Haemophilus Species

Daniel M. Musher

General Concepts

Clinical Manifestations

Type b Haemophilus influenzae can cause meningitis, epiglottitis, bacteremia, and cellulitis. Nontypable H influenzae can cause otitias media, sinusitis, tracheobronchitis, and pneumonia. Other Haemophilus species and the syndromes they cause include H parainfluenzae (pneumonia and endocarditis), H ducreyi (genital chancre), and H aegyptius (conjunctivitis or Brazilian purpuric fever).

Structure, Classification, and Antigenic Types

Haemophilus species are Gram-negative coccobacilli similar in ultrastructural features to other pathogenic bacilli. Haemophilus influenzae requires hemin (factor X) and NAD+ (factor V) for growth. Other Haemophilus species require only NAD+ and therefore grow on blood agar. Typable H influenzae isolates are classified on the basis of seven antigenically distinct capsular polysaccharides; isolates lacking these polysaccharides are called nontypable.


Type b H influenzae colonizes the nasopharynx, and may penetrate the epithelium and capillary endothelium to cause bacteremia. Meningitis may result from direct spread via lymphatic drainage or from hematogenous spread. Nontypable H influenzae colonizes the nasopharynx and, to a lesser extent, the trachea and bronchi and may infect mucosa damaged by viral disease or cigarette smoking. Lipooligosaccharide is largely responsible for inflammation; exotoxins do not play a role.

Host Defenses

Serum antibody to the capsule (in the case of typable H influenzae) or to somatic antigens is bactericidal and promotes phagocytosis.


Haemophilus influenzae colonizes healthy children and adults (although the rate of colonization is far greater for nontypable than for type b H influenzae) and is spread by direct contact, secretions, and/or aerosol. Haemophilus ducreyi is spread by venereal contact. There is no animal reservoir for these organisms.


Respiratory secretions and cerebrospinal fluid must be cultured on chocolate agar. Blood cultures are positive in meningitis. Capsular antigen may be detected in cerebrospinal fluid for early identification if Gram stain is unsuccessful. Haemophilus ducreyi grows on Mueller-Hinton agar with 5 percent sheep blood in a CO2 enriched atmosphere.


Recommended treatment includes ampicillin for strains of H influenzae that do not make -lactamase and a third generation cephalosporin or chloramphenicol for strains that do. Ampicillin or amoxicillin together with a substance, such as clavulanic acid, that blocks the activity of -lactamase is also effective, but does not reliably treat meningitis. Tetracyclines remain effective in treating sinusitis or respiratory infection proven to be due to nontypable H influenzae. Use of polyribosyl ribitol phosphate (PRP) vaccine and, more recently, protein-conjugated PRP has vastly reduced the frequency of infection due to type b H influenzae.


The genus Haemophilus includes a number of species that cause a wide variety of infections but share a common morphology and a requirement for blood-derived factors during growth that has given the genus its name. Haemophilus influenzae, the major pathogen, can be separated into encapsulated or typable strains, of which there are seven types (a through f including e') based on the antigenic structure of the capsular polysaccharide, and unencapsulated or nontypable strains. Type b H influenzae is by far the most virulent organism in this group, commonly causing bloodstream invasion and meningitis in children younger than 2 years. Nontypable strains are frequent causes of respiratory tract disease in infants, children, and adults.

Other Haemophilus species cause disease less frequently. Haemophilus parainfluenzae sometimes causes pneumonia or bacterial endocarditis. Haemophilus ducreyi causes chancroid. Haemophilus aphrophilus is a member of the normal flora of the mouth and occasionally causes bacterial endocarditis. Haemophilus aegyptius, which causes conjunctivitis and Brazilian purpuric fever, and Haemophilus haemolyticus used to be separated on the basis of their ability to agglutinate or lyse red blood cells, but both are now included among the nontypable H influenzae strains.

