Etiology

Because of the difficulty of documenting the microbiology of pneumonia in infants and young children, accurate data concerning the incidence and specific agents of bacterial and viral pneumonia in children have been lacking. While viruses have historically been identified as the most common causes of CAP, etiologic studies conducted between the 1970s and 1990s indicated that two bacterial pathogens were responsible for the majority of fatal cases of childhood pneumonia— Streptococcus pneumoniae and Haemophilus influenzae type b. In the intervening years, effective protein conjugate vaccines (PCVs) that protect against H. influenzae type b and S. pneumoniae have been introduced into the vaccine programs of many countries, including the United States, profoundly decreasing the incidence of documented bacterial pneumonia caused by these pathogens.

Recent prospective, population-based surveillance (the Centers for Disease Control and Prevention [CDC]’s Etiology of Pneumonia in the Community [EPIC] study) that included molecular diagnostic methods in addition to culture and serology, has provided invaluable data on CAP for hospitalized children. Children younger than 18 years were eligible if (1) they were diagnosed with CAP between January 1, 2010, and June 30, 2012; (2) resided in 1 of 22 counties, which comprised the study catchment areas of participating children’s hospitals in Nashville and Memphis, Tennessee, and Salt Lake City, Utah; and (3) had a chest radiograph performed within 72 hours before or after admission.

Overall 2638 children were enrolled in the EPIC study; the final cohort included the 2358 (89%) children with radiographic evidence of pneumonia. This study had several major findings. First, a pathogen was detected in 81% of children. Viruses alone were identified in two-thirds (66%) of children overall and in half (50%) of children older than 5 years. Second, the most commonly detected pathogens were respiratory syncytial virus (RSV; 28%), rhinovirus (27%), human metapneumovirus (13%), and adenovirus (11%); adenovirus accounted for 15% of infections in younger children but only 3% of infections in children older than 5 years. Third, typical bacteria accounted for 8% of CAP overall ( S. pneumoniae accounted for 5% of cases, Staphylococcus aureus and Streptococcus pyogenes accounted for 1% each). Fourth, atypical bacteria were quite common in older children but uncommon in younger children; Mycoplasma pneumoniae was detected in 19% of children older than 5 years versus 3% of young children.

Children 4 years and younger had a virus as the sole pathogen in nearly half the cases and viral-viral co-detection in an additional 15% to 25% of cases depending on age subgroup. Bacterial infections, either viral-bacterial coinfections or bacterial infections alone, accounted for less than 15% of infections in children younger than 4 years. In children younger than 2 years, the most common pathogens were RSV (42%), rhinovirus (29%), adenovirus (18%), and human metapneumovirus (14%), whereas influenza, parainfluenza, and coronaviruses each accounted for 6% to 7% of infections. In children 2 to 4 years of age, the most common pathogens were RSV (29%), rhinovirus (25%), human metapneumovirus (17%), and adenovirus (9%); Mycoplasma pneumoniae , coronaviruses, influenza, and parainfluenza were each also detected in 5% to 8% of children. Bacterial infections occurred in 5% to 6% of children younger than 4 years with CAP: Streptococcus pneumoniae (3% to 4%), S. aureus (1%), and group A Streptococcus (1%). H. influenzae type b was only rarely identified, although the majority of children were immunized.

Rhinovirus and M. pneumoniae were more prevalent among school-aged children. Rhinovirus was identified in 30% and 19% of children 5 to 9 years and 10 to 17 years, respectively. M. pneumoniae was identified in 16% and 23% of children 5 to 9 years and 10 to 17 years, respectively. Bacterial causes in these age groups included Streptococcus pneumoniae (3% to 4%), S. aureus (1%), and group A Streptococcus (<1%). Other studies of bacterial pathogens recovered through culture or polymerase chain reaction (PCR) of blood or pleural fluid have also identified S. pneumoniae as the most important bacterial cause of CAP in children, particularly CAP complicated by empyema.

