Comparing the Yield of Nasopharyngeal Swabs, Nasal Aspirates, and Induced Sputum for Detection of Bordetella pertussis in Hospitalized Infants.

Background. Advances in molecular laboratory techniques are changing the landscape of Bordetella pertussis illness diagnosis. Polymerase chain reaction (PCR) assays have greatly improved the sensitivity detection and the turnaround time to diagnosis compared to culture. Moreover, different respiratory specimens, such as flocked nasopharyngeal swabs (NPSs), nasopharyngeal aspirates (NPAs), and induced sputum, have been used for B. pertussis detection, although there is limited head-to-head comparison to evaluating the PCR yield from the 3 sampling methods. Methods. Hospitalized infants <6 months of age who fulfilled a broad syndromic criteria of respiratory illness were tested for B. pertussis infection by PCR on paired NPSs and NPAs; or paired NPSs and induced sputum. An exploratory analysis of B. pertussis culture was performed on induced sputum specimens and in a subset of NPSs. Results. From November 2014 to May 2015, 484 infants with paired NPSs and NPAs were tested; 15 (3.1%) PCR-confirmed pertussis cases were identified, 13 of which were PCR positive on both samples, while 1 each were positive only on NPS or NPA. From March to October 2015, 320 infants had NPSs and induced sputum collected, and 11 (3.4%) pertussis cases were identified by PCR, including 8 (72.7%) positive on both samples, 1 (9.1%) only positive on NPS, and 2 (18.2%) only positive on induced sputum. The 3 types of specimens had similar negative predictive value >99% and sensitivity >83%. Compared to PCR, culture sensitivity was 60% in induced sputum and 40% in NPSs. Conclusions. Flocked nasopharyngeal swabs, nasopharyngeal aspirates, and induced sputum performed similarly for the detection of B. pertussis infection in young infants by PCR.