Chapter 33. Uncomplicated Pneumonia

The World Health Organization defines pneumonia as the presence of cough and fast or difficult breathing (above 50 breaths per minute for children 2 to 12 months of age; above 40 breaths per minutes for children 12 months to 5 years of age).1 This broad definition may encompass other causes of lower respiratory tract infection including bronchiolitis as well as noninfectious causes of respiratory distress such as asthma. A more specific definition for pneumonia is “the presence of fever, acute respiratory symptoms, or both, plus evidence of parenchymal infiltrates on chest radiography.”2 This definition allows for the possibility of bacterial as well as viral causes of pneumonia.

Over 4 million children are diagnosed with pneumonia each year in the United States.3 The annual incidence of pneumonia of 34–56 per 1000 in children <5 years of age is higher than at any other time of life except perhaps in adults >75 years of age.3–6 This incidence decreases to approximately 16 cases per 1000 children 5 years of age or older.7 In 1996, the incidence of death from community-acquired pneumonia (CAP) in US children was <5 per 100,000; this figure represents a mortality rate of <1% per episode of pneumonia that requires hospitalization. Though mortality rates attributable to CAP have decreased by 97% over the past 50 years,8 more than 600,000 children annually require hospitalization for the management of CAP and its complications.3,9

A predisposition for respiratory tract infections seems to be a risk factor for developing pneumonia as evidenced by a study of 201 children with CAP between 3 months and 15 years of age.10 Compared to a cohort of healthy controls, children with CAP were more likely to have a history of recurrent respiratory infections during the past year, wheezing episodes, otitis media, and tympanocentesis before the age of 2 years.

Relatively common causes of childhood pneumonia are summarized in Table 33–1. Viruses are an important cause of childhood pneumonia.11–14 Respiratory viruses may be the primary cause of pneumonia or viral infection may predispose to bacterial superinfection by causing extensive tracheobronchial inflammation. Streptococcus pneumoniae remains the most common bacterial cause of lobar pneumonia outside of the perinatal period though its frequency may be decreasing as a result of the seven valent pneumococcal protein–polysaccharide conjugate vaccine (PCV7) licensed in 2000.15 Three randomized, double-blind controlled trials, one in the United States and two in developing countries, have demonstrated that vaccination with a pneumococcal conjugate vaccine decreased the incidence of CAP.16–18 The Kaiser Permanente trial in Northern California enrolled 37,868 children; 11% of control patients developed CAP during the study period.16 The decrease in pneumonia among PCV7 vaccinated patients was 20.5% when pneumonia was defined by the presence of clinical symptoms plus an abnormal chest radiograph.16 A subsequent observational study revealed that pneumonia-associated hospitalizations decreased by 39% in the United States from 2001 to 2004.19

Once considered to occur primarily among adolescents and young adults,20–22Mycoplasma pneumoniae is increasingly being recognized as a cause of lower respiratory tract disease in young children7,11,13,14,23–26. In a study of CAP in Finland, M. pneumoniae was the causative pathogen in 9% of children 0–4 years of age and in 40% of children 5–9 years of age.13 Although the frequency of pneumonia due to any respiratory pathogen decreases in children older than 5 years of age, the relative importance of M. pneumoniae as a cause of pneumonia increases with patient age. M. pneumoniae accounts for approximately 20–40% of cases of acute pneumonia in junior high and high school students and for up to 50% of cases among college students.27–31

Bacterial Pneumonia

S. pneumoniae is the leading bacterial cause of CAP in children. Pneumococci gain access to the lung by aspiration or inhalation. They lodge in bronchioles, proliferate, and initiate an inflammatory process that begins in alveolar spaces with an outpouring of protein-rich fluid. The fluid acts as culture medium for the bacteria and helps spread them to neighboring alveoli, typically resulting in lobar pneumonia.32 In some cases, bacteremia leads to hematogenous seeding of the lung with subsequent development of pneumonia. This latter mechanism is thought to be more common for Staphylococcus aureus.