Internationally, the Bill and Melinda Gates Foundation is sponsoring the Pneumonia Etiology Research for Child Health (PERCH) project. This case-control study involves seven countries in two geographic regions, Africa and Asia, where most cases of fatal childhood pneumonia occur. PERCH sites include Bangladesh, Gambia, Kenya, Mali, South Africa, Thailand, and Zambia. The study had evaluated children aged 1 to 59 months and uses conventional and molecular methods to identify an etiology for pneumonia. Preliminary data from Kenya demonstrate an etiology in more than 75% of cases, with bacteria identified in 9% (most commonly S. pneumoniae ), viruses in 53%, and mixed viral and bacterial infection in 15%. An additional pilot study conducted in New Caledonia as part of the PERCH trial identified a pathogen in 89% of 108 hospitalized cases, with viruses representing more than 80% of the pathogens detected, similar to the findings of the EPIC study in the United States.

Epidemiology

In the developed world, CAP remains a common cause of pediatric hospitalization and ambulatory medical visits. Age is an important determinant of disease rates, with younger children more likely to have CAP and be hospitalized. In the United States, hospitalization rates published for CAP in 2010 ranged from a high of 1169/100,000 for infants younger than 1 year to a low of 44.7/100,000 for children 13 to 18 years of age. Children 1 to 5 years had rates of 383/100,000 and those 6 to 12 years had rates of 69/100,000. Regional data from 2010 to 2012 focused on different age groups but yielded generally comparable estimates: younger than 2 years, 620/100,000; 2 to 4 years, 238/100,000; 5 to 9 years, 101/100,000; and 10 to 17 years, 42/100,000. In the outpatient setting, there are more than 1.5 million visits each year; age-specific estimates for outpatient visits (ambulatory practices and emergency departments) in 2006 and 2007 were as follows: 1 to 5 years, 48.8/1000; 6 to 10 years, 18.4/1000; and 11 to 18 years, 8.7/1000. Race is also an important factor in CAP. Black children have higher rates of hospitalization for CAP in contrast to white children; however, the difference has decreased since the introduction of pneumococcal conjugate vaccines. Immune status is also relevant when identifying risk for CAP. Children with congenital and acquired immune deficiency, including human immunodeficiency virus (HIV) infection, are at increased risk for lower respiratory tract infection caused by both common and uncommon respiratory pathogens, which are described in more detail in the chapters on immunocompromised hosts. Additionally children who are exposed to HIV in utero but are uninfected (HIV-exposed, uninfected) are also at increased risk of infectious illnesses and mortality, including pneumonia-related treatment failure, compared with unexposed children.

The epidemiology of CAP in the United States changed significantly after the licensure of the H. influenzae type b and pneumococcal conjugate vaccines. Overall, rates of invasive pneumonia have decreased in children younger than 5 years, the age group targeted for pneumococcal conjugate vaccines. Decreases also have been observed in adults and suggest a likely benefit from herd immunity.

Pathogenesis

The pathogenesis and host immune response differs significantly among viral, bacterial, and mycoplasma pathogens. Detailed information regarding pathogen-specific disease can be found in chapters describing each pathogen.

For viral pneumonia, the pathologic processes associated with infection by RSV, adenoviruses, and influenza virus A have been studied most extensively in humans, aided by animal models for each pathogen. For most viral lower respiratory tract pathogens, inoculation of the child starts with the upper respiratory tract after exposure to infectious secretions from others through sneezing or coughing or from fomites. Once in the lower respiratory tract, viral agents proliferate and spread by contiguity to involve the lower and more distal portions of the respiratory tract. Infected epithelium loses ciliary function, rounds up as a result of injury, and sloughs into the air passages, with subsequent stasis of mucus and accumulation of cellular and inflammatory debris. When infection extends to the terminal airways, alveolar lining cells themselves will lose structural integrity and function. As a result, surfactant production may be lost, hyaline membranes form, and pulmonary edema may develop. Mononuclear cells infiltrate submucosal and interstitial structures, further contributing to tissue edema, narrowing of air passages, and alveolar-capillary block of gas exchange. With certain pathogens, particularly RSV, bronchiolar spasm in the context of accumulating airway debris gives rise to ball-valve–mediated air trapping with each inspiration, resulting in hyperinflation. Complete obstruction of an airway by debris may result in distal atelectasis. Ventilation-perfusion mismatch may compound hypoxemia that results from alveolar edema.

Histopathologic descriptions of the viral pathology can include acute bronchiolitis, necrotizing bronchiolitis, interstitial pneumonia, alveolar pneumonia, and hemorrhagic bronchopneumonia. Acute bronchiolitis is characterized by relatively superficial and reversible destruction of ciliated respiratory epithelium, along with accompanying mononuclear infiltration in submucosal tissues. Necrotizing bronchiolitis extends to the deeper submucosal layers lining the respiratory tract and may not be as readily reversible. This destructive pathology is associated particularly with adenoviral pneumonia.

