CHAPTER 57: PERTUSSIS AND OTHER BORDETELLA INFECTIONS

Pertussis is an acute infection of the respiratory tract caused by Bordetella pertussis. The name pertussis means “violent cough,” which aptly describes the most consistent and prominent feature of the illness. The inspiratory sound made at the end of an episode of paroxysmal coughing gives rise to the common name for the illness, “whooping cough.” However, this feature is variable: it is uncommon among infants ≤6 months of age and is frequently absent in older children and adults. The Chinese name for pertussis is “the 100-day cough,” which accurately describes the clinical course of the illness. The identification of B. pertussis was first reported by Bordet and Gengou in 1906, and vaccines were produced over the following two decades.

Of the 10 identified species in the genus Bordetella, only four are of major medical significance. B. pertussis infects only humans and is the most important Bordetella species causing human disease. B. parapertussis causes an illness in humans that is similar to pertussis but is typically milder; co-infections with B. parapertussis and B. pertussis have been documented. With improved polymerase chain reaction (PCR) diagnostic methodology, up to 20% of patients with a pertussis-like syndrome have been found to be infected with B. holmesii, formerly thought to be an unusual cause of bacteremia. B. bronchiseptica is an important pathogen of domestic animals that causes kennel cough in dogs, atrophic rhinitis and pneumonia in pigs, and pneumonia in cats. Both respiratory infection and opportunistic infection due to B. bronchiseptica are occasionally reported in humans. B. petrii, B. hinzii, and B. ansorpii have been isolated from patients who are immunocompromised.

Bordetella species are gram-negative pleomorphic aerobic bacilli that share common genotypic characteristics. B. pertussis and B. parapertussis are the most similar of the species, but B. parapertussis does not express the gene coding for pertussis toxin. B. pertussis is a slow-growing fastidious organism that requires selective medium and forms small, glistening, bifurcated colonies. Suspicious colonies are presumptively identified as B. pertussis by direct fluorescent antibody testing or by agglutination with species-specific antiserum. B. pertussis is further differentiated from other Bordetella species by biochemical and motility characteristics.

B. pertussis produces a wide array of toxins and biologically active products that are important in its pathogenesis and in immunity. Most of these virulence factors are under the control of a single genetic locus that regulates their production, resulting in antigenic modulation and phase variation. Although these processes occur both in vitro and in vivo, their importance in the pathobiology of the organism is unknown; they may play a role in intracellular persistence and person-to-person spread. The organism’s most important virulence factor is pertussis toxin, which is composed of a B oligomer–binding subunit and an enzymatically active A protomer that ADP-ribosylates a guanine nucleotide-binding regulatory protein (G protein) in target cells, producing a variety of biologic effects. Pertussis toxin has important mitogenic activity, affects the circulation of lymphocytes, and serves as an adhesin for bacterial binding to respiratory ciliated cells. Other important virulence factors and adhesins are filamentous hemagglutinin, a component of the cell wall, and pertactin, an outer-membrane protein. Fimbriae, bacterial appendages that play a role in bacterial attachment, are the major antigens against which agglutinating antibodies are directed. These agglutinating antibodies have historically been the primary means of serotyping B. pertussis strains. Other virulence factors include tracheal cytotoxin, which causes respiratory epithelial damage; adenylate cyclase toxin, which impairs host immune-cell function; dermonecrotic toxin, which may contribute to respiratory mucosal damage; and lipooligosaccharide, which has properties similar to those of other gram-negative bacterial endotoxins.

Pertussis is a highly communicable disease, with attack rates of 80–100% among unimmunized household contacts and 20% within households in well-immunized populations. The infection has a worldwide distribution, with cyclical outbreaks every 3–5 years (a pattern that has persisted despite widespread immunization). Pertussis occurs in all months; however, in North America, its activity peaks in summer and autumn.

In developing countries, pertussis remains an important cause of infant morbidity and death. The reported incidence of pertussis worldwide has decreased as a result of improved vaccine coverage (Fig. 57-1). However, coverage rates are still <50% in many developing nations; the World Health Organization (WHO) estimates that 90% of the burden of pertussis is in developing regions. In addition, overreporting of immunization coverage and underreporting of disease result in substantial underestimation of the global burden of pertussis. The WHO estimates that there were 195,000 deaths from pertussis among children in 2008.