CAP-related complications may occur by several overlapping mechanisms. First, progressive disease may lead to subpleural bacterial replication, endothelial injury, fibrin deposition, impaired lung perfusion (causing necrosis), and extension of infection to contiguous sites (e.g., empyema, bronchopleural fistula).33 Second, as leukocytes control bacterial multiplication, dying bacteria release cell wall and other components, which lead to overwhelming lung inflammation and, in some cases, respiratory failure.34 Third, bacteremia may lead to metastatic infection (e.g., endocarditis, meningitis) and organ dysfunction. Mortality, though rare, usually results from sequelae of bacteremia.35 Local complications of pneumonia are discussed in Chapter 34.

Atypical Pneumonia

Organisms such as M. pneumoniae and Chlamydophila pneumoniae gain access to the respiratory tract through aerosolized droplets spread among close contacts. Specific attachment to the respiratory epithelial tissue occurs primarily through interaction between a host epithelial cell receptor and the organisms attachment proteins.36–38 Following attachment, hydrogen peroxide and superoxide radicals synthesized by the atypical pathogens act in concert with endogenous toxic oxygen molecules generated by host cells to induce oxidative stress in the respiratory epithelium.39 In the lower respiratory tract, the organism may be opsonized by complement or antibodies. Activated macrophages begin phagocytosis and undergo chemotactic migration to the site of infection. CD4+ T lymphocytes, B lymphocytes, and plasma cells infiltrate the lung. Further amplification of the immune response occurs in association with lymphocyte proliferation, production of immunoglobulins, and release of tumor necrosis factor alpha, gamma interferon, and various interleukins.40 Lymphocyte activation and cytokine production may either minimize disease by enhancing host defense mechanisms or exacerbate disease by stimulating immune-mediated lung injury.41 Extrapulmonary complications of M. pneumoniae infection may occur as a consequence of either immune-mediated injury or direct invasion. Effects of cross-reactive antibody remain the proposed mechanism for hemolysis and cutaneous manifestations.42

Neonatal Pneumonia

Neonatal pneumonia can be acquired in three main ways: Congenital, during birth, or after birth. Congenital causes of pneumonia include Toxoplasma gondii, rubella, herpes simplex virus, mumps, adenoviruses, Listeria monocytogenes, and Mycobacterium tuberculsosis. Group B streptococci are responsible for most cases of bacterial pneumonia acquired at delivery. Chlamydia trachomatis is acquired during passage through an infected birth canal though cases occasionally occur in infants born by cesarean section after prolonged membrane rupture. Pneumonia develops in 5–20% of infants born to mothers with C. trachomatis infection.

The classic presentation of pneumonia is the sudden onset of fever, cough, and tachypnea. However, the clinical presentation of pneumonia in children can be diverse, and differentiating upper respiratory tract from lower respiratory tract infection can be especially challenging. Additionally, childhood pneumonia may be preceded by symptoms of viral upper respiratory tract infection.

Clinical examination findings of bacterial pneumonia include tachypnea, rales, and retractions. Tachypnea is the most sensitive sign of pneumonia with a sensitivity ranging from 64% to 81%43–48 (i.e., most patients with pneumonia have tachypnea at initial evaluation); however, its specificity is poor (i.e., most patients with tachypnea do not have pneumonia). The sensitivity of rales is lower (43–57%) but the specificity is higher (60–70%).43 Thus, pneumonia should be considered in any child with unexplained tachypnea, even if the lungs are clear to auscultation.

Abdominal pain is a well-recognized presenting symptom in children with basilar pneumonias.49 This occurs because the T9 dermatome is shared by the lung and the abdomen.50 The abdominal symptoms can be quite impressive and pneumonia has been inadvertently detected on abdominal computed tomography in children evaluated for suspicion of appendicitis.

