Comparison of Bordetella species identification among differing rt-PCR assays in the United States.

In the United States, the general laboratory method for diagnosing pertussis, caused by Bordetella pertussis, is real-time PCR (rt-PCR) targeting insertion sequence 481 (IS481). Other Bordetella species (parapertussis, holmesii, and bronchiseptica) can also cause a pertussis-like syndrome, and some commercial laboratory assays include the insertion sequence 1001 (pIS1001) that can detect B. parapertussis/B. bronchiseptica (BppBb). Because IS481 exists in B. pertussis and B. holmesii, current commercial assays cannot differentiate these two species. We used a multiplex rt-PCR assay containing species-specific targets to Bordetella to evaluate clinical specimens detected as B. pertussis/B. holmesii (BpBh) or BppBb by commercial laboratories. A sample of 3,984 clinical specimens positive for IS481 or pIS1001 from two commercial laboratories during 2012-2019 were re-tested at CDC. Agreement of Bordetella species between the CDC and commercial laboratory assays, and the proportion of commercial laboratory specimens that were non-B. pertussis by CDC's assay was assessed. Overall agreement in Bordetella species detection and identification between the CDC and commercial lab assays was 85%. Agreement for identifying B. pertussis was 87% for 3,663 BpBh specimens and 98% for identifying B. parapertussis in 310 BppBb specimens. CDC's assay detected B. holmesii in 55/3,984 (1.4%) specimens. Most discrepant results (410/490, 82%) were BpBh specimens interpreted as indeterminate B. pertussis at CDC. We found a small portion of B. holmesii in a sample of IS481-positive clinical specimens originally identified by commercial laboratory rt-PCR assays, suggesting that commercial PCR assays are a reliable diagnostic tool for correctly identifying Bordetella species in most patients with suspected pertussis. IMPORTANCE:When testing specimens collected from patients with suspected pertussis, large-scale commercial laboratories in the United States employ an IS481-based assay that cannot differentiate between Bordetella pertussis and Bordetella holmseii. The level of B. holmesii causing pertussis-like illness in the United States is not well-understood given that only B. pertussis is nationally notifiable. After re-testing with a multiplex, species-specific rt-PCR assay, our data show low levels of B. holmesii identified in a sample of IS481-positive clinical specimens originally identified by commercial laboratory rt-PCR assays. These results reinforce the validity of large-scale commercial rt-PCR testing as a reliable diagnostic tool for pertussis in the United States.