Before the institution of widespread immunization programs in the developed world, pertussis was one of the most common infectious causes of morbidity and death. In the United States before the 1940s, between 115,000 and 270,000 cases of pertussis were reported annually, with an average yearly rate of 150 cases per 100,000 population. With universal childhood immunization, the number of reported cases fell by >95%, and mortality rates decreased even more dramatically. Only 1010 cases of pertussis were reported in 1976 (Fig. 57-2). After that historic low, rates of pertussis slowly increased. In recent years, pertussis epidemics have been reported with increasing frequency worldwide. The United States experienced widespread outbreaks of pertussis in 2005, 2010, and 2012 at levels not seen in 40–50 years (>40,000 reported cases in 2012).

Although thought of as a disease of childhood, pertussis can affect people of all ages and is increasingly being identified as a cause of prolonged coughing illness in adolescents and adults. In unimmunized populations, pertussis incidence peaks during the preschool years, and well over half of children have the disease before reaching adulthood. In highly immunized populations such as those in North America, the peak incidence is among infants <1 year of age who have not completed the three-dose primary immunization series. An increase in pertussis incidence among adolescents and adults began in the late 1990s and led to the introduction of an adolescent booster across North America by 2006. While the disease burden among adolescents has started to decrease, children 7–10 years of age have recently emerged as a high-risk group. In major outbreaks in 2010 and 2012, the incidence of pertussis among children 10 years of age, most of whom were fully immunized, was as high as that among infants <6 months of age. Although adults contribute a smaller proportion of reported cases of pertussis than do children and adolescents, this difference may be related to a greater degree of underrecognition and underreporting. A number of studies of prolonged coughing illness suggest that B. ­pertussis may be the etiologic agent in 12–30% of adults with cough that does not improve within 2 weeks. In one study of the efficacy of an acellular pertussis vaccine in adolescents and adults, the incidence of pertussis in the placebo group was 3.7–4.5 cases per 1000 person-years. Although this prospective cohort study yielded a lower estimate than the studies of cough illness, its results still translate to 600,000–800,000 cases of pertussis annually among adults in the United States.

Severe morbidity and high mortality rates, however, are restricted almost entirely to infants. In Canada, there were 16 deaths from pertussis between 1991 and 2001; all those who died were infants ≤6 months of age. Similarly, in the United States between 1993 and 2004, all pertussis deaths and 86% of hospitalizations for pertussis involved infants ≤3 months of age. Although school-age children are the source of infection for most households, adults are the likely source for cases in high-risk infants and may serve as the reservoir of infection between epidemic years.

Infection with B. pertussis is initiated by attachment of the organism to the ciliated epithelial cells of the nasopharynx. Attachment is mediated by surface adhesins (e.g., pertactin and filamentous hemagglutinin), which bind to the integrin family of cell-surface proteins, probably in conjunction with pertussis toxin. The role of fimbriae in adhesion and in maintenance of infection has not been fully delineated. At the site of attachment, the organism multiplies, producing a variety of other toxins that cause local mucosal damage (tracheal cytotoxin, dermonecrotic toxin). Impairment of host defense by B. pertussis is mediated by pertussis toxin and adenylate cyclase toxin. There is local cellular invasion, with intracellular bacterial persistence; however, systemic dissemination does not occur. Systemic manifestations (lymphocytosis) result from the effects of the toxins.

The pathogenesis of the clinical manifestations of pertussis is poorly understood. It is not known what causes the hallmark paroxysmal cough. A pivotal role for pertussis toxin has been proposed. Proponents of this position point to the efficacy of preventing clinical symptoms with a vaccine containing only pertussis toxoid. Detractors counter that pertussis toxin is not the critical factor because paroxysmal cough also occurs in patients infected with B. parapertussis, which does not produce pertussis toxin. It is thought that neurologic events in pertussis, such as seizures and encephalopathy, are due to hypoxia from coughing paroxysms or apnea rather than to the effects of specific bacterial products. B. pertussis pneumonia, which occurs in up to 10% of infants with pertussis, is usually a diffuse bilateral primary infection. In older children and adults with pertussis, pneumonia is often due to secondary bacterial infection with streptococci or staphylococci. Deaths from pertussis among young infants are frequently associated with very high levels of leukocytosis and pulmonary hypertension.