Interstitial pneumonia is a diffuse process in which the inflammatory mononuclear response predominantly involves the peribronchial alveolar septa. In alveolar pneumonia, the alveoli are filled with degenerating lining cells and mononuclear or polymorphonuclear inflammatory cells with or without hyaline membranes and may reflect bacterial superinfection, acute respiratory distress syndrome, or nonspecific changes associated with mechanical ventilation and oxygen toxicity. Hemorrhagic bronchopneumonia has been described in a fatal pediatric case of hMPV lower respiratory tract infection. When hyaline membranes are present, the process is described as diffuse alveolar damage, the histopathologic hallmark of acute respiratory distress syndrome. Acute bronchiolitis and interstitial pneumonia have classically been observed in most cases of fatal viral pneumonia, regardless of the infecting pathogen.

Both viral replication and age-specific immune responses contribute to the severity of RSV disease in infants. It appears that the severity of RSV in infants relates to a vigorous innate immune response and inadequate adaptive immune response to viral replication and destruction of the respiratory tract epithelium. RSV infection initiates tremendous innate inflammation through both chemokines and cytokines after stimulation of airway epithelial cells and phagocytic cells.

Three important factors that influence the pathologic expression of nonbacterial pneumonia in children are anatomy, preexisting pulmonary disease, and immunity. In young infants, the small caliber of the terminal airways and the absence of interconnections between alveolar spaces (pores of Kohn) contribute to the development of wheezing and lobular atelectasis. Preexisting pulmonary disease (e.g., bronchopulmonary dysplasia) and an inability to clear the excessive secretions triggered by infection in patients with bronchopulmonary dysplasia also may lead to bronchospasm, atelectasis, and respiratory failure.

The most important factors in the pathogenesis of bacterial pneumonias are the virulence of the pathogen, the absence of specific humoral immunity, and the presence of a preceding viral respiratory tract infection. Most bacterial pneumonias are a result of colonization of the nasopharynx, followed by either bacteremia or aspiration of organisms. The lung is protected from bacterial infection by a variety of mechanisms, including entrapment and removal of organisms deposited on airways by ciliated epithelium and mucus, ingestion and killing of airway bacteria by alveolar macrophages, by reticuloendothelial system function for organisms gaining access to the lung via the bloodstream, by neutralization of invading bacteria by local and systemic nonspecific innate and specific immune substances (i.e., specific antibody, complement, opsonins), and by removal of invading organisms from the lung by lymphatic drainage. Pulmonary infection may occur when one or more of these barriers are altered, inhibited, or overwhelmed.

Animal models document that the inflammatory responses in the lung may be caused by various bacterial cell wall components, including peptidoglycans, lipoteichoic acid (of gram-positive organisms), and lipopolysaccharides (endotoxin) of gram-negative bacteria. Bacterial components stimulate a profound innate inflammatory response initiated in part by activation of pattern-recognition receptors (most notably Toll-like receptors [TLRs]), followed by an intense inflammatory cascade mediated by cytokines and chemokines that results in upregulation of cell surface adhesion molecules in addition to multiple intracellular pathways.

As classically described, pneumonia caused by S. pneumoniae begins with acute inflammation and hyperemia of the lower respiratory tract mucosa, exudation of edema fluid, deposition of fibrin, and infiltration of alveoli by polymorphonuclear leukocytes (i.e., “red hepatization”), followed by predominance of fibrin deposition and macrophage activity (i.e., “white hepatization”). Resolution subsequently occurs with absorption of exudates and return of lung morphology and physiology to normal. In contrast, when pneumonia is caused by organisms capable of inciting even more profound tissue inflammation (e.g., S. aureus , gram-negative enteric bacilli or pseudomonads), destruction of tissue and formation of abscesses frequently occur.

