Pertussis, an endemic and epidemic disease with a characteristic paroxysmal cough, is most often caused by Bordetella pertussis [2] but is sometimes caused by B. parapertussis. Historical dogma and epidemic patterns of recognized pertussis had, until recently, led to the beliefs that pertussis is an epidemic disease of children; adult pertussis is rare; and vaccine-induced immunity is relatively short-lived, whereas immunity following infection is life-long. Studies done during the past decade indicate that B. pertussis infections and illnesses are common and endemic in adults and that these infections are the reservoir for pertussis in susceptible children. In addition, it is clear that immunity after infection is not lifelong, and it seems that this type of immunity is inferior to that induced by immunization.

The importance of adult pertussis was recognized in the prevaccine era [3, 4]. In 1916, Luttinger [3] told of "Pertussis Pete," an adult who spread pertussis among three families; in 1925, Madsen [4] referred to "grandmothers' whooping cough." In 1978, Nelson [5] noted that 15 of 22 infants (68%) younger than 12 weeks of age were infected by adult exposure to pertussis, and subsequent studies of severe and fatal pertussis in very young infants have indicated an adult as the source case in most instances [6-8].

In the prevaccine era, reported pertussis was a cyclic disease, with epidemic peaks every 2 to 5 years [9]. When pertussis was brought under control by vaccination of children in the 1960s in the United States and England and Wales, it was noted that the 2- to 5-year cyclic pattern continued. This finding was surprising because when a disease (such as measles) is controlled by immunization, the circulation of the causative organism decreases and the epidemic cycle lengthens [10]. Failure of the pertussis cycle to lengthen led to Fine and Clarkson's suggestion [11] that immunization controlled the disease in children but did not disrupt circulation of the organism. This observation and the knowledge that adults were a common source of disease in infants led our group and others to study the epidemiology of adult pertussis.

The breakthrough that allowed the success of recent studies was the advancement in serologic techniques for the diagnosis of B. pertussis infection. Changes in IgG and IgA antibodies to multiple B. pertussis antigens could be determined with high precision, and cases could be determined by the demonstration of high titers in single serum samples compared with those of population controls [10].

From September 1986 through February 1989, we studied UCLA students with cough that lasted 6 days or more [10]. During this 2.5-year period, we found that 26% of the evaluated students had pertussis and that illness was endemic throughout the study period. Similar studies [12-17] done in adults in the United States, Australia, and Germany have had generally similar findings. Twelve percent to 32% of persons with prolonged cough have been found to have pertussis. In our study, important clinical findings in persons with pertussis were that the median duration of cough illness before seeking care was 21 days, productive cough was rare, the most common clinical diagnosis was bronchitis, and in no case was the diagnosis of pertussis entertained [10]. In children with primary B. pertussis infections, leukocytosis caused by lymphocytosis is common; however, none of the infected students had absolute lymphocytosis [10].

In contrast with our findings in the UCLA students (94% of whom had been vaccinated in childhood), we found that German adults (most of whom had not been vaccinated during childhood) were more likely to have typical pertussis with whooping and post-tussive vomiting [12]. Twenty-six percent of these adults had had pertussis during childhood.

Immunoglobulin A antibodies to B. pertussis antigens usually result from infection, not immunization. With this fact in mind, we examined the prevalence and degree of IgA antibody to four B. pertussis antigens in young adults in the United States and Germany [18]. We found that the mean titers and the prevalence of antibody to the four antigens were similar, suggesting that the circulation of B. pertussis in adults in the two countries was similar even though pertussis was epidemic in Germany and rare in the United States. In another study [19], in which we obtained serum samples yearly for 5 years from 51 persons, 90% of these persons had serologic evidence of at least one case of pertussis [19].

A summary of what I have reviewed and speculation about the future are given in the Table 1 Many acellular pertussis vaccines have been shown to be safe and effective in infants and children. Immunogenicity and reactogenicity studies with acellular vaccines have been evaluated in adults, and an efficacy trial for an adolescent and adult vaccine is presently under way in the United States. I believe that the circulation of B. pertussis can be curbed if we extend our immunization program to include adolescents and adults.

Until pertussis in adults is universally prevented through immunization, clinicians must be on the alert for the clinical disease. It should be suspected when an illness thought to be a cold is associated with the development of a paroxysmal cough that worsens at night and is nonproductive. Between coughing attacks, the patient with pertussis has no symptoms; this fact is useful in differentiating pertussis from cough illnesses caused by respiratory viruses or allergic conditions. The laboratory diagnosis of pertussis in the adult is difficult because of the usual delay in suspicion of the disease. The diagnosis can be made by culture if the specimen (nasopharyngeal swab or aspirate) is obtained within 2 weeks of the onset of cough or serologically by use of enzyme-linked immunosorbent assay to show increases in IgA or IgG antibody titer or high single titers to pertussis toxin. Adults suspected of having pertussis should be treated with erythromycin for 14 days. Early treatment will shorten the course of the illness and eliminate the risk for transmission.

Smallpox and diphtheria were both initially brought under control in the United States by the immunization of adults as well as children. However, both of these diseases were a major cause of illness and death in adults, whereas adult illness from pertussis goes unrecognized. The morbidity rate of adult pertussis has not been adequately studied; however, my experience and that of other colleagues indicates that substantial, prolonged adult illnesses are not uncommon [1, 20].

Pediatricians and others who care for children in the United States have accepted the public health responsibilities of universal immunization. As a result, adult rubella (the cause of congenital rubella) has almost been eliminated, and immunization of infants and children for hepatitis B is expected to eliminate this predominately adult disease. If acellular pertussis vaccines prove to be effective in adults, I hope that those who give primary care to adults will accept the public health challenge and effectively carry out universal immunization for this burdensome disease.

James D. Cherry, MD, MSc

University of California, Los Angeles, School of Medicine

Los Angeles, CA 90095