Comparison of multiplex polymerase chain reaction, culture, and serology for the diagnosis of Bordetella pertussis infection

Background: Accurate diagnosis of Bordetella pertussis infection is difficult. Polymerase chain reaction (PCR) tests are more sensitive than culture, but the reported sensitivity is variable. We prospectively compared the performance of culture, serology, and a multiplex PCR for the detection of B. pertussis. Methods: A total of 193 paired nasopharyngeal (NP) swab specimens were examined by both culture and a multiplex PCR. Serology results were available in 103 patients. Medical charts of the patients with discrepant laboratory findings were reviewed and compared with the United States Centers for Disease Control and Prevention (CDC) clinical case definition. Results: Of the 193 specimens, 11 were positive on both culture and PCR, and 14 were positive on PCR only. Of the 103 specimens with serology results, 3 were positive with all three methods, and 69 were negative with all methods. Eleven of the 14 PCR-positive only cases and 6 of the 19 serology-positive only cases were defined as true pertussis cases according to an expanded standard which includes either (1) culture positive or (2) PCR or serology positive with clinical features fulfilling the CDC clinical case definition and the patients having received macrolides treatment for more than 10 days. The sensitivity and specificity of the multiplex PCR were 79% and 98%, respectively, while those for serology were 47% and 85%, and for culture 39% and 100%. Conclusions: Our data confirm the superior sensitivity of the multiplex PCR in detection of B. pertussis, compared with conventional culture and serology. Clinical validation indicates that the multiplex PCR offers specific detection of B. pertussis from NP specimens.