Historical Aspects

The word croup is derived from the Anglo-Saxon word kropan, “to cry aloud.” Until the 20th century, most crouplike illnesses were thought to be diphtheria. Diphtheritic croup is an ancient disease that has been traced to the time of Homer. The historical trail of diphtheria disappeared in the 5th century and did not reappear until 1100 years later. In the 16th century, epidemics were noted in Europe. Top credits Bretonneau with differentiating diphtheritic croup from spasmodic croup in 1826. In the middle third of the 20th century, the history of croup was marked by three important events: (1) the rapid decline in incidence of diphtheria associated with the use of toxoid, (2) the introduction and widespread use of antibiotics, and (3) the advent of tissue culture techniques, resulting in the establishment of viruses as etiologic agents. After these three events occurred, a prevalent academic view was that all croup was of viral etiology, and bacteria generally were dismissed as causative agents. However, a careful review of many publications from the first half of the 20th century indicates a causative role for several bacteria, in addition to Corynebacterium diphtheriae, in croup. *

* References .

† References .

In the 1940s, Davison separated spasmodic croup from other, more severe forms of croup. The clinical and pathologic aspects of this entity were poorly defined, and today it often is not separated clinically from more severe forms of croup.

The early history of epiglottitis is obscure, probably because of the importance of diphtheritic croup. In 1887, Baron described in detail a 30-year-old woman with epiglottitis who recovered after treatment with hot poultices and steam with tincture of benzoin. In 1900, Theisen described three cases in the United States. Not until the early 1940s did acute epiglottitis become recognized as a definite clinical entity caused by H. influenzae type B. In 1948, Rabe, in a study of 347 children with “infectious croup,” presented evidence for the division of the clinical illness into three etiologic categories—diphtheritic croup, viral croup, and H. influenzae type B croup (acute epiglottitis). Since the early 1990s, a dramatic decrease in the number of cases of epiglottitis has occurred in many countries because of the widespread use of H. influenzae type B conjugate vaccines. ^‡

‡ References .

Terminology

The terminology and classification of infectious illnesses involving the larynx and supraglottic and infraglottic regions have evolved over time. Classifications often have mixed etiologic categories with anatomic systems and have led to confusion. Croup often has been presented in articles under the heading of laryngotracheobronchitis when the authors actually were discussing laryngotracheitis and spasmodic croup. The term membranous croup has been used as the title for articles dealing with bacterial croup This use is confusing because membranous croup historically was diphtheria. Many articles dealing with bacterial croup also have been titled bacterial tracheitis . *

* References .

Etiology of Croup Syndromes

The etiologic agents in laryngitis, laryngotracheitis, spasmodic croup, laryngotracheobronchitis, and laryngotracheobronchopneumonitis are presented by frequency and severity of illness in Table 18.2 . Laryngitis is a common manifestation of infection with many respiratory viruses in older children, adolescents, and adults. Outbreaks of laryngitis in closed population groups (e.g., boarding schools and military training camps) most frequently are caused by adenovirus types 4 and 7, and community outbreaks most often are noted in association with epidemic influenza. Sporadic instances of laryngitis usually are caused by adenoviral infections. Laryngitis also has been reported in association with group A streptococcal infections; the incidence of this association has varied from 2% to 40%. Somenek and associates described an 11-month-old boy with severe membranous laryngitis caused by methicillin-resistant Staphylococcus aureus (MRSA).

Generally accepted today is that acute laryngotracheitis and spasmodic croup, which rarely are differentiated clinically, are caused by infection with many different viruses. Although numerous studies of respiratory viral infection exist, almost no attempt has been made to delineate the differences in etiologic spectrum by severity of illness.

Parainfluenza virus type 1 is the most common cause of acute laryngotracheitis and is responsible for frequent and clearly delineated fall and winter epidemics. Croup with parainfluenza type 2 virus seldom is severe but occasionally is related to small outbreaks. Parainfluenza virus type 3 is a frequent cause of sporadic but severe illness.

The most severe laryngotracheitis has been noted in association with influenza A viral infections. Respiratory syncytial virus and several different adenoviruses frequently are isolated in croup. Generally these illnesses are not severe, but lower respiratory tract involvement occasionally is a problem. Laryngeal, tracheal, and bronchial involvement commonly occurs in measles. Although rhinoviruses, Mycoplasma pneumoniae, enteroviruses, herpes simplex virus, and reoviruses have been associated with croup, they generally cause only minimal distress. However, O’Niel and associates reported two immunocompetent children who presented with severe acute laryngotracheitis caused by herpes simplex virus. Croup has been noted in association with infection with the novel coronavirus NL63, human bocavirus, and human metapneumovirus. Lee and Storch noted that coronavirus NL63 was a relatively common cause of severe croup. Recurrent croup may be due to infection with a human papillomavirus.

