Laboratory diagnosis of Bordetella infection : DNA hybridization and PCR validated against serology and culture in a phase 3 vaccine efficacy trial

Abstract

The present study has focused on developing and validating assays alternative to culture for detection of Bordetella pertussis. The sensitivity of culture is comparatively low when compared to serology, varying from 30% to 80%. An attempt to increase this sensitivity was made by development of assays detecting the DNA of the bacteria. A DNA hybridization assay using a DNA probe specific for B.pertussis involved preculture of nasopharyngeal aspirates applied on membranes, which were placed on agar plates. Bacteria were then detected by DNA hybridization of the membranes.

Analysis of nasopharyngeal aspirates from 179 patients showed that both sensitivity and specificity of DNA hybridization were comparable to culture for diagnosis of B.pertussis, and that the method resulted in a saving of 3-5 days. To improve the sensitivity, a PCR based assay, amplifying the pertussis toxin promoter region, was developed. The PCR generated one band for B.pertussis, B.parapertussis and B.hronchiseptica. Identification of the three species was performed in a second step using restriction enzyme cleavage of the amplicons. An evaluation in 166 aspirates indicated an increased sensitivity for pertussis case finding with retained specificity. Further improvement of the sensitivity was attempted with a nested PCR using the same inner primers as in the already evaluated method. The PCR was validated against serology andcultured by analysis of 2421 nasopharyngeal aspirates from patients participating in a phase 3 pertussis vaccine efficacy trial. The diagnostic sensitivity for detection of B.pertussis was 90.2%, and for B.parapertussis 74.0% using culture as reference. PCR also identified an additional 34% positive samples. Relating these cases to serological, epidemiological, and clinical data indicated a PCR specificity approaching 100%. As the PCR was referred to highly efficient culture procedures within the vaccine trial, including bedside culture, enrichment procedures, and minimized transport time, the increased sensitivity achieved by PCR may be seen as a minimum of what can be expected under routine conditions. The time required for diagnosis by PCR was lowered from 2-5 days shorter than that required for culture. For validation of the PCR method it was of importance that also the reference methods were optimized or validated. Therefore sampling by collection of nasopharyngeal aspirates and swabs was compared, indicating that aspiration gave a culture sensitivity higher than or comparable to that which was found when swabs were used. The aspiration technique was preferred by the majority of parents and nurses.

The aspiration technique was chosen for the pertussis vaccine efficacy trial 1992-95. Immunomagnetic capture using paramagnetic beads coated with polyclonal antibodies directed towards the surface antigens of B.pertussis and B.parapertussis was found to be an efficient method for enrichment, easier to perform than the standard method, and lowering the risk for false positivity by contamination of the samples. This application was used in a nested PCR protocol, using approximately the same target region. Analysis of 55 nasopharyngeal aspirates indicated that the method had the potential to constitute a simpler, safer, and less time consuming sample treatment. The ELISA format used to estimate the degree of amplification considerably simplified the handling of results, and increased the objectivity of interpretation. Serology data are of utmost importance for evaluation of culture negative and PCR positive results. Within the vaccine trial, efforts have been put into standardizing and validating the IgG and IgA anti-FHA and -PT ELlSAs used for diagnostic pertussis serology. It was of importance to establish a method transferring ELISA absorbance values to units with respect to sensitivity and repeatability. A study comparing various such methods argued for the use of calibration using reference sera. The results indicated that the highest repeatability, as well as the highest diagnostic sensitivity, was obtained using the reference line unit calculation mode. This mode was chosen for the pertussis vaccine efficacy trial, and it is recommended as a reference method both in routine diagnostics and in vaccination trials.