Clinical Manifestations

Haemophilus species cause a variety of clinical syndromes (Fig. 30-1). Until the implementation of widespread vaccination programs, type b H influenzae was the most common cause of meningitis in children between the ages of 6 months and 2 years. In this situation, headache is followed rapidly by development of a stiff neck, with progression to coma and, in the absence of treatment, death. Emergent treatment reduces the incidence of, but does not eliminate, sequelae such as deafness and learning disabilities. Type b H influenzae also causes cellulitis and epiglottitis, a condition in which the epiglottitis becomes inflamed and swells, closing off the upper airway. Suffocation can be prevented in some cases only by performing a tracheostomy. Nontypable H influenzae strains commonly cause infection of the middle ear (otitis media), which manifests as an earache with fever in babies and young children. In adults, these organisms cause bronchitis and pneumonia, especially if some underlying disease of the bronchi and lungs is present. Nontypable H influenzae strains also commonly cause acute or chronic sinusitis in patients of all ages.

FIGURE 30-1 Clinical presentation of Haemophilus infections.

Chancroid is a venereal disease caused by H ducreyi. Lesions that resemble a syphilitic chancre result from sexual contact with an infected individual; they are usually found on the genitals. Unlike syphilitic chancres, the lesions are painful and are associated with a remarkable degree of swelling of lymph nodes in the inguinal area.

Structure, Classification, and Antigenic Types

Haemophilus species are Gram-negative coccobacilli that share common ultrastructural features with other Gram-negative bacilli. Their cell walls contain lipooligosaccharide, which resembles the lipopolysaccharide of Gram-negative bacilli but has shorter side chains (hence the designation oligosaccharide rather than polysaccharide). Haemophilus species have generally been thought not to make toxins or other extracellular products that account for their ability to produce infection. These organisms require hemin (factor X) and/or nicotinamide adenine dinucleotide (NAD+) (factor V) for growth. Whereas NAD+ is released into the medium by red blood cells and is available to the bacteria in blood agar, hemin is bound to red blood cells and is not released into the medium unless the cells are broken up, as in chocolate agar. Haemophilus influenzae requires both factors X and V; accordingly, it grows on chocolate agar but not on blood agar (Fig. 30-2), although it may appear on a blood agar plate as tiny satellite colonies around the colonies of other bacteria that have lysed red blood cells. Haemophilus parainfluenzae requires only factor V and therefore is able to grow on blood agar (however, recent reports suggest that many isolates identified as H parainfluenzae actually are H paraphrophilus). The long-prevailing notion that H ducreyi grows only in clotted rabbit blood has been dispelled by recent studies that show slow growth of this organism in Mueller-Hinton agar containing 5 percent sheep blood. All Haemophilus species grow more readily in an atmosphere enriched with CO2; H ducreyi and some nontypable H influenzae strains will not form visible colonies on culture plates unless grown in CO2-enriched atmosphere.

FIGURE 30-2 Growth of H influenzae requires both NAD+ (factor V) and hemin (factor X). Bacterial colonies occur only where both substances have diffused. (The hemin within intact erythrocytes is not accessible to the bacteria unless the erythrocytes are lysed by other bacteria or by heating the medium to make chocolate agar.)

Haemophilus influenzae strains are classified as either serotypable (if they display a capsular polysaccharide antigen) or nontypable (if they lack a capsule). The word "type" as applied to H influenzae refers to this serotyping scheme. There are six generally recognized types: a, b, c, d, e, and f. A seventh type has been designated e' because its polysaccharide is closely related to that of type e. Antiserum to type e' H influenzae is not routinely available. These types may be identified by an agglutination reaction that uses antisera raised in rabbits; with this method, however, cross-reactions with somatic antigens may cause nontypable strains to be designated erroneously as typable. This kind of error is eliminated by using counterimmunoelectrophoresis, in which migration under an electric current removes somatic (protein) antigens from the reaction, leaving only capsular polysaccharides to react with antibody.

The presence of a polyribosyl ribitol phosphate (PRP) capsule is an important virulence factor: it renders type b H influenzae resistant to phagocytosis by polymorphonuclear leukocytes in the absence of specific anticapsular antibody. Susceptibility to the bactericidal effect of serum depends on the presence of antibodies to a number of antigenic sites, including the lipooligosaccharide or outer membrane proteins designated as P1 and P2. Type b H influenzae is plainly the most virulent of the Haemophilus species; 95 percent of bloodstream and meningeal Haemophilus infections in children are due to this organisms. In contrast, in adults, nontypable strains of H influenzae are the most common cause of Haemophilus infection, presumably because most adults have acquired antibody to PRP.