Recent animal data provide a scientific basis for the observation that signs and symptoms of a viral respiratory tract infection frequently precede development of bacterial pneumonia. The respiratory virus may act by simultaneous destruction of respiratory epithelium, cytokine-mediated exuberant local inflammatory responses, and upregulation of functional bacterial cell surface antigen receptor molecules. Staphylococcal or pneumococcal pneumonia may occur during or shortly after infection caused by influenza virus. Severe pneumococcal pneumonia has been associated with outbreaks of influenza, and increased mortality rates from pneumonia occur during epidemics of influenza, as documented from new postmortem studies of fatalities during the influenza pandemic of 1918. As indirect proof of the role of preceding viral infection, the use of conjugate pneumococcal vaccines appears to have decreased some of the lower respiratory tract morbidity seen with influenza infection. Respiratory viruses other than influenza have also been associated with bacterial pneumonia, including RSV, parainfluenza, rhinovirus, hMPV, and adenovirus. The pneumococcal conjugate vaccines have reduced chest radiograph–defined alveolar consolidation associated with RSV by 12%, with parainfluenza types 1 to 3 by 44%, and with hMPV by 40%, suggesting that concurrent pneumococcal infection was frequent in virus-associated pneumonias. By documenting the decrease of bacterial pneumonia in immunized patients, the conjugate pneumococcal vaccine has been valuable in identifying the role of viral and pneumococcal coinfection. Immune compromise, anatomic defects, and neurologic comorbidities may predispose children to recurrent lower respiratory tract infection, but a detailed discussion of these risk factors is beyond the scope of this chapter.

Mycoplasma is primarily a mucosal pathogen, attaching extracellularly to the epithelial cells of the host airways. Attachment leads to the generation of an inflammatory response with production of cytokines with subsequent recruitment of lymphocytes and neutrophils to the airway mucosa, leading to the production of inflammatory infiltrates within the airways. Mycoplasma attaches to airway cells by means of unique attachment organelles, resulting in cytopathic effects on ciliated epithelium of bronchi and bronchioles that can include loss of cilia, vacuolization, loss of metabolic function, and ultimately cell death. Edema with bronchiolar and alveolar inflammatory infiltrates containing macrophages, lymphocytes, neutrophils, and plasma cells has been described. Diffuse alveolar damage may also occur.

The nonpermanent immunity that develops with natural infection provides a vigorous cell-mediated immune response that, on reexposure to Mycoplasma, may lead to more severe clinical illness and pulmonary injury. This concept of immune-mediated lung disease provides the basis for increasing pulmonary disease severity with increasing age, which correlates with observed, specific cell-mediated immunity that also increases with age. However, one study documented a similar incidence of serologic diagnosis of Mycoplasma pneumoniae infection in hospitalized preschool- and school-aged children with radiographic evidence of pneumonia. Although this finding suggests that M. pneumoniae may truly cause lower respiratory tract infection in all age groups, preschool children had clinical cure rates at 10 days into β-lactam therapy that were equivalent to cure rates in school-aged children treated with a macrolide or fluoroquinolone, raising additional questions of the natural history of mycoplasma disease in preschool children.

Clinical Manifestations

The clinical presentation of a child with pneumonia is most often based on symptoms directly related to decreased oxygenation of blood (hypoxemia), as well as symptoms directly related to lung inflammation, characteristically caused by a pathogen in a particular age group (from very young infants to adolescents). A benign viral lower respiratory tract infection caused by rhinovirus may produce only a cough, with no tachypnea or clinical toxicity. In contrast, a bacterial pneumonia caused by MRSA may result in tachypnea, dyspnea with grunting respirations and retractions, cyanosis, high fever, hypotension, and altered mental status. Detailed, pathogen-specific clinical presentations are provided within the chapters discussing each pathogen.

Of importance is the recent recognition of the high frequency of coinfections caused by both a virus and bacteria or caused by multiple viruses. In these situations, the clinical presentation and progression of disease may be an amalgamation of two or more pathogens. The association of influenza virus infection with subsequent bacterial pneumonia is perhaps the best studied. In these situations, the first signs of infection are those of viral disease, with coryza and sore throat. The progression of signs and symptoms and development of secondary fever may herald bacterial superinfection, which can be associated with consolidative pneumonia or necrotizing pneumonia that is characteristic of S. pneumoniae or S. aureus .

Clinical manifestations of hypoxemia in children may occur after infection by either viruses or bacteria. Although these clinical presentations have been best studied in the developing world, the pathogens responsible for lower respiratory tract infection, particularly bacterial pathogens, are seldom identified in these studies, making pathogen-specific descriptions of disease presentation or clinical course difficult. Regardless of pathogen, however, those with more severe hypoxemia will have various degrees of respiratory distress ( Box 22.1 ).