Bacteria, other than H. influenzae in epiglottitis and C. diphtheriae in membranous croup, generally were dismissed as causative agents in croup until more recently. Many publications on laryngotracheobronchitis from the first half of the 20th century indicate a role for several common bacterial pathogens. *

* References .

† References .

In the reports from the preantibiotic era, Streptococcus pyogenes was the pathogen implicated most frequently. Since 1979, S. aureus has been the agent implicated most commonly. Other important bacteria are Streptococcus pneumoniae and H. influenzae . More recently, Moraxella catarrhalis has been found to be the causative agent in several cases. In most instances, bacterial croup is likely to be the result of bacterial superinfection in viral disease. ^‡

‡ References .

Noninfectious causes of crouplike illnesses include airway hermangioma and laryngeal or esophageal foreign bodies.

Etiology of Supraglottitis

Acute supraglottitis in children in the prevaccine era almost always was caused by H. influenzae type B. ^§

§ References .

In countries where H. influenzae type B immunization is routine, supraglottitis is a rare illness in children. Nevertheless, cases still occur in association with some infectious agents. The following pathogens have been implicated in this regard: S. pneumoniae; S. aureus; H. parainfluenzae; groups A, B, C, and G streptococci; Pseudomonas aeruginosa (in a patient with severe combined immunodeficiency syndrome); untypable H. influenzae; and Bacillus spp. Candida tropicalis was isolated from the blood of a -year-old girl with supraglottitis who had been the recent recipient of an autologous bone marrow transplant.

Candida albicans was noted in a case in a newborn whose mother had vaginal candidiasis and in a 6-year-old boy with chronic mucocutaneous candidiasis. Candida spp. epiglottitis was noted in two children infected with human immunodeficiency virus (HIV), and C. albicans epiglottitis was observed recently in a 2-year-old girl who was immunocompromised as a result of receiving chemotherapy for a primitive neuroectodermal tumor.

Necrotizing epiglottitis in a 5-year-old boy with hemophagocytic lymphohistiocytosis associated with a mixed infection ( Enterococcus faecalis, Eikenella corrodens, Prevotella melaninogenica, and Neisseria spp.) has been described. Reed and colleagues described a 17-year-old boy with an aryepiglottic abscess associated with a mixed infection ( Capnocytophaga ochraceae, Bacteroides stercoris, and P. melaninogenica ).

Since 2009 in England and Wales, there has been an increase in invasive meningococcal diseases due to capsular group W sequence type 11 complex. Between 2010 and 2013, there were five cases of epiglottis due to this meningococcal complex. Lake and colleagues noted necrotizing epiglottitis due to nontoxigenic Corynebacterium diphtheriae in a 3-year-old girl who had nondiagnosed lymphoblastic leukemia. A 14-month-old girl with prolonged stridor and epiglottitis associated with a mixed infection with herpes simplex virus type 1, H. influenzae , S. pneumoniae , and S. aureus has been reported. A 16-month-old child with type 1 herpes simplex virus stomatitis complicated by stridor and respiratory distress had epiglottis; his arytenoepiglottic folds were edematous and covered with vesicular lesions resembling those in the oral mucosa. An 18-year-old girl had supraglottitis that also was caused by herpes simplex virus. In addition, parainfluenza type 3 and influenza type B viruses were isolated from the nasopharynx of two children with supraglottic inflammation. Haemophilus paraphrophilus was recovered from the epiglottic surface of a single patient, as was M. catarrhalis in another patient.

In adults, H. influenzae type B also has been the major cause of epiglottitis, but other organisms occur more commonly in adults than in children. *

* References .

In Finland, Bizaki and coworkers reviewed 308 adult cases of acute supraglottitis during the 2-decade period from 1989 to 2009. The incidence increased from 1.88 per 100,000 in the first decade to 4.73 in the second decade. The causative organisms were S. pyogenes, S. pneumoniae, S. milleri, staphylococci Pseudomonas spp., H. influenzae, and C. albicans .