The relative place of H influenzae biogroup aegyptius, the cause of Brazilian purpuric fever, remains to be determined. Most, but not all, strains that cause this syndrome contain a unique 24 megadalton plasmid and a 79 kilodalton outer membrane protein, either or both of which may mediate virulence. These organisms also have at least some of the genetic material that codes for encapsulation of type b H influenzae.

By using a series of biochemical reactions, H influenzae also may be classified into six biotypes designated I through VI. Most type b H influenzae strains fall into biotypes I or II, whereas most non typable H influenzae strains fall into biotypes II through VI. Several interesting clinical correlations have recently emerged. Biotype I isolates appear to have a predilection for causing pneumonia. Nearly all genital isolates, as well as bloodstream isolates from infected neonates or from women with puerperal sepsis, are biotype IV. In addition, biotype III, which agglutinates red blood cells in vitro and includes H influenzae biogroup aegyptius, has been implicated as a common cause of conjunctivitis. There is no explanation for these clinically observed associations.


The pathogenesis of H influenzae infections is not completely understood, but the presence of the type b polysaccharide capsule is a major factor in virulence. Encapsulated organisms can penetrate the epithelium of the nasopharynx and invade blood capillaries directly. Nontypable strains are less invasive, but they, as well as typable strains, induce an inflammatory response that causes disease; production of exotoxins is not thought to play a role in pathogenicity. Nontypable H influenzae strains colonize the nasopharynx of most normal individuals, but type b H influenzae strains are found in only 1 to 2 percent of normal children. Outbreaks of type b infection occur, especially in nurseries and child care centers; prophylactic administration of antibiotics may be used. Vaccination with type b polysaccharide appears to be effective in preventing infection, and vaccines are now available for routine use.


The pathogenesis of meningitis due to type b H influenzae has been well studied. These organisms colonize the nasopharynx and spread from one human to another by direct contact or via secretions and/or aerosol. They penetrate epithelial layers and capillary endothelium by unknown mechanisms, reaching the meninges either directly via lymphatic drainage from the nasopharynx or indirectly by causing bacteremia with subsequent seeding of the highly vascular choroid plexus.

Most cases of H influenzae meningitis in adults are due to nontypable strains. The pathogenesis of these infections differs from that of type b H influenzae. Nontypable strains are unencapsulated and therefore less virulent, and they are unable to penetrate directly into capillaries. Rather, they gain entry to the central nervous system by direct extension, often associated with infection of the sinuses or middle ear and/or with trauma involving the sinuses or skull. Thus, about 50 percent of adults with H influenzae meningitis have a history of prior head trauma with or without a documented cerebrospinal fluid leak, and another 25 percent have chronic otitis media. Also, H influenzae is second only to Streptococcus pneumoniae as a cause of recurring meningitis, an unusual syndrome attributed to a connection between the sinuses and the subarachnoid space, usually via a tear in the dura. The clinical picture of meningitis caused by typable or nontypable H influenzae is similar to that caused by other bacteria, such as S pneumoniae. Since Haemophilus species do not produce substances that obviously damage mammalian tissues, bacterial replication is probably the usual pathway for disease production, with triggering of the complement cascade by classic and alternative pathways, followed by accumulation of inflammatory cells.

Cellulitis and Epiglottitis

Cellulitis and epiglottitis are discussed together because their pathogenesis is probably quite similar. Both are due to type b H influenzae, are likely to cause associated bacteremia, and occur more frequently in children than adults. Epiglottitis can be regarded as a cellulitis of the relatively loose submucosal connective tissues of the epiglottis. In this syndrome, a sore throat rapidly progresses to difficulty in breathing, stridor, obstruction of the air ways, and respiratory arrest. Local extension from the colonized nasopharynx through soft tissues is probably responsible for epiglottitis. Cellulitis often involves the face or neck. It sometimes seems to start at the buccal mucosa and extend outward, supporting the idea that it also results from local extension. The often-repeated teaching that facial cellulitis due to Haemophilus causes a distinctive bluish tinge enabling it to be distinguished from cellulitis caused by other bacteria, defies reason and is best ignored.

