Most studies that support the statement "Immunity after natural infection with pertussis appears to be lifelong" were performed before introduction of the pertussis vaccine, that is, when diagnosis of pertussis was primarily based on clinical criteria. Thus, repeated pertussis infection or exposure in adults with previous incident pertussis infections would not be expected to result in a clinical syndrome resembling whooping cough; this leads to the impression that "immunity after pertussis appears to be lifelong." Although pertussis infection in adults previously infected with pertussis is possible, estimating the magnitude and clinical implications of this phenomenon is difficult.

Studies from the past decade suggest that pertussis infection (based on serologic evidence) or identification (based on culture or polymerase chain reaction [PCR]) occurs in 10% to 25% of adults with prolonged cough illness (1). However, studies also suggest that most adult pertussis infections based on serologic evidence are asymptomatic (2). Should serologic changes or pertussis recovery in asymptomatic hosts be considered "infection" or "carriage"? How does one distinguish infection from carriage when pertussis is identified in adults with non–whooping cough syndromes? For example, viral upper respiratory tract infections may make carriage of pertussis more likely, similar to what has been reported for Streptococcus pneumoniae carriage (3). Interpretation of these pertussis studies could have been aided by testing for other respiratory pathogens.

Another limitation of the pertussis literature is lack of a uniform gold standard for pertussis infection. The specificity of serologic criteria can vary because of use of single compared with paired serum specimens, as well as potential cross-reactivity of immunoglobulins with B. parapertussis and Haemophilus influenzae. The sensitivity of serologic criteria is also unclear, since up to 10% of patients with pertussis identified on PCR do not meet any serologic criteria (4). Conversely, PCR testing can also yield false-positive results (based on negative controls) and false-negative results (when compared with serologic evidence).

If previous natural pertussis infection does not confer lifelong immunity, what are the resulting clinical implications? There are few for physicians practicing in populations where pertussis vaccination is common, since the literature (at least in the United States) has regularly reported the lack of lifelong immunity conferred by vaccination and the increased prevalence of pertussis carriage or infection among adults with prolonged cough illness. However, in populations without widespread pertussis vaccination in which pertussis is endemic, physicians should be alert to the possibility that adults remainan important reservoir for pertussis (despite no or atypical clinical manifestations of disease).