S. pneumoniae was reported to be isolated from the blood of 18 adults with supraglottitis, 10 of whom were receiving immunosuppressive therapy or were infected with HIV-1, and H. parainfluenzae was isolated from the blood of 5 patients. In Denmark, between 1995 and 2002, H. influenzae type F was recovered from 13 cases of epiglottitis in adults. Numerous other infectious agents have been implicated in case reports of adults with supraglottitis. ^†

† References .

Epiglottitis also can result from noninfectious causes. Hot foods and water can cause thermal epiglottitis, as can poisoning with corrosive agents, including cocaine alkaloid. Tsai and Wang noted acute epiglottitis in a 23-year-old man after receiving traditional Chinese gua sha therapy. Hereditary angioedema may manifest with findings typical of epiglottitis.

Epidemiology of Croup

Croup accounts for approximately 15% of lower respiratory tract disease seen in pediatrics. In a large 11-year study in a pediatric practice in Chapel Hill, North Carolina, Denny and associates noted the incidence of croup by age and sex. The highest attack rate occurred in children 7 to 36 months of age. Few cases occurred after the sixth birthday. Hoekelman studied the occurrence of illness prospectively in 246 full-term, first-born, well infants during their first year of life. Three infants (1.2%) had croup during the study year. The analysis of a pediatric practice with approximately 3000 active records and approximately 10,000 yearly visits of children younger than 5 years disclosed five cases of croup in a group of 50 consecutive hospitalized patients. Although croup occurs occasionally in older children, most cases occur within the first 3 years of life.

Croup occurs more commonly in boys than in girls. In a 6-year study in Alberta, Canada, involving 20,079 emergency department visits, 61.3% were boys. In the Chapel Hill study, an increase was noted in the number of croup cases beginning in September, with a peak in October and November, and then a decrease during the next 7-month period. In a 2-year emergency department study in Toronto involving 1700 cases, the peak month of visits and hospital admissions was found to be October. Marx and associates reviewed the National Hospital Discharge Survey data for hospitalizations for croup between 1979 and 1993. They also examined Centers for Disease Control and Prevention laboratory-based surveillance data and published reports with virus isolation studies. Major peaks in hospitalizations for croup occurred in October of odd-number years at the time of peak parainfluenza virus type 1 activity. Minor peaks in hospitalizations for croup occurred each year in February when influenza A, influenza B, and respiratory syncytial viral infections were common occurrences. Epidemic peaks of acute laryngotracheitis reflect community-wide activity with parainfluenza 1 and 2 viruses or influenza A or B outbreaks.

In the Alberta, Canada, study in 1990 to 2000 and 2004 to 2005, biennial trends were apparent. Peak emergency department visits for croup occurred in November for odd years and in February for the other years. For the odd years, the overall rate increased from 43.4/1000 in 1999 to 2000 to 49.6/1000 in 2003 to 2004. A similar increase was noted in the even years (30.9/1000 in 2000 to 2001 to 34.1/1000 in 2004 to 2005).

In the Toronto study, the time of the visit to the emergency department was analyzed. The peak number of visits occurred between 10 pm and 4 am . During this period, approximately 17% of the children seen were admitted to the hospital. In contrast, of children seen between noon and 6 pm , approximately 50% were admitted to the hospital. Quite similar findings were noted in Alberta in the 6-year period between 1999 and 2005. A study of croup hospitalizations in Ontario over 14 years from 1988 to 2002 noted a biennial mid-autumn peak and an annual summer trough throughout the entire study period. A striking finding in this study was a marked decrease in hospitalizations for croup after the winter of 1993 to 1994. This decrease continued in linear fashion for the remainder of the study duration. The authors suggest that the decreasing trend in croup hospitalizations was likely due to the increasing use of corticosteroid treatment in patients who presented to emergency departments with croup.

Because croup is caused by the same viruses that cause other respiratory illnesses, the method of spread probably is similar for all (see discussion of the common cold in Chapter 7 ). In children, most spread involves close person-to-person contact, with large droplets of virus-containing nasal secretions being applied to the nose from the hands of the future host or by close-range airborne acquisition. Parainfluenza viruses are common causes of colds in adults, so older individuals with trivial illnesses may be the source of more severe childhood croup.

