleural lung surfaces are visibly involved and can be well localized.^[38] Risk-benefit ratio merits careful consideration given the risk of such complications as pneumothorax, bronchopleural fistula, and hemothorax, as well as sampling a nondiagnostic site. This is a high-risk procedure and should not be considered a routine procedure in the diagnosis or treatment of pneumonia in the neonate.

This procedure is performed for diagnostic and therapeutic purposes in children with pleural effusions (eg, when pleural fluid is impinging on lung or cardiac function) (see the image below).

In the presence of radiographically visible fluid, careful positioning of the infant and thoracentesis after sterile preparation of the sampling site may yield diagnostic findings on Gram stain, direct microscopy, and/or culture. If the Gram stain or the culture result from the pleural fluid is positive or the WBC count is higher than 1000 cells/mL, by definition, the patient has an empyema, which may require drainage for complete resolution. Ultrasonography may reveal smaller fluid pockets and facilitate safer sampling under direct visualization. Although data from neonates are insufficient to draw conclusions, studies in older populations suggest a very high correlation with culture of lung tissue and/or blood. Other therapeutic decisions can be made based on the properties of the effusion (see Complications).

The risk of pneumothorax or laceration of intercostal vessels is real but can be minimized by the use of proper technique, including use of the Z-technique (stretching the skin down over the entry site, so that release after the procedure will permit the return of tissues to their usual location with occlusion of the path of the needle), entry over the superior rib margin (to minimize inadvertent puncture of intercostal vessels) at a dependent site where fluid is most likely to collect, continuous aspiration once the skin is penetrated, and no further advancement once fluid is obtained.

For more information, see Infections of the Lung, Pleura, and Mediastinum.

No specific histologic findings are reported in most patients with pneumonias beyond evidence of inflammation and cellular infiltration and exudation into alveolar spaces and the interstitium. Sputum, lavage, or biopsy material may yield diagnostic findings.

Tissue samples of lung tissue in human infants have typically been obtained from an unrepresentative population. The sample population usually includes only infants with severe pulmonary disease that results in death or threatens to do so or infants who die of other causes and have coincidental sampling of the lung. Consequently, direct observations regarding histologic changes in mild or moderate pneumonia are sparse and are often supplemented by extrapolation from animal disease models, human adults with similar diseases, or more severe cases in human infants that resulted in death or biopsy. Despite these limitations, certain observations in congenital pneumonia recur, whether or not a specific pathogen is implicated.^[39]

Macroscopically, the lung may have diffuse, multifocal, or very localized involvement with visibly increased density and decreased aeration. Frankly hemorrhagic areas and petechiae on pleural and intraparenchymal surfaces are common. Airway and intraparenchymal secretions may range from thin and watery to serosanguineous to frankly purulent, and they are frequently accompanied by small to moderate pleural effusions that display variable concentrations of inflammatory cells, protein, and glucose.

Frank empyema and abscesses are unusual in newborn infants. Particulate meconium or vernix may be visible, especially in the more proximal airways, following aspiration episodes. Superimposed changes, such as air leak, emphysema, and sloughed airway mucosa, may be seen as a consequence of volutrauma, pressure-related injury, oxygen toxicity, and other processes that reflect the vigorous respiratory support often provided to these infants in an attempt to manage derangements of gas exchange caused by the underlying illness.

With conventional microscopy, inflammatory cells are particularly prominent in the alveoli and airways. Mononuclear cells (macrophages, natural killer cells, small lymphocytes) are usually noted early, and granulocytes (eosinophils, neutrophils) typically become more prominent later. Microorganisms of variable viability or particulate debris may be observed within these cells. If systemic neutropenia is present, the number of inflammatory cells may be reduced. Alveoli may be atelectatic from surfactant destruction or dysfunction, partially expanded with proteinaceous debris (often resembling hyaline membranes), or hyperexpanded secondary to partial airway obstruction from inflammatory debris or meconium.

Hemorrhage in the alveoli and in distal airways is frequent. Vascular congestion is common; vasculitis and perivascular hemorrhage are seen less frequently. Inflammatory changes in interstitial tissues are less common in newborns than in older individuals.

Microscopic examination of tissue following immunohistochemical staining or other molecular biologic techniques can identify the herpes virus and an increasing number of other organisms. In patients with TB, acid-fast bacilli are present and can be detected using the Ziehl-Neelsen stain or can be grown on the Lowenstein-Jensen medium. Caseating granulomas are highly suspicious, even in the absence of detectable organisms. Findings of foamy alveolar casts are practically diagnostic for P jirovecii pneumonia, and the cup-shaped organisms are often found using Gomori methenamine silver staining or direct immunofluorescence.

Fungal elements may be seen using Gomori methenamine silver staining or periodic acid-Schiff (PAS) staining. Aspergillus and Zygomycetes species may be seen using simple hematoxylin and eosin (H&E) staining. The specific morphology of the organisms may be diagnostic, but, occasionally, culture or immunostaining is required.