Respiratory Disease

Nontypable H influenzae is a major pathogen that colonizes the human respiratory tract. Adherence of bacteria to mammalian tissues, which is mediated by pili (fimbriae), is thought to be an important precursor to colonization, and infection of the upper airways is associated with the presence of pili. Respiratory infections caused by these organisms include sinusitis, otitis media, acute tracheobronchitis, and pneumonia.

Purulent material aspired from acutely infected paranasal sinuses in children or adults or from behind an infected tympanic membrane in babies and young children commonly contains nontypable H influenzae. Studies of outer membrane protein profiles have shown that middle ear and nasopharyngeal isolates are identical, supporting the notion that colonization of the eustachian tube, followed by obstruction and infection, is probably responsible. Repeated bouts of otitis media are thought to be due to different strains; each infection may be associated with emergence of antibody to distinctive surface proteins. The decreasing frequency of otitis media with age is due in part to anatomic changes and in part to immunity to H influenzae.

The situation is somewhat more complex for bronchopulmonary disease. Nontypable H influenzae is found in the nasopharynx and in sputum cultures of nearly one-half of adults with chronic bronchitis. Not surprisingly, this organism is also recovered from the large airways via bronchoscopy, since upper-airway bacteria are carried along by the bronchoscope. On the basis of these observations, as well as of other observations on the presence of antibody in serum samples of patients with chronic bronchitis, some British investigators concluded that H influenzae actually plays a causative role in what they call chronic bronchitisour chronic obstructive pulmonary disease. With the use of semiquantitative techniques, H influenzae have been shown to be very scarce in the sputum of patients with chronic bronchitis who are in a stable clinical state (and perhaps results from oropharyngeal contamination of the specimen), but that, at least in some patients, large numbers are present during an exacerbation. This observation fits with studies in which transtracheal aspiration has revealed the trachea to be free of Haemophilus isolates in stable patients with emphysema, but to contain these organisms in some patients who have purulent sputum.

In any case, nontypable H influenzae is certainly a prominent cause of acute tracheobronchitis or pneumonia in patients who have underlying chronic bronchitis, emphysema, or obstructive pulmonary disease. Other debilitating diseases such as malnutrition, lung cancer, and alcoholism are also often present. Symptoms of acute bronchitis include increased shortness of breath, cough, and production of purulent sputum; in more severe cases, fever and an increased white blood cell count may also be present. Pneumonia may result, with patchy or segmental pulmonary infiltrates detectable by radiography. Gram stain of sputum reveals a profusion of Gram-negative coccobacilli (Fig. 30-3), but no other pathogenic bacteria. A properly plated specimen yields a nearly pure growth of Haemophilus. When S pneumoniae organisms are present in small numbers (for example, 1 S pneumoniae colony to 100 Haemophilus colonies), it is less certain that Haemophilus is the sole pathogen. Blood cultures are positive in 10 to 15 percent of patients with Haemophilus pneumonia. Nontypable H influenzae is second only to S pneumoniae as the cause of bacterial pneumonia in middle-aged men. Although type b H influenzae is more virulent, pneumonia due to this organism is much less common, probably because of its vastly lower incidence of colonization. Compared with nontypable H influenzae pneumonia, the underlying pulmonary disease may not be so prominent. In type b H influenzae disease, the onset is more acute and blood cultures are more likely to be positive. Haemophilus influenzae received its name because it was first isolated from the lungs of individuals who died during an epidemic of influenza virus infection in 1890. It is not possible to determine whether bacterial infection was due to typable or nontypable isolates.

FIGURE 30-3 Gram-stained sputum showing profuse Gram-negative coccobacilli with no other bacterial forms present (original magnification X 440). Cultures showed nontypable H influenzae as the overwhelmingly predominant isolate. Quantitative culture showed 7 X 108 CFU/ml of sputum.

