Diagnosing pertussis: the role of polymerase chain reaction

Whereas the diagnosis of classical pertussis has traditionally been based on clinical criteria, increasing numbers of atypical presentations suggest the need for an extensive laboratory-based approach.

To assess the relative efficacy of clinical and laboratory methods in the diagnosis of Bordetella pertussis by patient age and immunization status.

We compared the clinical and laboratory diagnosis of B. pertussis in 87 pre-vaccinated, 78 recently vaccinated, and 75 post-vaccinated children with suspected pertussis. Serum and nasopharyngeal swabs were obtained for serology, culture and polymerase chain reaction.

PCR and culture identified 41% and 7% of patients with B. pertussis, respectively (P < 0.001). All positive cultures were PCR-positive. Positive PCR was less common among those recently vaccinated than among those in the pre- (P < 0.001) and post-vaccinated groups (P < 0.05). Positive culture was more common among those pre-vaccinated than among those recently vaccinated (P < 0.01). Positive tests for immunoglobulin M and A were more common among the post-vaccinated than the pre- and recently vaccinated (P < 0.001), respectively. Logistic regression analyses revealed that clinical criteria have no significant association with infection in recently and post-vaccinated children. Among the pre-vaccinated children, whoop and cough duration were associated with a positive PCR (odds ratio 7.66 and 0.5, P < 0.001). Seventy-six percent of pre-vaccinated, 39% of recently vaccinated and 40% of post-vaccinated children with positive PCR did not meet the U.S. Centers for Disease Control diagnostic criteria for B. pertussis.

PCR is a useful tool for pertussis diagnosis, particularly in pre-vaccinated infants. The yield of culture and serology is limited, especially among pre- and recently vaccinated children. In pre-vaccinated infants with whoop and less than 2 weeks of cough, PCR testing should be implemented promptly.