Epidemiology of Supraglottitis

In the prevaccine era, the epidemiology of invasive H. influenzae type B disease varied markedly among population groups. Epiglottitis also differed by population group, but this variation was not related necessarily to the rate of overall H. influenzae type B invasive disease in the population. For example, among Alaskan Eskimos and Navajo Native Americans, whose risk for acquiring invasive H. influenzae type B disease was 4 to 10 times that of most other American populations, epiglottitis was not recognized among 295 patients with invasive H. influenzae type B illness.

Of H. influenzae type B invasive disease, the percentage of cases of epiglottitis varied markedly among localities. For example, in Israel, only 0.3% of invasive H. influenzae type B infections were epiglottitis, whereas in Ireland, Wales, northeast England, Australia (Sydney), and Denmark, the percentage of epiglottitis cases varied between 16% and 32%. In Minnesota and Dallas County, Texas, only 6% and 3%, respectively, of invasive H. influenzae type B infections were epiglottitis. In contrast with these findings in the United States, Europe, and Australia, the most common manifestation of invasive H. influenzae type B infection in Sweden was epiglottitis. In Sweden, the incidence of epiglottitis in children 14 years of age or younger in 1981 to 1983 was 10/100,000 per year. The annual incidence in Minnesota and Dallas County, Texas, in children 5 years of age or younger in 1982 to 1984 was 5.4 and 4.4 cases, respectively, per 100,000 per year.

In children in the United States in the prevaccine era, the peak occurrence of epiglottitis was during the third year of life, and 72% of all cases occurred in children 1 to 5 years of age. The disease occurred more commonly in boys than in girls; in eight studies with 611 cases, 58% of cases were in boys.

In specific geographic areas, marked differences in the yearly percentage of cases of epiglottitis occurred, but no intercity, national, or international cycles of illness were demonstrated. Seasonal prevalence varied by locality but was not marked. The greatest number of cases in three studies occurred in the winter and spring, whereas Baxter observed more cases during the summer and in November, and Cohen and Chai found no seasonal pattern.

Epiglottitis is a disease that occurs most commonly in temperate climates. In several U.S. military hospitals, a wide geographic variation in incidence was noted; no cases were found among 4625 admissions at Gorgas Hospital in Panama, whereas 1/600 admissions to Elmendorf Hospital in Alaska were for epiglottitis. In the past, acute epiglottitis had not been reported in either Taiwan or Hong Kong.

Epiglottitis also occurs in adults but less commonly than in children. Annual rates in Rhode Island, Denmark, Finland, and northern California were 1.0, 0.9, 0.2, and 1.8/100,000, respectively. In both Ireland and Israel, rates of reported epiglottitis in adults appear to be increasing. Whether these increases are true increases or are the result of increased awareness is not apparent.

In the present H. influenzae type B conjugate vaccine era, the incidence of epiglottitis has fallen dramatically in all countries employing routine immunization, as has the incidence of all invasive disease caused by H. influenzae . *

* References .

It should be noted that sporadic cases of epiglottitis mostly caused by non– H. influenzae type B bacteria continue to occur in the United States. In 2006, 377 cases were reported including 369 cases in children.

In 2006, an increase in incidence of invasive H. influenzae type B disease, including acute epiglottitis, was reported in England. Many of the cases occurred in children who had received H. influenzae type B conjugate vaccine at 2, 3, and 4 months of age but not a booster dose in the second year of life. This resurgence of invasive H. influenzae type B disease led to the implementation of a booster dose at 12 months of age, which helped to control the burden of invasive H. influenzae type B in all age groups.

Pathology and Pathogenesis of Croup

Laryngoscopic studies in acute laryngotracheitis reveal redness and swelling of the lateral walls of the trachea, just below the vocal cords. Because the subglottic trachea is surrounded by a firm cartilaginous ring, the inflammatory swelling can occur only by encroaching on the patency of the airway; the subglottic space often is reduced to a slit 1 to 2 mm wide. As the disease progresses, the tracheal lumen becomes obstructed further by a fibrinous exudate, and its surface is covered by pseudomembranes composed of the exudative material. The vocal cords frequently are swollen, and their mobility is impaired.

Histologic study of postmortem material from the larynx and trachea reveals marked edema and cellular infiltration in the lamina propria, submucosa, and adventitia. The cellular infiltrate includes histiocytes, lymphocytes, plasma cells, and polymorphonuclear leukocytes. The cotton rat model of laryngotracheitis caused by human parainfluenza virus type 3 reveals pathologic findings similar to those noted in human postmortem studies. This model also indicates the time course of events. Early in the course of an infection (first 2 to 4 days), moderate mucosal and submucosal inflammatory infiltrates were noted in the subglottic and proximal tracheal regions. The infiltrates initially contained lymphocytes and neutrophils, and subsequently a mononuclear infiltrate with lymphocytes and macrophages in the submucosa was noted. Cell injury was most marked on days 6 to 8 after infection. The ciliated epithelial cells were blunted, with loss of cilia in large patches. By 12 days after infection, the region reverted to a nearly normal appearance.

