A retrospective performance evaluation of an adenovirus real-time PCR assay.

ABSTRACT

Human adenoviruses (AdVs) cause a wide range of diseases. To date, there are at least 60 known human AdV types and, as these exhibit high levels of genetic variation this could impact potentially upon their detection by polymerase chain reaction (PCR)-based technology. Here, the sensitivity of a pan-AdV real-time PCR assay (AdV-PCR) used widely for testing clinical samples was determined retrospectively. An in silico analysis was performed initially using the 370 AdV sequences available on the Genbank database. To investigate for potential false-negative results, two additional AdV-PCR assays were used to re-evaluate 779 respiratory samples submitted for virus testing and 1,012 nasal swab samples collected as part of an ongoing community-based study. The results were then compared to those obtained by AdV-PCR. In silico analysis showed the presence of mismatches in the AdV-PCR primers and probe for most AdV sequences available on Genbank. Notably, 215 of the 370 (58%) sequences had at least three mismatches with the AdV-PCR forward primer. Of the 779 clinical samples, 88 were identified as AdV-positive, of which 84 were positive by the AdV-PCR. The four samples providing false-negative results in the AdV-PCR had high cycle threshold values in the other methods suggesting that sampling at low load, rather than sequence variation, was responsible for the negative results. No false-negative AdV-PCR results were observed for the community-based study samples. Reassuringly, the results show that despite the high level of sequence variation in the AdV-PCR assay oligonucleotide targets, the assay remains suitable for routine detection of human AdV strains.