Brazilian Purpuric Fever

In the past few years, a syndrome of fulminating illness with substantial mortality characterized by nausea, vomiting, hemorrhagic skin lesions, fever, prostration, and shock has been recognized under the name Brazilian purpuric fever. Haemophilus influenzae biogroup aegyptius can be cultured from the blood of affected patients. Many have had a history of conjunctivitis in the weeks preceding onset of the disease. As noted above, most of the responsible organisms differ from other H influenzae biogroup aegyptius strains that only cause conjunctivitis in having a unique plasmid and a 79 kD outer membrane protein, as well as containing genetic material that hybridizes with the capsular locus of H influenzae.


Type b H influenzae use to be a relatively common cause of septic arthritis in children and results from hematogenous dissemination. Interestingly, this organism only rarely caused osteomyelitis; the reasons for this discrepancy are unknown.

Nontypable H influenzae biotype IV tends to colonize the female genital tract and may cause puerperal fever and/or neonatal sepsis. Infection in the mother is relatively mild, but it may be fulminating in the newborn infant. It is unknown whether nontypable H influenzae biotype IV has any special virulence beyond the factors that promote adherence to vaginal epithelial cells.

Some patients have bacteremia due to type b H influenzae without an apparent focus of infection. There is good evidence that if this condition is left untreated in children, some source for the infection (e.g., meningitis) will become apparent within 24 to 48 hours. However, in this situation one cannot be certain that the meninges were not seeded secondarily to the bacteremia. In adults, there appears to be a syndrome of bacteremia due to nontypable H influenzae for which no focus ever becomes apparent.

Host Defenses

For many years it was believed that bactericidal antibody directed against PRP capsule of type b H influenzae was entirely responsible for host resistance to infection. However, more recent studies have stressed a role for antibody to somatic antigens as well. For example, antibody to PRP can often be detected in the sera of children on admission to the hospital with sepsis due to type b H influenzae. In addition, adsorption of immune serum with PRP alone does not remove its protective capabilities, whereas adsorption with whole organisms does. Finally, immunization with ribosomes is protective in animal models of infection. Separation of the outer membrane of type b H influenzae into its many protein constituents by polyacrylamide gel electrophoresis (PAGE) combined with analysis of antibody responses during infection has suggested that antibody to any of a number of individual membrane proteins may be associated with immunity. Bactericidal antibodies that react with individual outer membrane proteins or with lipooligosaccharide constituents have been identified. These findings support, on a molecular basis, the potential importance of antibody to noncapsular antigens in immunity to type b H influenzae infection. Opsonizing antibody may also play a role in protection and may be directed against PRP or somatic constituents (Fig. 30-4).

FIGURE 30-4 Macrophage or polymorphonuclear leukocyte phagocytosing H influenzae coated with antibodies specific for the capsule and somatic antigen.

Recent studies of nontypable H influenzae strains have shown that bactericidal antibody to outer membrane proteins develops in infants in response to otitis media caused by these organisms. Normal adults generally have both bactericidal and opsonizing antibodies directed against nontypable H influenzae. Although levels of opsonizing antibody may be low in adults who develop acute nontypable H influenzae infection, substantial levels of bactericidal activity are present in serum at the time infection is diagnosed. It is not clear why this should occur. In some instances a blocking effect by secretory IgA in bronchial secretions might be responsible. Alternatively, the extensive structural damage to the bronchi and lungs that predisposes to serious nontypable H influenzae infection may allow proliferation of the bacteria unchecked by normal serum defense mechanisms.


Haemophilus organisms spread directly among individuals without a known contribution from environmental sources or animal reservoirs. Nontypable H influenzae strains are found in the nasopharynx of many healthy subjects, depending upon the frequency and intensity with which they are sought. By contrast, type b H influenzae is found only in 1 to 2 percent of healthy children, and its spread to previously uncolonized children in the early years is associated with a substantially increased risk of infection. Families and day care centers are important sources for dissemination of these organisms.