Wheezing is more typical of viral respiratory infections and reactive airways disease, but may occur in childhood pneumonia, especially those cases caused by atypical bacteria such as M. pneumoniae and C. pneumoniae.11,51 Children infected with these atypical bacteria may present with the gradual onset of headache, malaise, fever, and pharyngitis with the subsequent development of cough and dyspnea, indicating the development of bronchopneumonia. However, more acute presentations occur in up to 25–40% of patients.52 A typical course for infections caused by atypical pathogens includes a dry, nonproductive cough, which develops approximately 3–5 days after the initial symptoms; it can become productive of mucopurulent or blood-streaked sputum. The cough usually brings the patient to medical attention 5–7 days after onset of symptoms. Coughing from either tracheobronchitis or pneumonia can persist for 2 or more weeks after resolution of fever and other constitutional signs and symptoms, resulting in a month-long illness that may resemble pertussis.53 Physical findings associated with lower respiratory tract disease such as wheezing may be absent during the first week of the illness, the time when cough and constitutional symptoms are most troublesome. As sputum production begins in the second week of illness, fever, myalgias, and other constitutional symptoms have resolved, but auscultation of the chest frequently reveals crackles or wheezes or both, thus the designation “walking pneumonia.”

An association between atypical bacteria and asthma has been proposed in a study which identified M. pneumoniae in 50% of children hospitalized with their first asthma attack and in 20% with a prior episode of wheezing.51 In that study,51 children with M. pneumoniae or C. pneumoniae infection during their first asthma attack were far more likely to have asthma recurrences.

Pneumonia caused by C. trachomatis typically manifests between 4 and 12 weeks of life. Infants are usually afebrile with tachypnea and diffuse rales; wheezing is not usually present. Approximately half of infants with pneumonia also have conjunctivitis.

Pneumonia due to an infectious etiology is by far the most likely diagnosis in a patient presenting with fever, cough, and focal consolidation. However, a number of noninfectious conditions may mimic pneumonia (Table 33–2) including foreign body aspiration, heart failure, malignancy (e.g., lymphoma), atelectasis, pulmonary embolus, pulmonary hemorrhage (e.g., idiopathic pulmonary hemosiderosis, Wegner's granulomatosis), and sarcoidosis. Collagen vascular diseases including systemic lupus erythematosus, scleroderma, and rheumatoid arthritis may cause interstitial lung disease with pulmonary fibrosis. Other causes of interstitial lung disease include environmental irritants such as noxious gases, radiation, and certain drugs. Hydrocarbons (e.g., gasoline, kerosene, charcoal lighter fluid) may be accidentally ingested by young children, leading to a secondary pneumonitis. Congenital lung anomalies, which can lead to pneumonia, include pulmonary sequestration, congenital lobar emphysema, and congenital cystic adenomatoid malformations.

Aspiration pneumonia should be considered in children with underlying neuromuscular disorders, swallowing difficulties, or gastroesophageal reflux. Infection is typically caused by oral flora and may include anaerobes such as Peptostreptococcus spp., Fusobacterium spp., and Bacteroides spp. Gram-negative rods that cause aspiration pneumonia include Escherichia coli, Pseudomonas aeruginosa, and Klebsiella pneumoniae.

Specific animal exposure may suggest a less common infectious etiology of pneumonia (Table 33–3). Exposures to consider include pigeons (Chlamydia psittaci, Cryptococcus neoformans), animal urine (leptospirosis), and rabbits (Francisella tularensis). Bacillus anthracis is associated with wool sorting and may be seen in animal handlers and veterinarians as well as in the context of bioterrorism.

Plague is an infection of rodents caused by Yersinia pestis and transmitted to humans by infected fleas. Symptoms of pneumonic plague include high fever, chills, myalgias, malaise, headache followed within 24 hours by cough and hemoptysis. Human-to-human transmission occurs by inhaling infected droplets from patients with pulmonary infection.