Both humoral and cell-mediated immunity are thought to be important in pertussis. Antibodies to pertussis toxin, filamentous hemagglutinin, pertactin, and fimbriae are all protective in animal models. Pertussis agglutinins were correlated with protection in early studies of whole-cell pertussis vaccines. Serologic correlates of protection conferred by acellular pertussis vaccines have not been established, although antibody to pertactin, fimbriae, and (to a lesser degree) pertussis toxin correlated best with protection in two efficacy trials. The duration of immunity after whole-cell pertussis vaccination is short-lived, with little protection remaining after 10–12 years. Recent studies have demonstrated early waning of immunity—i.e., within 2–4 years after the fifth dose of acellular pertussis vaccine in children who received acellular pertussis vaccine for their primary series in infancy. These data suggest that boosters may be needed more frequently than every 10 years, as previously thought. Although immunity after natural infection was thought to be lifelong, seroepidemiologic evidence demonstrates that it clearly is not and that subsequent episodes of clinical pertussis are prevented by intermittent subclinical infection.

Pertussis is a prolonged coughing illness with clinical manifestations that vary by age (Table 57-1). Although not uncommon among adolescents and adults, classic pertussis is most often seen in preschool and school-age children. After an incubation period averaging 7–10 days, an illness develops that is indistinguishable from the common cold and is characterized by coryza, lacrimation, mild cough, low-grade fever, and malaise. After 1–2 weeks, this catarrhal phase evolves into the paroxysmal phase: the cough becomes more frequent and spasmodic with repetitive bursts of 5–10 coughs, often within a single expiration. Posttussive vomiting is frequent, with a mucous plug occasionally expelled at the end of an episode. The episode may be terminated by an audible whoop, which occurs upon rapid inspiration against a closed glottis at the end of a paroxysm. During a spasm, there may be impressive neck-vein distension, bulging eyes, tongue protrusion, and cyanosis. Paroxysms may be precipitated by noise, eating, or physical contact. Between attacks, the patient’s appearance is normal but increasing fatigue is evident. The frequency of paroxysmal episodes varies widely, from several per hour to 5–10 per day. Episodes are often worse at night and interfere with sleep. Weight loss is not uncommon as a result of the illness’s interference with eating. Most complications occur during the paroxysmal stage. Fever is uncommon and suggests bacterial superinfection.

After 2–4 weeks, the coughing episodes become less frequent and less severe—changes heralding the onset of the convalescent phase. This phase can last 1–3 months and is characterized by gradual resolution of coughing episodes. For 6–12 months, intercurrent viral infections may be associated with a recrudescence of paroxysmal cough.

Not all individuals who develop pertussis have classic disease. The clinical manifestations in adolescents and adults are more often atypical. In a German study of pertussis in adults, more than two-thirds had paroxysmal cough and more than one-third had whoop. Adult illness in North America differs from this experience: the cough may be severe and prolonged but is less frequently paroxysmal, and a whoop is uncommon. Vomiting with cough is the best predictor of pertussis as the cause of prolonged cough in adults. Other predictive features are a cough at night, sweating episodes between paroxysms of coughing, and exposure to other individuals with a prolonged coughing illness.

Complications are frequently associated with pertussis and are more common among infants than among older children or adults. Subconjunctival hemorrhages, abdominal and inguinal hernias, pneumothoraces, and facial and truncal petechiae can result from increased intrathoracic pressure generated by severe fits of coughing. Weight loss can follow decreased caloric intake. In a series of more than 1100 children <2 years of age who were hospitalized with pertussis, 27.1% had apnea, 9.4% had pneumonia, 2.6% had seizures, and 0.4% had encephalopathy; 10 children (0.9%) died. Pneumonia is reported in <5% of adolescents and adults and increases in frequency after 50 years of age. In contrast to the primary B. pertussis pneumonia that develops in infants, pneumonia in adolescents and adults with pertussis is usually caused by a secondary infection with encapsulated organisms such as Streptococcus pneumoniae or Haemophilus influenzae. Pneumothorax, severe weight loss, inguinal hernia, rib fracture, carotid artery aneurysm, and cough syncope have all been reported in adolescents and adults with pertussis.