Haemophilus influenzae meningitis cannot be distinguished on the basis of clinical presentation, physical examination, or cerebrospinal fluid abnormalities from meningitis due to other common bacterial pathogens. The cerebrospinal fluid in untreated patients contains an average of 2 X 107 bacteria/ml, so that microscopic examination, especially in the absence of prior antibiotic therapy, should reveal the infecting organisms. Detection of capsular material in the cerebrospinal fluid by counter immunoelectrophoresis is helpful in cases in which the Gram stain is not conclusive; this technique is especially important in patients who have received enough antibiotic to suppress the growth of organisms in cultures of cerebrospinal fluid, but not enough to be curative.

The bacteriologic diagnosis of pneumonia or acute febrile purulent tracheobronchitis due to H influenzae is made by finding myriad small, somewhat pleomorphic, Gram-negative coccobacilli in Gram-stained sputum (Fig. 30-3) and by culturing H influenzae as the overwhelmingly predominant isolate; the mean number of viable organisms per milliliter of infected sputum is about 5 X 108. Blood cultures may be positive in 10 to 15 percent of patients with pneumonia and are negative in those with acute febrile tracheobronchitis.

Endocarditis due to H parainfluenzae tends to be associated with large vegetations that embolize to large arteries such as femoral or carotid, causing a limb to turn blue and cool, or producing a stroke. The etiologic diagnosis of endocarditis is, of course, established by blood culture. Recent studies have suggested that many isolates previously identified as H parainfluenzae are, in fact, H paraphrophilus. Chancres due to H ducreyi are tender, somewhat irregular, and slightly indurated; they may be confused with primary syphilitic chancres, traumatic lesions of the penis (especially with bacterial superinfection), fixed drug eruptions, or ulcerated herpetic lesions. The diagnosis is established by culturing the causative organism on Mueller-Hinton agar supplemented with 5 percent sheep blood and incubating it for 96 hours in a CO2-enriched atmosphere.


Outbreaks of serious infection due to type b H influenzae can be prevented by vaccination or prophylactic therapy. Initial trials of vaccination with type b H influenzae PRP were disappointing, because this polysaccharide in its pure form is not immunogenic in infants, the group most at risk of infection. Later studies showed that injection of PRP conjugated to a protein, such as diphtheria toxoid, that serves as an adjuvant results in good antibody responses in infants. Clinical trials with these vaccines have been successful, and preparation of PRP linked to outer membrane proteins or ribosomes are currently in widespread use.

Once an outbreak of type b H influenzae infection has been documented, infants and toddlers who are in intimate contact with colonized or infected individuals have a greatly increased, albeit still small, likelihood of developing serious infection. The use of rifampin prophylaxis to prevent or eradicate nasopharyngeal colonization has been recommended. This measure is controversial, however, because if widely applied it might encourage the emergence of rifampin-resistant organisms, and also because the cost to prevent each potential case of meningitis is high.

The mainstay of therapy for H influenzae infection used to be ampicillin, since isolates were uniformly susceptible to 0.5 g/ml. In the last few years an increasing proportion of H influenzae isolates have produced -lactamase. In most medical centers, 25 to 30 percent of type b isolates and a somewhat smaller percentage of nontypable isolates are now resistant to penicillin or ampicillin; in some centers, 50 to 60 percent of type b H influenzae isolates are ampicillin resistant. Very rarely, an isolate resists ampicillin but does not produce -lactamase; decreased penetration into the bacterium is thought to be responsible. Treatment with a combination of amoxicillin and clavulanic acid (a substance that covalently binds -lactamase) is effective against -lactamase-producing strains, but has not been recommended for treating meningitis. Chloramphenicol was long considered the drug of choice for meningitis caused by a penicillin-resistant H influenzae strain, and it is still highly effective. Third-generation cephalosporins, such as ceftriaxone or cefotaxime, are effective against H influenzae and penetrate the meninges well; these drugs are useful in treating H influenzae meningitis. The addition of corticosteroids may reduce the incidence of complications such as deafness.