Hantavirus pulmonary syndrome is a disease that begins with a viral prodrome and can rapidly progress to respiratory failure and circulatory collapse. Hantavirus is carried by rodents such as deer mice and is spread through the inhalation of aerosolized urine or dried excreta.

Radiologic Imaging

The clinical manifestations of CAP are diverse and are shared by other infectious and noninfectious respiratory conditions, therefore chest radiography (CXR) should be used to confirm the presence and determine the location of a pulmonary infiltrate in all children with suspected CAP evaluated in the emergency department as well as those requiring hospitalization. Both posterior–anterior (PA) and lateral views should be obtained since nonlobar infiltrates may be missed using only frontal view CXRs.54 We recognize that routine performance of CXRs in children with suspected CAP in the ambulatory setting is neither practical nor possible. Furthermore, in children with relatively mild symptoms, it is unlikely that the CXR results will alter management or outcome. In the office setting, CXRs should be performed in moderately ill patients, including those with hypoxia, and in those whose symptoms persist despite antibiotic therapy.

The radiographic pattern of the infiltrate does not reliably differentiate among various causes. For example, an alveolar infiltrate (disease affecting the terminal air space characterized by a lobar or segmental distribution with a tendency to coalesce) (Figures 33–1 and 33–2) may represent viral pneumonitis rather than bacterial pneumonia. While interstitial infiltrates are more typical of M. pneumoniae (Figure 33–3), patchy, unilateral, segmental, or subsegmental consolidation also occur (Figure 33–4). The presence of specific findings may serve to raise the suspicion of specific pathogens (Table 33–4). Pneumatoceles (air-filled cavities resulting from alveolar rupture) are detected in approximately two-thirds of patients with pneumonia caused by S. aureus (Figure 33–5).55 The presence of hilar adenopathy, especially bilateral, out of proportion to the extent of parenchymal consolidation should raise suspicion for infection caused by M. tuberculosis, endemic fungi (e.g., Histoplasma capsulatum), and M. pneumoniae. Nodular infiltrates, though more common with infections caused by Pneumocystis jiroveci (formerly P. carnii) and endemic fungi, may also be caused by viruses and atypical bacteria.

An important caveat is that variation exists among radiologists on the presence or absence of radiographic features used for diagnosis of CAP. When radiographs for 34 children with a diagnosis of CAP and 34 children without CAP were reviewed without knowledge of the prior interpretation or clinical history, there was variation in the radiographic confirmation of pneumonia in 24% of cases and variation regarding the location of the lesion in 50% of cases.56 In a study of adults with suspected CAP, there was fair to good interobserver reliability in detecting the presence of an infiltrate or pleural effusion; however, the interobserver reliability of other radiographic characteristics such as the pattern of the infiltrate (alveolar vs. interstitial) and the presence of air bronchograms and hilar lymphadenopathy was poor.57 Computed tomography (CT) of the chest should be performed to more clearly define specific abnormalities such as nodular infiltrates, lung abscesses, necrotizing pneumonia, and pleural effusions.

Blood Cultures

Blood cultures are positive for pathogenic bacteria in <2% of patients with pneumonia who are well enough to be managed in the outpatient setting.16,58,59 Blood cultures are more commonly positive in patients requiring hospitalization for pneumonia (7–10%) and in patients with pneumonia complicated by pleural effusion (up to 25%).55,60,61 Despite the low overall yield of blood cultures in patients who require hospitalization, information on the causative organism may provide information that allows for directed narrowing or broadening of the spectrum of antibiotic therapy. Therefore, we believe that blood cultures are not required for patients with CAP who are well enough to be treated in the outpatient setting but should routinely be obtained from children with CAP requiring hospitalization.