The older literature indicates that classic laryngotracheobronchitis and the same disease with pneumonia represent extension of disease from the trachea to the bronchi and alveoli. The progressive obstructive disease with exudate and pseudomembrane obstruction at the bronchial and bronchiolar levels usually is the result of secondary bacterial involvement. In bacterial croup, the tracheal wall is infiltrated with inflammatory cells, and ulceration, pseudomembranes, and microabscess formation occur. *

* References .

Spasmodic croup is an enigma because it occurs in association with respiratory viral infections similar to those that cause more severe laryngotracheitis. Using direct laryngoscopy, Davison noted that the subglottic tissues in spasmodic croup showed noninflammatory edema.

Although the eventual site of clinically important pathologic change in laryngotracheitis is within the larynx and trachea, the initial acquisition of infection is similar to that of other respiratory viral infections and occurs within the upper air passages, including the nasal and pharyngeal epithelial surfaces. After acquisition of virus, infection of the cells of the local respiratory epithelium develops and spreads locally to involve the larynx and trachea. The initial symptoms of nasal stuffiness and throat irritation reflect the primary sites of involvement. Studies in organ culture systems and in the cotton rat model have shown that several respiratory viruses inhibit tracheal ciliary function and eventually lead to marked destruction of the epithelium and evidence of viral infection in the lamina propria. In uncomplicated croup, failure of gas exchange within the lung, in addition to hypoxia resulting from subglottic tracheal obstruction, may occur.

Parainfluenza virus types 1, 2, and 3 all are significant causes of respiratory infections in children (see Chapter 179 ). Infection with parainfluenza virus type 3 occurs in most infants, whereas infections with types 1 and 2 generally occur more frequently in older children. Only a third of children have antibodies against these two viral types by the time they reach 4 years of age.

Because parainfluenza viral infections are common occurrences in young children, and because only a few get croup, host factors probably are important in the pathogenesis. In contrast to the numerous studies of the pathogenesis of bronchiolitis and respiratory syncytial virus infection, few similar studies relating to parainfluenza viruses and croup have been reported.

These pathologic data suggest that the findings in laryngotracheitis are directly virus related (parainfluenza virus types 1, 2, and 3 and influenza virus types A and B) because of the direct cytopathic effect of a virus, are related to the concomitant host response, or perhaps both. In the cotton rat model, the use of topical steroid therapy led to a significant reduction in the degree of inflammatory infiltrates and cell injury, which suggests that the host response contributes to the pathologic findings. Schaap-Nutt and associates described the replication kinetics and cytokine secretion of the parainfluenza viruses in human tracheobronchial airway epithelium. They found serotype-specific differences of the innate immune response, suggesting that parainfluenza type 1 is able to remain undetectable for several days after infection compared to parainfluenza type 2 and 3. Parainfluenza 2 inhibits the early immune response to a lesser extent than the other serotypes, and parainfluenza 3 triggers a steady increase of the basolaterally secreted cytokines. These findings are consistent with both the epidemiology and clinical findings in croup. The pathogenesis of laryngotracheobronchitis and laryngotracheobronchopneumonitis is similar to that described previously but with the extension of infection to the lower respiratory tract and usually the occurrence of secondary bacterial infection.

Many studies suggest that allergic factors play a role in recurrent croup. Welliver and associates found that children with croup caused by parainfluenza viral infections had titers of IgE-specific antibody in their nasopharyngeal secretions that were 3.6-fold higher than titers of similarly infected children with only upper respiratory tract infections. The children with croup had a cell-mediated immune response to parainfluenza virus that was 1.6-fold greater than that of the children with only upper respiratory tract infections.