In addition to the above-named drugs, tetracycline and sulfa drugs are effective in treating upper and lower respiratory infections caused by Haemophilus. Erythromycin should not generally be used to treat H influenzae infections; many isolates are resistant, and documentation of susceptibility in routine clinical laboratories is subject to error.

The spread of soft chancre due to H ducreyi is best prevented by use of a condom during sexual intercourse. Two-thirds of H ducreyi isolates produce -lactamase. All isolates are susceptible in vitro to erythromycin, and excellent clinical results have been obtained.


Doern GV, Jones RN: Antimicrobial susceptibility testing of Haemophilus influenzae,

Branhamella catarrhalis, and Neisseria gonorrhoeae. Antimicrob Agents Chemother 32: 1747, 1988

Eskola J, Peltola H, Takala AK et al: Efficacy of Haemophilus influenzae type b polysaccharide-diphtheria toxoid conjugate vaccine in infancy. N Engl J Med 317:717, 1987

Groeneveld K, van Alphen L, Eijk PP, et al: Endogenous and exogenous reinfections by Haemophilus influenzae in patients with chronic obstructive pulmonary disease: The effect of antibiotic treatment on persistence. J Infect Dis 161:512, 1990

Hammond GW, Slutchuk M, Scatliff J et al: Epidemiologic, clinical, laboratory, and therapeutic features of an urban outbreak of chancroid in North America. Rev Infect Dis 2:867. 1980

Harabuchi Y, Faden H, Yamanaka N, et al: Nasopharyngeal colonization with nontypable Haemophilus influenzae and recurrent otitis media. J Infect Dis 170:862, 1994

Harrison LH, de Silva GA, Pittman M et al: Epidemiology and clinical spectrum of Brazilian purpuric fever. J Clin Microbiol 27:599, 1989

Jorgensen JH, Doern GV, Maher LA, et al: Antimicrobial resistance among respiratory isolates of Haemophilus influenzae, Moraxella catarrhalis, and Streptococcus pneumoniae in the United States. Antimicrob Agents and Chemother 34(11):2075, 1990

Lesse AJ, Gheesling LL, Bittner WE, et al: Stable, conserved outer membrane epitope of strains of Haemophilus influenzae biogroup aegyptius associated with Brazilian Purpuric Fever. Infect and Immun, p. 1351, 1992

Mason EO, Jr, Kaplan SL, Lamberth LB et al: Serotype and ampicillin susceptibility of Haemophilus influenzae causing systemic infections in children: 3 years of experience. J Clin Microbiol 15:543, 1982

Murphy TF, Apicella MA: Nontypable Haemophilus influenzae: a review of clinical aspects, surface antigens, and the human immune response to infection. Rev Infect Dis 9:1, 1987

Murphy TF, Berstein JM, Dryja DM et al: Outer membrane protein and lipooligosaccharide analysis of paired nasopharyngeal and middle ear isolates in otitis media due to nontypable Haemophilus influenzae: pathogenic and epidemiological observations. J Infect Dis 156:723, 1987

Musher D, Goree A, Murphy T et al: Immunity to Haemophilus influenzae type b in young adults: correlation of bactericidal and opsonizing activity of serum with antibody to polyribosylribitol phosphate and lipooligosaccharide before and after vaccination. J Infect Dis 154:935, 1986

Musher DM, Kubitschek KR, Crennan J et al: Pneumonia and acute febrile tracheobronchitis due to Haemophilus influenzae. Ann Intern Med 99:444, 1983

Osterholm MT, Pierson LM, White KE et al: The risk of subsequent transmission of Haemophilus influenzae type b disease among children in day care. N Engl J Med 316:1, 1987

Sell SH, Wright PF (eds): Haemophilus influenzae: Epidemiology, Immunology, and Prevention of Disease. Elsevier Biomedical, New York, 1982 St. Geme JW,III, Falkow S: Infect and Immun, p. 4036, 1990

Wallace RJ, Jr, Baker CJ, Quinones FJ et al: Nontypable Haemophilus influenzae (biotype 4) as a neonatal, maternal, and genital pathogen. Rev Infect Dis 5:123, 1983