Whereas the older child and adolescent can verbalize the sensation of dyspnea, the younger child or infant may present with hypoxemia with increased respiratory effort and decreased level of attentiveness, decreased consolability, poor color, and decreased spontaneous movement, which has been best described in young infants with bronchiolitis. In resource-poor areas of the world, cyanosis has been documented to be associated with severe hypoxia. A systematic review of published studies, primarily in the developing world, found that cyanosis and use of accessory muscles associated with the head tilting downward with each inspiration had a higher specificity for predicting hypoxemia in children than other signs but were not sufficiently sensitive to be able to be used in the clinical diagnosis of hypoxemia.

Tachypnea is a nonspecific clinical sign that may be a result of fever, anxiety, metabolic acidosis, or the hypoxemia that accompanies respiratory failure. Whereas in the developing world rapid breathing as perceived by the mother, as well as chest retractions, nasal flaring, and crepitations, were statistically associated with hypoxemia, the sensitivity and specificity of tachypnea in diagnosing hypoxemia were both only about 70% to 75% in contrast to that of pulse oximetry. Likewise among children younger than 5 years of age undergoing chest radiography for possible pneumonia at a pediatric emergency department in Boston, the respiratory rates for those with documented pneumonia by chest radiograph were not statistically higher than those whose chest radiographs were normal. However, in the developing world, a high, age-specific respiratory rate (tachypnea), particularly when measured at 24 hours into antimicrobial therapy (amoxicillin or penicillin G), has been linked to clinical treatment failure in children with severe pneumonia.

In the developed world, realizing that clinical presentation of pneumonia is neither sensitive nor specific for the diagnosis, pulse oximetry is widely used as the basis for management decisions, to confirm clinical suspicions of pneumonia with hypoxemia. Pulse oximetry technology is widely available, accurate in most settings, simple to perform, and has the ability to quantitate the severity of lung disease. Hypoxemia is a risk factor for poor outcome in both adults and children with lower respiratory tract infection. Although pulse oximetry is used together with clinical assessment, recommendations have been made to hospitalize a child whose Sp o ₂ is less than 90% in room air, although prospectively collected data to assess the risks and benefits of this particular value have not been studied in the developed world. Given the lack of scientific data to support specific Sp o ₂ -based recommendations, it is understandable that some practice variation exists.

Viral pathogens are far more commonly associated with lower respiratory tract infection, particularly for preschool-aged children. The clinical presentation generally follows a short upper respiratory tract infection prodrome (coryza, pharyngitis, mild fever) by a few days, with a gradual onset of cough, usually nonproductive. Clinical toxicity is usually mild for most viral pathogens. Some degree of bronchiolitis may occur with lower respiratory tract infection, leading to wheezing with difficult air entry and prolonged expiration with air trapping within the distal airways (primarily caused by RSV, hMPV, and rhinovirus). Lower respiratory tract disease is most often bilateral, affecting all lobes, in contrast to the more common presentation of focal, unilateral pneumonia caused by bacteria. Disease is usually benign, lasting only 3 to 5 days and resolving spontaneously. With young infants under 1 year of age, particularly those infected by RSV, increasing lung parenchymal inflammation can lead to decreasing oxygenation and respiratory failure. Hypoxemia and significant reactive airway disease may occur, not uncommonly leading to hospitalization. For the most common viral pathogens responsible for clinically significant lower respiratory tract disease, respiratory failure that leads to intubation is extremely unusual but may occur with influenza A or B viruses, particularly in children with comorbid medical conditions. Very uncommon viral pathogens, including SARS coronavirus and influenza H5N1 (avian influenza) may cause rapidly progressive respiratory failure with a high mortality rate.