In a subsequent review, investigators noted that an atopic disposition might be associated with the development of croup. Children with croup caused by a parainfluenza virus had specific IgE antibodies and released histamine into the airway more frequently than did control patients with parainfluenza viral upper respiratory tract infections. They also noted that children with recurrent croup had several atopic features, such as positive skin tests to environmental allergens, and were more likely to develop asthma when they grew older. Also, some children with a history of recurrent croup develop stridor with histamine challenge. Rennie and associates found that children with C/T genotype of the CD14 G-1359T SNP were more likely to have a history of croup, and children with genotype homozygous C/C had low prevalence of croup in this study. This suggests that CD14 polymorphisms may influence the innate response to viral infections associated with croup.

The above-mentioned studies by Welliver and associates and other reports make it easy to accept a role of atopy in recurrent croup. What is difficult to explain, however, is why many apparently primary infections result in only spasmodic croup. As one of us (JDC) suggested previously, initial sensitization may be parainfluenza virus group–specific and not type-specific. This idea suggests that early infection (primary infection) with parainfluenza virus type 3 would set the stage for spasmodic croup with parainfluenza viral types 1 and 2. The primary infection itself could have been mild owing to transplacentally acquired antibody.

Anatomy and Pathophysiology of Supraglottitis

The thin, elastic, leaflike epiglottic cartilage is attached to the anterior surface of the thyroid cartilage by the thyroepiglottic ligament ( Fig. 18.1 ). The hyoepiglottic ligament also provides support and anchors the epiglottis to the hyoid bone. The superior aspect of the epiglottis arches slightly posteriorly. Stratified squamous epithelium covers the anterior surface of the epiglottis and the superior third of the posterior portion; respiratory epithelium covers the remaining posterior surface. The stratified squamous epithelium is loosely adherent and creates a large potential space for the accumulation of inflammatory cells and edema fluid.

Anatomic relationships of the supraglottic larynx. Posterior view ( left ) and sagittal view ( right ).

The arytenoepiglottic folds arise from the epiglottis and terminate posteriorly near the paired arytenoid cartilages. These structures commonly are involved in the supraglottic infection and occasionally are the site of serious disease without epiglottitis per se. Immediately anterior to the epiglottis are the valleculae epiglotticae, where saliva pools before deglutition.

Supraglottic cellulitis with marked edema involving the epiglottis, arytenoepiglottic folds, ventricular bands, and arytenoids is the hallmark of this illness. As the edema increases, the epiglottis curls posteriorly and inferiorly. Inspiration tends to draw the inflamed supraglottic ring into the laryngeal inlet, whereas expiration is unopposed. This “ball-valve” mechanism is thought to produce slight hypoxia without hypercapnia. Diffuse infiltration with polymorphonuclear leukocytes, hemorrhage, edema, and fibrin deposition can be seen microscopically; this infiltration can progress to microabscesses, with H. influenzae type B occasionally seen in the tissue. Frank abscess formation has been documented in adults. Infection of the supraglottic larynx may spread inferiorly to involve the paraglottic space, but as a rule, neither upward extension into the laryngeal lymphatics nor downward extension into the subglottic region occurs.

Infection of the supraglottic structures probably arises from direct invasion by H. influenzae type B, with subsequent development of bacteremia. The bacteremia appears to be relatively short in duration and of low concentration, as suggested by several observations: (1) the serum concentration of the capsular polysaccharide is directly proportional to the concentration and duration of bacteremia, (2) children with epiglottitis have less H. influenzae type B capsular polysaccharide in their sera than do patients with meningitis, and (3) the density of H. influenzae type B in blood is significantly lower in patients with epiglottitis than in patients with meningitis. In 23 patients with epiglottitis, the geometric mean number of organisms was 123 colony-forming units (CFUs)/mL, whereas the geometric mean number in 43 patients with meningitis was 2203 CFU/mL ( P < .001).

What predisposes the epiglottis to infection is unknown. Possibly, mild trauma to the epiglottis occurs during food intake. This trauma could result in damage to the mucosal surface, which, in turn, could allow the invasion of organisms that already were present in the upper respiratory tract. Also possible is that a viral infection damages the mucosal surface and thus predisposes the patient to secondary bacterial infection.

Acute-phase sera in most children with epiglottitis lack specific H. influenzae type B bactericidal and hemagglutinating antibody. Seroconversion regularly occurs after infection.

