Validation of nested Bordetella PCR in pertussis vaccine trial.

A nested PCR, using a 239-bp sequence in the pertussis toxin promoter region, was developed and evaluated. The assay differentiates Bordetella pertussis, Bordetella parapertussis, and Bordetella bronchiseptica by restriction enzyme analysis of the amplified fragments. The diagnostic performance of the PCR was evaluated in a Swedish pertussis vaccine efficacy trial which took place from 1992 to 1995, including study children and household members and using culture and serology for laboratory confirmation of suspected cases. In total 2,421 nasopharyngeal aspirates were analyzed. The total diagnostic sensitivity for B. pertussis was 90.2% (194 of 215). During the study period samples were processed with and without the cation-exchange resin Chelex. The PCR diagnostic sensitivity for B. pertussis among the Chelex-treated aspirates was 94.9% (75 of 79), and that for B. pertussis among 124 aspirates in a consecutive non-Chelex-treated material was 89.5% (111 of 124). After Chelex treatment of the 13 PCR-negative samples, an additional six became PCR positive, giving a final sensitivity of 94.3%. In addition, PCR was positive for B. pertussis with 57 of 1,744 samples negative by culture but with available serological data. The specificity of PCR with these samples was supported by a significant increase in antibody levels between acute and convalescent sera in 45 cases and by epidemiological or clinical data in all but two of the remaining cases. PCR was also positive for B. pertussis with 26 of 415 aspirates from episodes lacking serology. The diagnostic sensitivity of PCR for B. parapertussis was 74.0% (37 of 50). There were an additional seven culture-negative B. parapertussis PCR findings, six from cases with significant antibody increases against filamentous hemagglutinin only and one from a case lacking serology. There were no samples positive for B. bronchiseptica. In conclusion, PCR detection of B. pertussis and/or B. parapertussis enabled us to identify 90 positive nasopharyngeal aspirates, in addition to the 262 culture-positive samples (an increase of 34%). Relating these cases to serology and clinical data indicated a PCR specificity approaching 100%.