In contrast, bacterial pneumonia may have a more rapid progression from cough to dyspnea and is more often associated with signs and symptoms of systemic toxicity, including high fever, hypotension, myalgias, malaise, headache, gastrointestinal complaints, restlessness, apprehension, and, on occasion, manifestations of secondary sites of infection. As with viral disease, signs and symptoms of bacterial pneumonia vary with the bacterial pathogen and the age of the child. Pleural fluid collections adjacent to infected lung may be extensive and can become infected, resulting in bacterial empyema that is a well-recognized complication of bacterial CAP ( Fig. 22.1 ). Some bacteria are associated with a specific pattern of disease, such as the lobar pneumonia of S. pneumoniae ( Fig. 22.2 ) and the empyema, abscess ( Fig. 22.3 ), necrotizing pneumonia, and pneumatocele formation caused by S. aureus ( Fig. 22.4 ); however, any of these manifestations may result from infection caused by any of the bacterial pathogens. In general, children with bacterial pneumonia have cough, tachypnea, and, with progressive disease, symptoms that are characteristic of hypoxemia, including dyspnea, shallow or grunting respirations, and flaring of the alae nasae. For lower lobe pneumonia, inflammation may be associated with gastrointestinal symptoms, including abdominal pain that may mimic that of appendicitis. Signs of pneumonia may be subtle in young infants. Percussion usually is not valuable in an infant or older child if distribution of the pneumonia is patchy. Dullness to percussion is associated more often in young children with the presence of pleural fluid than with the involvement of the parenchyma of the lung. Auscultatory findings classically include rales, or, for the child with consolidation, “tubular” breath sounds not associated with rales, lacking the normal breath sounds associated with gradual aeration and expansion of the lung with each inspiration. Abnormal findings in older children include dullness to percussion, decreased tactile and vocal fremitus on palpation, and the presence of egophony. Intercostal retraction indicates recruitment of accessory muscles, often necessary to assist respiration when hypoxemia is present.

Computed tomography scan of a 1-year-old boy with influenza complicated by an empyema caused by methicillin-resistant Staphylococcus aureus . (A) A large right-sided pleural effusion with compression of the adjacent lung is seen, as is pleural enhancement with areas of septation and cavitation (B).

Chest radiograph of a 2-year-old boy with pneumococcal pneumonia and bacteremia demonstrates consolidation in the right lower lobe with obliteration of the right hemidiaphragm.

Radiologic images of a 1-year-old boy with a lung abscess caused by methicillin-susceptible Staphylococcus aureus . (A) Chest radiograph reveals a cavitary right lower lobe mass consistent with lung abscess. (B) Chest computed tomography shows a rounded cavitary lesion that measures approximately 5 × 5 cm. It has minimum peripheral enhancement.

Chest radiograph of a 6-year-old boy with multilobar pneumonia caused by methicillin-resistant Staphylococcus aureus reveals right upper and left lower lobe opacities with multiple areas of air-filled cavities or pneumatoceles. There is a small right pleural effusion. A chest tube appears in the left lung. The clinical picture was consistent with a necrotizing pneumonia.

Irritation of the pleura and accumulation of pleural fluid is accompanied by chest pain that may be severe and may limit chest movement. As the effusion enlarges, dyspnea may increase, but pleuritic pain may diminish and become a dull ache. The pain of pleural irritation may be present at the site of inflammation or may be referred. Pleural irritation over the right upper lobe may elicit meningismus, a sign of meningeal irritation.

Atypical pneumonia in children is most often caused by Mycoplasma pneumoniae ; disease caused by either Legionella pneumophila or Chlamydophila pneumoniae is quite rare. Unfortunately, no prospective natural history studies that involve routine screening of both outpatients and inpatients for Mycoplasma spp. infection in all pediatric age groups have been performed. Most studies of CAP in children that collected clinical information on Mycoplasma were retrospective, including some reporting outbreaks. Some reported data in children are derived from prospective antimicrobial therapy studies for adults, with insufficient numbers of evaluable children to allow a basic understanding of the disease in children. Mycoplasma spp. clinical presentation is indistinguishable from that of viral pneumonia, prompting recommendations to clinicians to test for Mycoplasma before starting antimicrobial therapy. Presenting complaints are generally related to slowly progressive systemic symptoms over the course of 3 to 7 days, with malaise, pharyngitis, and headache, followed by cough. Cough usually is irritative and nonproductive, reflecting tracheobronchitis, a manifestation of cell injury and host response to lung inflammation, which rarely progresses in children to hypoxemia requiring hospitalization and supplemental oxygen. Cough may last for 2 to 4 weeks or longer and is associated with clinical findings of rales, rhonchi, and wheezes that may be so prominent in the context of a child who does not appear ill as to merit the common name of “walking pneumonia.” Similar to certain viral pathogens, M. pneumoniae may be associated with increased wheezing in children with asthma, particularly younger infants, although a recent meta-analysis suggested that, overall, wheezing was less likely to be present in mycoplasma infection compared with viral infection. The clinical presentation in preschool children is generally less severe than in older children, presumably as a result of less immune-mediated inflammation.