If the classic symptoms of pertussis are present, clinical diagnosis is not difficult. However, particularly in older children and adults, it is difficult to differentiate infections caused by B. pertussis and B. para-pertussis from other respiratory tract infections on clinical grounds. Therefore, laboratory confirmation should be attempted in all cases. Lymphocytosis (an absolute lymphocyte count of >10⁸–10⁹/L) is common among young children, in whom it is unusual with other infections, but not among adolescents and adults. Culture of nasopharyngeal secretions remains the gold standard of diagnosis, although DNA detection by PCR has replaced culture in many laboratories because of increased sensitivity and quicker results. Appropriate PCR methodology must include primers to differentiate among B. pertussis, B. parapertussis, and B. holmesii. The best specimen is collected by nasopharyngeal aspiration, in which a fine flexible plastic catheter attached to a 10-mL syringe is passed into the nasopharynx and withdrawn while gentle suction is applied. Since B. pertussis is highly sensitive to drying, secretions for culture should be inoculated without delay onto appropriate medium (Bordet-Gengou or Regan-Lowe), or the catheter should be flushed with a phosphate-buffered saline solution for culture and/or PCR. An alternative to the aspirate is a Dacron or rayon nasopharyngeal swab; again, inoculation of culture plates should be immediate or an appropriate transport medium (e.g., Regan-Lowe charcoal medium) should be used. Results of PCR can be available within hours; cultures become positive by day 5 of incubation. B. pertussis and B. parapertussis can be differentiated by agglutination with specific antisera or by direct immunofluorescence.

Nasopharyngeal cultures in untreated pertussis remain positive for a mean of 3 weeks after the onset of illness; these cultures become negative within 5 days of the institution of appropriate antimicrobial therapy. The duration of a positive PCR in untreated pertussis or after therapy is not known but exceeds that of positive cultures. Since much of the period during which the organism can be recovered from the nasopharynx falls into the catarrhal phase, when the etiology of the infection is not suspected, there is only a small window of opportunity for culture-proven diagnosis. Cultures from infants and young children are more frequently positive than those from older children and adults; this difference may reflect earlier presentation of the former age group for medical care. Direct fluorescent antibody tests of nasopharyngeal secretions for direct diagnosis may still be available in some laboratories but should not be used because of poor sensitivity and specificity. Pseudo-outbreaks of pertussis have been reported as a result of false-positive PCR results. Greater standardization of PCR methodology can alleviate this problem.

As a result of the difficulties with laboratory diagnosis of pertussis in adolescents, adults, and patients who have been symptomatic for >4 weeks, increasing attention is being given to serologic diagnosis. Enzyme immunoassays detecting IgA and IgG antibodies to pertussis toxin, filamentous hemagglutinin, pertactin, and fimbriae have been developed and assessed for reproducibility. Two- or fourfold increases in antibody titer are suggestive of pertussis, although cross-reactivity of some antigens (such as filamentous hemagglutinin and pertactin) among Bordetella species makes it difficult to depend diagnostically on seroconversion involving a single type of antibody. Late presentation for medical care and prior immunization also complicate serologic diagnosis because the first sample obtained may in fact be a convalescent-phase specimen. Criteria for serologic diagnosis based on comparison of results for a single serum specimen with established population values are gaining acceptance, and serologic measurement of antibody to pertussis toxin is becoming more widely standardized and available for diagnostic purposes, particularly in outbreak settings and for surveillance.

A child presenting with paroxysmal cough, posttussive vomiting, and whoop is likely to have an infection caused by B. pertussis or B. parapertussis; lymphocytosis increases the likelihood of a B. pertussis etiology. Viruses such as respiratory syncytial virus and adenovirus have been isolated from patients with clinical pertussis but probably represent co-infection.

