We designed a multiplex PCR test for the detection and identification of 20 different human respiratory virus types and subtypes, including conventional respiratory viruses, common cold viruses, and newly emerging respiratory viruses. Following nucleic acid extraction, the first step of the assay is a multiplex RT-PCR using 14 consensus primers that amplify consensus regions of the various viral genomes. All PCRs gave specific and robust products, with predominant bands for most viral targets being amplified and only the occasional minor nonspecific band. The second step of the RVP assay involved detection of specific amplicons by using a multiplex TSPE reaction, in which the TSPE primers annealed to specific amplicons and were extended by tag polymerase incorporating biotinylated dCTP. The TSPE reaction used chimeric primers containing a virus-specific sequence and a unique tag sequence that allowed each product to be addressed to a specific microbead that contained a complementary anti-tag oligonucleotide (Fig. ). TSPE products captured on microbeads were detected with a streptavidin-phycoerythrin conjugate, and signals produced for each bead were analyzed by the Luminex 100 flow cell instrument and expressed as MFI units. A total of 21 microbeads were used in the assay, each capturing a specific PCR product (Table ). The use of consensus PCR primers together with type- and subtype-specific TSPE primers allowed the detection of the following viruses: influenza A and B viruses; influenza A virus subtypes H1, H3, and H5, including the H5N1 Asian lineage; parainfluenza virus types 1, 2, 3, and 4; RSV types A and B; MPV; adenovirus; rhinovirus; enterovirus; and CoVs OC43, 229E, NL63, HKU1, and SARS-CoV.
Assignment of virus types and subtypes to individual microbeads for detection by the Luminex x-MAP system
The background signals for all 21 beads were low, generally in the range of 50 to 100 MFI units, whereas positive signals were in the 2,000- to 8,000-MFI-unit range, giving signal-to-cutoff ratios in the range of 100 to 500 for all targets (Fig. ). The low background level on all of the remaining beads (excluding the positive bead) indicates the absence of cross-talk or signal on other beads, indicating the high specificity of hybridization of the Universal Array tag and anti-tag oligonucleotides used for signal sorting. The excellent specificity afforded by the PCR and TSPE primers, together with the excellent signal-to-cutoff ratios, allowed the detection of multiple targets. This detection is shown in Fig. for an influenza A virus, subtype H1N1, that has a strong signal on both the matrix gene bead and the influenza A virus, subtype H1, hemagglutinin gene bead. The high specificity of the assay allowed for the detection of 15 specimens that were positive for two viruses (data not shown).
FIG. 2. RVP test results obtained for 11 NP specimens showing TSPE signals recorded for all microbeads. The 11 positive specimens are indicated across the bottom, with the signals for each of the 21 beads indicated by colored bars. The first sample on the left (more ...)
The analytical sensitivity for each viral target was determined by testing serial dilutions of stock virus seeds (50% tissue culture infectious doses [TCID50]/ml) or in vitro-generated RNA transcripts from plasmids containing cloned amplicons (genome equivalents). For all 20 virus types and subtypes tested, the RVP assay detected between 0.1 and 100 TCID50 of virus. The RVP test had the following analytical sensitivities: 0.1 TCID50 for rhinovirus, enterovirus, CoV 229E, and influenza A virus subtypes H1 and H3; 0.5 TCID50 for influenza B virus, parainfluenza virus type 3, and MPV; 1 TCID50 for RSV type A and parainfluenza virus type 4; 10 TCID50 for parainfluenza virus type 2, RSV type B, and CoVs NL63 and OC43; and 100 TCID50 for adenovirus, parainfluenza virus type 1, and SARS-CoV. The corresponding analytical sensitivities in genome equivalents were 50 to 250 for all virus types/subtypes.