Other Microbiologic Investigations

Gram stain and culture of expectorated sputum is infrequently performed in children with CAP because children cannot always provide adequate specimens for testing. Gram stain and culture of expectorated sputum should be considered in patients with cavitary lesions, failure of outpatient therapy, and intensive care unit hospitalization if a good quality specimen can be provided. Expectorated sputum should be visibly purulent and transported to the laboratory as soon as possible (within 2 hours) since delayed transport times decrease the viability of many important pathogens and allow for the overgrowth of commensal flora. Sputum culture results must be interpreted in concert with the Gram stain. Laboratory assessment of sputum quality is based on the number of squamous epithelial cells (<10/high power field) and white blood cells (WBCs) (>25 per high power field) on Gram stain. In our experience, cooperative children 8 years of age or older with a productive cough can usually provide an adequate sputum sample. Sputum production can be induced with administration of nebulized hypertonic (5%) saline.

In young infants, C. trachomatis can be detected by direct fluorescent antibody testing or by culture of a nasopharyngeal (NP) or conjunctival swab. Polymerase chain reaction (PCR) testing of a NP aspirate is very sensitive for B. pertussis whereas culture is time-consuming and direct fluorescent antibody testing has poor sensitivity. In a patient with diffuse wheezing, M. pneumoniae and C. pneumoniae can be detected by PCR testing of a NP aspirate; if PCR is not available, acute and convalescent serum antibody titers are appropriate. Viruses can be detected in NP aspirates by PCR or immunofluorescence, but identification of a virus does not exclude the possibility of bacterial superinfection. Nasopharyngeal bacterial cultures are of no value in the etiologic diagnosis of childhood pneumonia since S. pneumoniae and other lower respiratory tract pathogens commonly colonize the nasopharynx in healthy children.

Urinary antigen tests for the detection of S. pneumoniae are routinely used to diagnose pneumococcal pneumonia in adults.62S. pneumoniae can be isolated from sputum culture in up to 80% of adults with a productive cough and a positive pneumococcal urinary antigen test.63,64 These tests have good sensitivity in identifying children with pneumococcal bacteremia (sensitivity, 77–100%).65 They were also positive in 47 of 62 (76%) patients with lobar pneumonia; however, since the etiology of pneumonia could not be confirmed in these patients, the relevance of this finding is not clear.65 Additionally, false positives, attributable to pneumococcal nasopharyngeal colonization, occur in 15% of children. Pneumococcal urinary antigen tests are not routinely recommended since the results are difficult to interpret in the absence of a true gold standard.

Acute Phase Reactants

Acute phase reactants, including peripheral WBC count with differential, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR), do not reliably distinguish bacterial from viral infections. Korppi et al. found that the WBC count, CRP, and ESR were significantly higher in children with pneumococcal pneumonia compared with viral or atypical pneumonia.66 However, the number of patients with pneumococcal disease was relatively small (n = 29), there was considerable overlap in values between the two groups, and the sensitivity and positive predictive value for their cutoffs were low. The sensitivity for a CRP >6.0 mg/dL or an ESR >35 mm/h was 26% and 25%, respectively. The positive predictive value for CRP >6.0 mg/dL or an ESR >35 mm/h was 43% and 38%, respectively.66 Nohynek et al. found wide variation in WBC count, CRP, and ESR values between children with CAP attributable to bacteria and viruses; the values did not differ significantly between the two groups.67 Therefore, we feel that acute phase reactants can be measured at baseline for patients requiring hospitalization and may then be useful in conjunction with clinical findings in assessing response to antibiotic therapy; however, they are not useful in differentiating bacterial versus viral etiologies.

Other Laboratory Tests

Other laboratory tests include serum electrolytes and arterial or venous blood gas measurements. Serum electrolytes should be measured in ill, appearing patients and in those with clinical evidence of dehydration. Inappropriate secretion of antidiuretic hormone, leading to hyponatremia, is relatively common in children with severe pneumonia.68,69 The likely mechanism is latent vasopressin-dependent impairment of renal water excretion.70 Resolution of hyponatremia correlates with clinical recovery. Other electrolyte abnormalities, including hypernatremia and hyper- or hypokalemia, are much less common. Blood gas measurements allow detection of hypoxemia and metabolic acidosis. Most instruments also provide crude (but rapid) measurements of the patient's hemoglobin and serum sodium; abnormalities in these parameters detected by bedside instruments should always be confirmed in the clinical laboratory.

