- Top of page
- Materials and Methods
- Transparency Declaration
In a head to head comparison of four multiplex PCR assays with DFA and culture in children, multiplex PCR offered significantly improved sensitivity in the detection of the traditionally diagnosed respiratory viral agents (INFA, INFB, PIV (1–3), RSVA, RSVB, ADV and hMV), in addition to detecting coronaviruses, BoV, enteroviruses and rhinoviruses, which increased the overall positivity rate from 38.4% to 66.9%.
Among all the multiplex assays tested, Seeplex RV15 was the most sensitive for detecting all targets except for enteroviruses and rhinoviruses. All multiplex assays had good sensitivity for the detection of influenza A (93.7–98.4%). Influenza B sensitivity was good in all multiplex assays (100%) except for RVP Fast (64.9%). The superior performance of Seeplex RV15 for RSV (100% sensitivity) reflected strong performance for both RSVA and RSVB, whereas the decreased sensitivity of other assays reflected a weaker performance for either target (i.e. Resplex II v2.0: RSVA 90.4%, RSVB 79.3%, and RVP: RSVA 85.5%, RSVB 98.3%). Similarly, variability in the sensitivity of individual targets of the four parainfluenza viruses resulted in variation in the overall sensitivity. Again, Seeplex RV15 showed good sensitivity for all four types (85.7–100%), while Resplex II v2.0 had reduced sensitivity for PIV4 (60%), RVP had reduced sensitivity for PIV1 (71.4%) and 3 (71.4%), and RVP Fast had reduced sensitivity for PIV1 (46.7%), 2 (77.8%) and 3 (42.8%). Sensitivity for detecting hMPV was good for Seeplex RV15, RVP and RVP Fast (92.3–97.4%), and acceptable for Resplex II v2.0 (82%). However, performance for adenovirus, an important respiratory pathogen, was very variable, ranging from 52.4% (RVP Fast) to 100% (Seeplex RV15), probably reflecting the variation in serotype coverage among the assays.
Of the additional viral agents tested in the multiplex assays, the coronaviruses were consistently detected across all assays except for CoV OC43 by RVP (53.8%) and CoV HKU1 by RVP Fast (16.7%). Seeplex RV15 and RVP Fast detected 100% of bocavirus infections, while the sensitivity of Resplex II v2.0 was only 75%. Detection of enterovirus and rhinovirus was the most inconsistent. Although the specific targets for each multiplex assay are proprietary, it is known that the highly conserved regions of the 5′NTR region of either rhinoviruses or enteroviruses, will also amplify members of the other genus. Thus, some assays, such as the RVP and RVP Fast assays, have combined the enterovirus and rhinovirus targets, because developing specific targets for each genus outside of the 5′NTR region may compromise sensitivity of detection, especially of the rhinoviruses. This is possibly the case with the Seeplex RV15 assay, which separates enteroviruses and rhinoviruses, but has a lower sensitivity than the other assays. Though the Resplex II v2.0 assay differentiates between enteroviruses and rhinoviruses, the occurrence of 38.4% of positive specimens testing positive for both targets, suggests that there may be cross-reactivity between them.
Specificity was excellent for all assays, using our composite reference standard. Without using individual single-plex assays to adjudicate the single test positives, we cannot determine whether the slightly lower specificity observed for a few targets in several assays was due to higher sensitivity of detection or detection of false positives.
Multiplexed respiratory panels provide clinicians with more diagnostic and treatment information for managing patients. In the case of influenza A, knowledge of the subtype is important with respect to predicting the activity of antiviral agents such as the adamantanes and neuraminidase inhibitors. In addition to increased sensitivity and number of viruses detected, multiplex assays permit the improved identification of cases of infection with multiple agents, which may be clinically significant, especially in immune compromised individuals. In our study, we found that two or more viruses were present in 10.9% of specimens (16.3% of positive specimens). Bocavirus and coronaviruses were the viruses most commonly associated with multiple agent infection, followed by human metapneumovirus, the parainfluenza viruses, adenovirus and the entero/rhinoviruses. Influenza A/B and RSV were the least likely to be detected in the presence of another virus. The role of multiple viral agents in affecting the clinical course of disease is at present unknown and worthy of further study.
With respect to the technical performance of the different multiplex assays, the following issues were identified: Seeplex RV15 was designed as a two-step RT-PCR format necessitating a separate RT (cDNA) assay, though a new one-step procedure has been developed. It was the only assay that required three PCR master mixes with five targets in each one plus the internal control. It was also the only assay that incorporated positive controls for all 15 viral targets, which is considered an additional quality control feature of the assay. Seeplex RV15 was the assay with the shortest post-PCR step, especially for a small number of specimens when using the Lab 901 Screen Tape® system (maximum five specimens per run). In contrast, the Resplex II v2.0, RVP and RVP Fast use a 96-well microtitre plate format on the Luminex platform, permitting high throughput analysis.
Practical considerations in most laboratories regarding the feasibility and the direct and indirect costs of introducing multiplex molecular testing for respiratory viruses have led to a relatively slow routine implementation of this methodology. Mahony et al.  have shown that RVP employed as the first-line diagnostic tool in children was the least costly strategy, compared with DFA and culture, DFA alone or DFA plus RVP, when the prevalence of infection was ≥11%. The cost of molecular testing is offset by its more efficient use of labour than conventional DFA and culture and by savings to the healthcare system when additional testing and hospitalization can be avoided by knowledge of a test result with high sensitivity and specificity. Operationally, molecular methods also allow virology laboratories to continue to function, even in the event that viral culture cannot be carried out due to biosafety issues.
In our study, newer assays or versions of the multiplex assays (Resplex II Plus Panel PRE (21 targets) and Seeplex Influenza A/B Subtyping (six targets)) showed good sensitivity and specificity relative to a tri-plex influenza real-time RT-PCR (Astra influenza Screen & Type) for pandemic H1N1 2009 INFA virus detection. This is important, as traditional seasonal H1 subtyping molecular assays will not react with the pandemic strain and it is expected that the 2009 pandemic strain may become the predominant circulating seasonal H1 strain in the immediate post-pandemic period.
We have shown that multiplex PCR increases the sensitivity of detection of respiratory viruses in children by 74.3% over DFA and viral isolation, while maintaining excellent specificity. However, it will be important to develop more effective clinical and laboratory algorithms for their timely and optimal use and to study their impact on patient care in different populations in different clinical settings. Influenza, RSV, parainfluenza virus, adenovirus and hMPV have been well established as leading causes of respiratory infection among infants and children [16–18]. However, the role of rhinoviruses, enteroviruses, bocavirus and coronaviruses as co-pathogens in upper respiratory tract infection or as agents of lower respiratory tract infection, has been less well investigated, and will be aided by studies using this technology .