Diagnosis

Diagnostic testing for CAP is summarized in this section and reviewed in detail in the 2011 IDSA and PIDS guidelines. Pulse oximetry is recommended in both the outpatient and inpatient settings, when feasible, to identify children with hypoxemia who may require additional testing or admission.

Inpatient Setting

For children who require hospitalization, chest radiograph to evaluate the extent of infiltrate and to identify the presence or absence of pleural fluid should be obtained ( Fig. 22.5 ). Imaging in children with influenza and secondary fever or clinical deterioration may reveal necrotizing pneumonia or empyema ( Fig. 22.6 ). Blood cultures are more likely to be positive in the sicker children. The rate of bacteremia ranged from 1% to 11% in published studies. A systematic review reported an overall bacteremia prevalence of 5.1% in children with CAP, with rates of 4.1% and approximately 10% in nonsevere and severe CAP, respectively. In the children fully immunized with PCV vaccines, blood cultures are recommended only in children with moderate to severe illness. When positive, blood culture results can guide definitive antibiotic treatment. Rapid microarray assays performed on positive blood cultures now permit rapid differentiation of S. aureus from other gram-positive bacteria, thus minimizing the need to change the antibiotic regimen from aminopenicillins to vancomycin or antistaphylococcal penicillins in stable patients while awaiting organism identification.

Radiologic images of a 2-year-old girl with large parapneumonic effusion caused by Streptococcus pneumoniae. (A) Chest radiograph shows complete right hemithorax opacification with significant mediastinal shift to the left. (B) Computed tomography (CT) shows a massive right pleural effusion occupying the entire right hemithorax associated with leftward mediastinal shift. (C) The coronal CT view also demonstrates heterogeneous enhancement of the atelectatic right lung along with a lobulated collection of air, consistent with a mixture of compression-induced atelectasis and pneumonia.

Computed tomography image of a 1-year-old boy with influenza complicated by empyema caused by methicillin-resistant Staphylococcus aureus . There is a large right-sided pleural effusion with compression of the adjacent lung as well as pleural enhancement with some areas of cavitation. The left lung is hyperinflated with consolidation of the lower segment of the upper lobe.

Complete blood count may be useful in hospitalized children to identify anemia and thrombocytopenia associated with hemolytic uremic syndrome, a rare complication of bacterial pneumonia. Sputum for Gram stain and culture is recommended in adults but rarely performed in children. For older children and adolescents with severe CAP, sputum testing should be attempted.

Urinary antigen tests for S. pneumoniae are not recommended for children with CAP because false-positive results are common, occurring in 15% (95% confidence interval [CI], 11% to 22%) of febrile, nonbacteremic children in one emergency department–based study. The use of pneumococcal antigen testing for pleural fluid specimens, however, has been demonstrated to be reliable and improves the identification of a specific etiology for empyema, with a reported sensitivity of 83% to 96% and specificity of 95% to 100% (see Chapter 23 ).

M. pneumoniae is a common cause of CAP, particularly in children aged 10 to 17 years. The role of diagnostic testing for M. pneumoniae is not well defined because limited studies have been conducted on the natural history of M. pneumoniae infection and the impact of antibiotic treatment, particularly in young children. Testing may be most useful in school-aged children to guide antibiotic therapy. A variety of diagnostic tests are available. Traditional culture for M. pneumoniae is impractical for most clinical laboratories, given the slow growth and complex nutritional requirements of the organism. Cold-agglutinin titers greater than 1 : 64 are common in adults, but it is unclear how often this occurs in children. In addition, infections with other organisms may cause increases in cold-agglutinin titers, limiting the utility of this diagnostic test. Serologic diagnosis is most often used by clinical laboratories. Several serologic tests are available, including a rapid test for M. pneumoniae immunoglobulin (IgM), with sensitivity in published studies ranging from 74% to 96% and specificities ranging from 85% to 98%. PCR technology is currently available for M. pneumoniae in regional reference laboratories as well as in hospital laboratories as part of a multiplex PCR test panel that is currently commercially available. PCR tests have high specificity but variable sensitivity, making these tests more challenging to interpret than serology. In one observational study of children 3 months to 16 years of age, the prevalence of M. pneumoniae by real-time PCR did not differ significantly between asymptomatic (21% of 405 children) and symptomatic (16% of 321 children, P = .11) children. Longitudinal sampling in a subset of children found that most children with M. pneumoniae —15 of 21 (71%) asymptomatic children and 19 of 22 (86%) symptomatic children—tested negative after 1 month. However, 19% of all children with longitudinal follow-up demonstrated persistence of M. pneumoniae in the upper respiratory tract for up to 2 months. Thus interpretation of positive results, including the duration of positivity, from a variety of specimens such as nasal wash fluid, throat swab, sputum, and pleural fluid will all require further validation before clinical management based on PCR can be universally realized.