Treatment of CAP is usually empiric since definitive information about the causative pathogen is seldom available (Table 33–5).

Outpatient Therapy

For outpatients, amoxicillin remains appropriate first-line therapy. Alternate antibiotics for patients with penicillin allergy include clindamycin, levofloxacin, third-generation cephalosporins or macrolide derivatives such as clarithromycin or azithromycin. Amoxicillin-clavulanate, while providing coverage against beta-lactamase producing Haemophilus influenzae and Moraxella catarrhalis, does not provide expanded coverage against S. pneumoniae, the most likely bacterial pathogen. Pneumococcal resistance to beta-lactam antibiotics is mediated by alterations in penicillin-binding proteins; increasing the dose of amoxicillin or penicillin can overcome this mechanism of resistance but adding a beta-lactamase inhibitor such as clavulanic acid does not. If a patient with CAP initially managed in the outpatient setting fails to improve, a CXR should be performed to assess for the presence of local complications such as pleural effusions that require hospitalization for further management. The typical duration of treatment is 10 days (5 days can be used for azithromycin).

Inpatient Therapy

For hospitalized patients, ampicillin remains appropriate first-line therapy for uncomplicated pneumonia. For moderately or severely ill patients, cefotaxime, ceftriaxone, and levofloxacin provided broader coverage against penicillin-resistant pneumococci. Azithromycin, clarithromycin, or levofloxacin should be added to cover atypical pathogens for patients who do not respond to beta-lactam therapy; these may be administered orally. Vancomycin should be considered in patients with hypotension or necrotizing pneumonia where methicillin-resistant S. aureus (MRSA) is possible. Other options for coverage against MRSA include clindamycin or linezolid. The typical duration of therapy for uncomplicated pneumonia is 10–14 days. Patients may be switched to oral therapy at the time of discharge.

Changing Drug Resistance Patterns and Selection of Empiric Therapy

A key question is whether the changing patterns of pneumococcal drug resistance should change our empiric antibiotic choices for uncomplicated CAP. The alarming increase in the prevalence of penicillin resistance among S. pneumoniae has been well publicized. Prior to PCV7 licensure, approximately one-third of pneumococcal isolates obtained from children were resistant to penicillin.71 This changing antibiotic resistance pattern has prompted clinicians to administer other antibiotics to patients with CAP, including more frequent use of cephalosporin and macrolide class antibiotics.72 However, virtually all penicillin-resistant isolates were also resistant to second-generation cephalosporin antibiotics and 60% were resistant to erythromycin, a macrolide antibiotic.71 A reduction in drug-resistance among pneumococci has been noted in studies conducted shortly after PCV7 licensure.73–75 Kyaw et al.,76 using data from Active Bacterial Core surveillance from 1996 to 2004, identified a significant decrease in the incidence of penicillin-nonsusceptible pneumococcal infections (by 57%) and in the incidence of strains nonsusceptible to multiple antibiotics (by 59%) beginning in 1999.76 However, the incidence of invasive pneumococcal disease caused by penicillin-nonsusceptible vaccine-related and nonvaccine serotypes increased by 54% and 195%, respectively.76