In adolescents and adults, who often do not have paroxysmal cough or whoop, the differential diagnosis of a prolonged coughing illness is more extensive. Pertussis should be suspected when any patient has a cough that does not improve within 14 days, a paroxysmal cough of any duration, a cough followed by vomiting (adolescents and adults), or any respiratory symptoms after contact with a laboratory-­confirmed case of pertussis. Other etiologies to consider include infections caused by Mycoplasma pneumoniae, Chlamydia pneumoniae, adenovirus, influenza virus, and other respiratory viruses. Use of angiotensin-converting enzyme (ACE) inhibitors, reactive airway disease, and gastroesophageal reflux disease are well-described noninfectious causes of prolonged cough in adults.

Chemoprophylaxis

Because the risk of transmission of B. pertussis within households is high, chemoprophylaxis is widely recommended for household contacts of pertussis cases. The effectiveness of chemoprophylaxis, although unproven, is supported by several epidemiologic studies of institutional and community outbreaks. In the only randomized placebo-controlled study, erythromycin estolate (50 mg/kg per day in three divided doses; maximum dose, 1 g/d) was effective in reducing the incidence of bacteriologically confirmed pertussis by 67%; however, there was no decrease in the incidence of clinical disease. Despite these disappointing results, many authorities continue to recommend chemoprophylaxis, particularly in households with members at high risk of severe disease (children <1 year of age, pregnant women). Data are not available on use of the newer macrolides for chemoprophylaxis, but these drugs are commonly used because of their increased tolerability and their effectiveness.

Immunization

(See also Chap. 5) The mainstay of pertussis prevention is active immunization. Pertussis vaccine became widely used in North America after 1940; the reported number of pertussis cases subsequently fell by >90%. Whole-cell pertussis vaccines are prepared through the heating, chemical inactivation, and purification of whole B. pertussis organisms. Despite their efficacy (average estimate, 85%; range for different products, 30–100%), whole-cell pertussis vaccines are associated with adverse events—both common (fever; injection-site pain, erythema, and swelling; irritability) and uncommon (febrile seizures, hypotonic hyporesponsive episodes). Alleged associations of whole-cell pertussis vaccine with encephalopathy, sudden infant death syndrome, and autism, although not substantiated, have spawned an active anti-immunization lobby. The development of acellular pertussis vaccines, which are effective and less reactogenic, has greatly alleviated concerns about the inclusion of pertussis vaccine in the combined infant immunization series.

Although whole-cell vaccines are still used extensively in developing regions of the world, acellular pertussis vaccines are used exclusively for childhood immunization in much of the developed world. In North America, acellular pertussis vaccines for children are given as a three-dose primary series at 2, 4, and 6 months of age, with a reinforcing dose at 15–18 months of age and a booster dose at 4–6 years of age.

Although a wide variety of acellular pertussis vaccines were developed, only a few are still widely marketed; all contain pertussis toxoid and filamentous hemagglutinin. One acellular pertussis vaccine also contains pertactin, and another contains pertactin and two types of fimbriae. In light of phase 3 efficacy studies, most experts have concluded that two-component acellular pertussis vaccines are more effective than monocomponent vaccines and that the addition of pertactin increases efficacy still more. The further addition of fimbriae appears to enhance protective efficacy against milder disease. In two studies, protection conferred by pertussis vaccines correlated best with the production of antibody to pertactin, fimbriae, and pertussis toxin.

Adult formulations of acellular pertussis vaccines have been shown to be safe, immunogenic, and efficacious in clinical trials in adolescents and adults and are now recommended for routine immunization of these groups in several countries, including the United States. In this country, adolescents should receive a dose of the adult-formulation diphtheria–tetanus–acellular pertussis vaccine at the preadolescence physician’s visit, and all unvaccinated adults should receive a single dose of this combined vaccine. In addition, in the United States, pertussis immunization is specifically recommended for health care workers and for women during each pregnancy to increase passive transfer of maternal antibodies to the fetus. Pertussis vaccine coverage among U.S. adolescents was 78.2% in 2011, but coverage among adults is low (2.1% as of 2007). Further improvements in adult vaccine coverage may permit better control of pertussis across the age spectrum, with collateral protection of infants too young to be immunized. However, more effective vaccines with longer-lasting protection will ultimately be needed to control this disease.