Conflicting evidence exists relating to a possible genetic difference between patients with H. influenzae type B epiglottitis and other individuals. Whisnant and associates surveyed human leukocyte antigens (HLAs) and erythrocytic antigens among 30 children with epiglottitis and 20 patients with meningitis. HLA-A11 was found in 3% and 17% of the patients with epiglottitis and meningitis, respectively ( P < .01). HLA-B5 occurred in 3% and 13% of patients with meningitis and epiglottitis, respectively ( P < .05), whereas HLA-B40 occurred more often (23% vs. 10%) in patients with meningitis than in those with epiglottitis ( P < .05). Moreover, the frequency of HLA-A28 and HLA-B17 was higher among the patients with epiglottitis than in uninfected control subjects. However, the results of another study did not confirm these observations.

The distribution and frequency of MNS erythrocyte antigens in patients with epiglottitis may differ from those observed in others. For example, the NNSS genotype occurred in 6.4% of patients with epiglottitis and in 0.5% of healthy control subjects ( P < .0002), but this difference was not confirmed. However, the results of two studies suggest that the MNS genotype occurs less often in patients with H. influenzae type B meningitis than in patients with epiglottitis. In one study, white children with H. influenzae type B meningitis lacked G2m(n), an allotype antigen of immunoglobulin (Ig) G2 subclass heavy chains, more frequently than did control subjects. In another study in a white population in Finland, however, this difference was not noted. The frequency of the Km(1) immunoglobulin allotype in children with epiglottitis did not differ significantly from the prevalence of that marker in blacks or whites. However, in blacks with H. influenzae type B meningitis, the Km(1) marker occurred less frequently in patients than in control subjects. The identification of one outer membrane protein subtype of H. influenzae type B that was associated relatively infrequently with epiglottitis suggests that isotype-specific differences in the propensity of H. influenzae type B to cause epiglottitis may exist.

Clinical Presentation

Differential Diagnosis

The therapeutic approaches to the various acute obstructions in the region of the larynx vary markedly. Establishing a correct diagnosis is essential and frequently lifesaving. Table 18.3 lists the differential points of eight conditions with symptoms and signs of acute upper airway obstruction.

The most frequent serious differential diagnostic problem is the recognition of acute epiglottitis and its separation from the less fulminant laryngotracheitis. In epiglottitis (a rare disease in children today because of universal immunization with H. influenzae type B conjugate vaccines), the important differential points are lack of a croupy cough; the presence of a swollen, cherry-red epiglottis; the sitting posture of the child with the chin pushed forward and a reluctance or refusal to lie down; and the relatively greater apprehension and anxiety of the patient than the degree of chest retraction suggests. In contrast, a child with acute laryngotracheitis has a normal epiglottis on examination, always has a typical barking cough, is comfortable in a supine position, and frequently appears to have only minimal apprehension, despite retractions in which the sternum appears to be indenting 2 inches or more.

Early in the course of epiglottitis, the diagnosis can be confirmed only by the observation of the epiglottis, which can be done without difficulty. Later in the disease, the posture of the child and the history of rapidly progressing disease render the differential from laryngotracheitis readily apparent, so examination of the epiglottis directly (a dangerous procedure if the child is forced to lie down) or indirectly by a lateral neck radiograph rarely is indicated and usually is contraindicated. A study in adults found that bedside sonography could be used as a diagnostic tool for acute epiglottitis in the emergency department.

Laryngotracheobronchitis and laryngotracheobronchopneumonitis can be recognized by signs of lower respiratory involvement (rales, air trapping, wheezing, and pulmonary infiltrates on the radiograph). Bacterial disease should be suspected in laryngotracheobronchitis and laryngotracheobronchopneumonitis and when symptoms and signs become worse in laryngotracheitis. A lateral radiograph can be useful in the evaluation because it may reveal soft tissue densities within the trachea. Lateral neck and chest radiographs are regarded by many physicians as definitive tests to determine whether to rule out epiglottitis and laryngotracheitis. In a careful study by Stankiewicz and Bowes, the sensitivity and specificity of both radiographs were low, however.

Although a rare occurrence today, laryngeal diphtheria always should be considered and ruled out in croup. Recently Hsia and associates noted a child with Guillain-Barré syndrome who initially presented with a crouplike cough.

Spasmodic croup is rarely confused with acute laryngotracheitis, but a perusal of the literature indicates that the two entities most commonly are considered laryngotracheitis, which is unfortunate because prognostic considerations for the two entities are different. Spasmodic croup always is of sudden onset at night, occurs without fever, and is relieved by simple therapeutic modalities.

The possibility of a foreign body and angioneurotic edema always must be considered in upper airway obstructive disease. Differential points are presented in Table 18.3 . Rarely acute upper airway obstruction occurs in adolescents as a result of psychogenic and emotional factors.