The precise contributions of antibiotic-resistant S. pneumoniae and specific antimicrobial agents to clinical outcomes are not known. Some studies have failed to find a difference in clinical outcomes for children with pneumonia caused by penicillin-susceptible versus penicillin-nonsusceptible S. pneumoniae.77,78 However, these studies have generally included patients with low levels of drug resistance; it is possible that higher levels of drug resistance are more likely to lead to treatment failure than lower levels of drug resistance. Additionally, pediatric studies have included few patients receiving discordant therapy (i.e., receipt of antibiotics that are inactive in vitro against the isolated organism), a limitation that precludes specific assessment of the effect of drug resistance on outcomes. In contrast, Yu et al. reported that clinical outcome following discordant therapy depended on the class of antibiotic administered.79 Among 844 adult patients with pneumococcal bacteremia, cefuroxime (a second-generation cephalosporin) therapy was associated with higher mortality in the setting of cefuroxime resistance. However, therapy with penicillin or third-generation cephalosporins, such as cefotaxime and ceftriaxone, was not associated with increased mortality in the setting of penicillin or third-generation cephalosporin resistance, respectively. Other studies have documented treatment failure with macrolide-class antibiotics in the treatment of adults with pneumococcal pneumonia.80,81 In summary, the degree to which empiric antibiotic selection affects the clinical outcome of children with pneumonia is not known. In adults, use of second-generation cephalosporin or macrolide antibiotics is associated with higher rates of treatment failure and mortality compared with use of other antibiotic classes.

Some experts advocate using higher doses of cefuroxime to overcome cefuroxime resistance. While this strategy may be effective, there is no data to suggest that such a strategy would be any more effective than using penicillin, amoxicillin, or ampicillin. When broader pneumococcal coverage is desired, third-generation cephalosporins may be more appropriate than second-generation cephalosporins.

Hospitalization Decisions

Selection of the initial site of care, whether outpatient or inhospital, is one of the most important management decisions as it directly affects the intensity of subsequent testing and therapy. CAP-related admission rates vary significantly among nearby regions; within Pennsylvania, admission rates for children with CAP differ by as much as fivefold between adjacent counties.82 These data suggest that physicians do not use consistent criteria to make site of care decisions. Unnecessary hospitalization has disadvantages, including increased likelihood of exposure to ionizing radiation and increased health care costs. However, outpatient management of high-risk patients may increase CAP-associated morbidity.

Generally agreed upon indications for hospitalization include percutaneous oxygen saturation ≤92%, marked tachypnea (>70/min in infants or >50/min in older children), difficulty in breathing, grunting, dehydration, or poor oral intake, and when the family is unable or unwilling to provide close observation and recognize clinical worsening.4

Follow-Up Radiographs

Routine follow-up CXRs are not warranted in children who recover uneventfully from an episode of CAP. CXRs performed 3–7 weeks after an episode of radiographically confirmed CAP revealed residual or new changes in 59 (30%) of 196 children; persistence of interstitial infiltrates and the interval development of atelectasis were the most commonly noted findings.83 Follow-up 8–10 years later did not reveal any illnesses attributable to the initial episode of CAP.83 Among 41 children with CAP, repeat CXRs 4–6 weeks later were normal in 37 children and revealed resolving infiltrates in the remaining four children.84 In a prospective study of adults hospitalized with severe CAP, CXRs were repeated 7 and 28 days after admission.85 At day 7, 75% of patients still had abnormal findings on CXR. At day 28, 47% of patients still had abnormal findings on CXR. Delayed resolution of radiographic abnormalities was associated with multilobar disease, dullness to percussion on examination, higher CRP levels, and documented pneumococcal infection. However, delayed resolution of radiographic abnormalities did not portend failure of therapy or a worse clinical outcome. In summary, routine follow-up CXRs do not seem to provide additional clinical value. A subset of patients, such as those with lobar collapse or recurrent pneumonia involving the same lobe, may benefit from repeat CXRs.

• Tachypnea is the most sensitive sign of pneumonia.
• Abdominal pain is a well-recognized presenting symptom in children with basilar pneumonia.
• Blood cultures are not required for patients with CAP who are well enough to be treated in the outpatient setting, but should routinely be obtained from children with CAP requiring hospitalization.
• M. pneumoniae may cause lobar or segmental infiltrates on CXR.