Pathogen-specific diagnoses for bacterial CAP remain challenging, although considerable progress has been made in the identification of pathogens through molecular testing of respiratory specimens. Both the PERCH and EPIC studies described previously use conventional and molecular diagnostics for viral and bacterial pathogens. A combination of these techniques identifies a viral or bacterial pathogen in 75% to 80% of hospitalized cases. Nucleic acid amplification techniques have been useful in identifying pathogens in culture-negative empyema. The available methods and reagents are expanding rapidly and include fluorescent antibody techniques, enzyme-linked immunosorbent assay, direct DNA probes, and PCR. Clinical specimens may be tested directly or after preincubation in tissue culture systems.

Biological markers, or biomarkers, have been explored as a method to improve assessment of disease severity and prediction of etiology to facilitate management decisions, including whether to initiate antibiotic therapy or admit a child to the hospital. Currently available biomarkers, including white blood cell (WBC) count, C-reactive protein (CRP), procalcitonin (PCT), and proadrenomedullin, reflect host inflammatory response to infection. WBC counts and CRP concentrations have insufficient diagnostic accuracy to serve as biomarkers in predicting etiology in children with CAP. While the WBC count is elevated in many children with bacterial pneumonia, most children with CAP and elevated WBC counts do not have bacterial infection. Furthermore the degree of elevation does not reliably distinguish bacterial from viral infection. CRP is produced by the liver in response to cytokines released at the site of inflammation. A meta-analysis that included eight studies and 1230 patients with CAP found that serum CRP concentrations exceeding 3.5 to 6.0 mg/dL weakly predicted bacterial pneumonia. Furthermore there was significant variation in the sensitivity (range, 17–100%) and specificity (range, 40–88%) of elevated CRP concentrations in predicting bacterial pneumonia. Among 161 children with CAP, CRP concentrations of greater than 6 mg/dL had a sensitivity of 26% and a specificity of 83% in identifying children with pneumococcal pneumonia. PCT, a non–thyroid tissue–derived precursor of the thyroid hormone calcitonin, demonstrates greater promise, but additional data, including more rigorous methods to prove or disprove bacterial infection (e.g., comparing PCT to a gold standard for bacterial infection), are necessary before PCT can be routinely recommended to inform decisions regarding antibiotic use or hospitalization. PCT concentrations in most studies are significantly higher among children with bacterial pneumonia compared with pneumonia caused by viral or atypical pathogens. Concentration cutoffs ranging from greater than 0.5 ng/mL to greater than 2 ng/mL have been proposed to distinguish bacterial from nonbacterial pneumonia, although these studies included relatively small numbers of patients with proven bacterial infection, and the PCT values lack sufficient sensitivity for routine clinical use at this time. Among 75 children with CAP, 37 met criteria for presumed pneumococcal CAP. The authors found that a PCT of less than 0.5 ng/mL ruled out pneumococcal CAP in more than 90% of cases, suggesting that PCT may be more helpful in informing the decision to not prescribe antibiotic therapy. Proadrenomedullin (ProADM), a midregion fragment of the parent precursor of adrenomedullin, demonstrates promise. ProADM improved the prognostic accuracy of the Pneumonia Severity Index, an established prediction rule in adults with CAP, in some studies but offered only additional risk stratification among high-risk patients in another. Few studies have explored its predictive value in children. Among 88 children with CAP, Alcoba et al. found that ProADM greater than 0.16 nmol/L had 100% (95% CI, 40% to 100%) sensitivity and 70% (95% CI, 59% to 80%) specificity for bacteremia. ProADM concentrations were twofold higher in complicated ( n = 11; 0.18 nmol/L, P = .039) compared with uncomplicated ( n = 77; 0.08 nmol/L) pneumonia.