There are many causes of recurrent croup in infants and children, and many of these cases should have operative evaluation by an otolaryngology service. Diagnostic possibilities include subglottic stenosis, laryngomalacia, bronchomalacia, tracheomalacia, vascular compression of distal trachea, stenosis of left main bronchus, subglottic cyst, tracheoesophageal fistula, unilateral vocal cord immobility, tracheal bronchus, and esophagitis and gastroesophageal reflux disease.

Specific Diagnosis in Croup Syndromes

The epidemiologic history frequently is an important factor in establishing a specific diagnosis. Obtaining bacterial culture specimens from the throat, laryngeal region, and blood is helpful in diagnosing epiglottitis and important in identifying laryngotracheitis, laryngotracheobronchitis, and laryngotracheobronchopneumonitis when secondary infection is suspected. The white blood cell count should be obtained because it can be helpful when secondary bacterial infection is considered.

A specific etiologic diagnosis can be made by the isolation of virus or its identification by a direct antigen test from a nasopharyngeal specimen. The diagnostic virologic facilities of many medical centers enable rapid identification of parainfluenza viruses, respiratory syncytial viruses, adenoviruses, most rhinoviruses, and influenza viruses to be established.

Specific Diagnosis in Supraglottitis

The clinical picture of sore throat, dysphagia, drooling, anxiety, and inspiratory distress without significant stridor and the characteristic sitting position should suggest the presumptive diagnosis in most cases. The definitive anatomic diagnosis is made by visualization of the epiglottis, and the etiologic diagnosis is confirmed by culture of an organism from the blood or the surface of the epiglottis.

In the typical case, the epiglottis is fiery red and greatly swollen. In children, the epiglottis can be seen by simple depression of the tongue with a tongue blade. In older children and adults, indirect or direct laryngoscopy usually is necessary to confirm the diagnosis. On occasion, the obstruction is caused by swelling of the ventricular bands and the arytenoepiglottic folds so that the epiglottis may appear relatively normal.

Major controversy exists concerning the safety of using a tongue depressor to examine a child with suspected epiglottitis because sudden cardiorespiratory arrest has been noted to occur. However, most instances of cardiorespiratory arrest occurred after the child was forced into a supine position rather than because of the examination itself. In many instances, patients with presumptive epiglottitis can be examined while they are in an upright position by using a tongue blade or indirect laryngoscopy.

Case management should be individualized. In the child with moderate or advanced disease, the clinical diagnosis should be apparent without having to do an intraoral examination. In this situation, intraoral examination should not be performed, but the child should be prepared for the establishment of an airway. This preparation should be rapid but controlled so that intubation can be performed in an operating room.

The diagnosis of epiglottitis can be established by the classic appearance on a lateral neck radiograph ( Fig. 18.2 ). However radiographic procedures could lead to a delay in providing the necessary definitive therapy. The use of the lateral neck radiograph should be reserved for subacute cases in which the specific diagnosis after completion of a clinical examination is not clear.

A lateral neck radiograph from a child with acute epiglottitis shows the swollen epiglottis (thumb sign) encroaching on the airway.

(Courtesy Dr. Ines Boechat.)

The lateral film of the neck, delineating the soft tissues and taken with the patient upright, gives the best view of the upper airway anatomy (see Fig. 18.2 ). The hypopharynx is dilated; normal cervical lordosis may be replaced by a straight or kyphotic contour. The valleculae are narrowed and may be obliterated. A thickened mass of tissue stretching from the valleculae to the arytenoids emphasizes the appropriateness of the term supraglottitis . In adults with epiglottitis, the widths of the epiglottis and arytenoepiglottic folds uniformly exceed 8 and 7 mm, respectively.

When performed, radiography of the neck in the anteroposterior projection usually reveals that tracheal narrowing is absent. However, some children with acute supraglottitis have localized subglottic narrowing indistinguishable from that found in acute laryngotracheitis.

All patients with suspected epiglottitis should have a blood culture, and a culture specimen should be obtained from the surface of the epiglottis when an artificial airway is established. Today, cultures are of increasing importance because of the change in epidemiology of H. influenzae type B infection and the resulting increased likelihood that a different organism may be causing the illness. A white blood cell count with a differential also may provide useful information. In children who have received antimicrobial treatment before cultures were obtained, performing a direct antigen test for H. influenzae type B on the blood and urine is worthwhile.