We evaluated the performance of the RVP assay by testing 294 respiratory tract specimens that were submitted to the clinical virology laboratory for routine investigation of respiratory viruses. Aliquots of each specimen were tested by routine DFA plus culture, followed by the RVP test. DFA and culture were performed in the clinical virology laboratory, and the RVP test was performed in the research laboratory by technologists blinded to previous results obtained for the specimens. For the 294 specimens, there were 228 concordant results, including 123 positives by DFA/culture and the RVP test and 105 negatives by both tests (Table ). DFA/culture detected 128 positive specimens, and the RVP test detected 123 of these, for an unadjusted sensitivity of 96.1% for the seven conventional respiratory viruses (influenza A and B viruses, parainfluenza virus types 1 to 3, RSV, and adenovirus) routinely detected in most clinical laboratories. The RVP test detected an additional 61 positive specimens, 14 of which were negative by DFA/culture for the seven viruses tested, and 47 were positive for viruses not tested for by DFA/culture. These 61 additional positive specimens included 2 for influenza A virus, 1 for parainfluenza virus type 1, 2 for parainfluenza virus type 2, 1 for parainfluenza virus type 4, 2 for RSV, 8 for MPV, 39 for rhinovirus/enterovirus, 6 for OC43 CoV, 2 for NL63 CoV, 1 for HKU1 CoV, and 3 specimens that were positive for two viruses, including 1 specimen that was positive for MPV and rhinovirus/enterovirus and 2 specimens that were positive for OC43 and rhinovirus/enterovirus. All of the 66 specimens that gave discordant results, including the 5 DFA/culture-positive specimens that were negative by the RVP test and the 61 specimens that were positive by the RVP test and negative or positive for viruses not tested for by DFA/culture, were tested by a second PCR that targeted a different area of the viral genome. Table shows the results for the 5 specimens that had given DFA/culture-positive, RVP test-negative discordant results and for the 14 specimens that had given DFA/culture-negative, RVP test-positive discordant results. Three of the 5 DFA/culture-positive, RVP test-negative specimens (numbers 167, 191, and 187) were confirmed to be positive by PCR, indicating two false positives (numbers 286 and 62) by DFA/culture. All of the additional 61 RVP test-positive specimens were confirmed as true positives by the second PCR. If a true positive is defined as being positive by two or more tests (DFA, culture, the RVP test, and/or confirmatory PCR), then there were 183 positives and 111 negatives. To determine how the RVP test performed compared to DFA and culture, we eliminated the 47 specimens that were positive for a virus not tested for by DFA and culture (i.e., parainfluenza 4; MPV; CoVs OC43, 229E, NL63, and HKU1; and rhinovirus/enterovirus) and used the remaining 247 specimens for analysis. Among these 247 specimens, there were 137 positives and 110 negatives. The sensitivity and specificity of DFA/culture were 91.9% (126/137) and 98.2% (108/110), respectively. The RVP test had a sensitivity of 97.8% (134/137) and a specificity of 96.4% (107/110). If, however, all confirmed respiratory viruses detected by the RVP test are included in the analysis, then the RVP assay detected 180 out of 183 positive specimens and had an overall sensitivity of 98.4%, whereas DFA/culture detected only 126 out of 183 specimens and had a sensitivity of 68.8%.
Distribution of DFA/culture results and RVP test results for 294 NP specimens
PCR results for 5 DFA/culture-positive, RVP test-negative and 14 DFA/culture-negative, RVP test-positive specimens
Of particular interest was the finding that 15 out of 294 (5.2%) specimens were positive for two viruses in this group of specimens. The dual infections included the following combinations: one type 1 parainfluenza virus plus one rhinovirus/enterovirus, one type 2 parainfluenza virus plus one rhinovirus/enterovirus, two type 3 parainfluenza viruses plus one rhinovirus/enterovirus, three RSVs plus one rhinovirus/enterovirus, one adenovirus plus one rhinovirus/enterovirus, one MPV plus one OC43 CoV, three MPVs plus one rhinovirus/enterovirus, two OC43 CoVs plus one rhinovirus/enterovirus, and one adenovirus plus one KHU1 CoV. No specimen was positive for three respiratory viruses. Testing additional specimens from a separate study has confirmed a dual positivity rate for NP specimens collected from our combined pediatric and adult population of 5% to 8% using the RVP assay during the 2005-to-2006 season.