Analytical Sensitivity of Seven Point-of-Care Influenza Virus Detection Tests and Two Molecular Tests for Detection of Avian Origin H7N9 and Swine Origin H3N2 Variant Influenza A Viruses

LETTER

Rapid virological diagnosis is important for early case identification, initiation of specific antiviral therapy, and implementation of infection control measures for patients suffering from influenza. Most point-of-care (POC) test kits commonly used at the outpatient setting which provide results within 15 to 30 min are based on antibodies that specifically bind to the influenza virus nucleoprotein (1), while rapid molecular tests based on PCR targeting the relatively conserved M gene are frequently used for hospitalized patients. With the emergence of human infections due to the swine origin influenza A H3N2 variant (H3N2v) virus in North America in 2009 and more recently the avian origin influenza A H7N9 (H7N9) virus in China, it would be important to ascertain the analytical sensitivities of these rapid tests for the two novel influenza viruses. Therefore, seven commercially available POC antigen tests and two multiplex molecular tests for swine origin H3N2v and avian origin H7N9 viruses were evaluated.

Two swine origin H3N2v isolates (A/Indiana/08/2011 and A/Wisconsin/12/2010), two avian origin H7N9 isolates (A/Zhejiang/DTID-ZJU01/2013 [2] and A/Anhui/1/2013 [3]), and one H1N1pdm09 isolate (A/HK/415742/09/H1N1) were used for this evaluation. The viruses were cultured for 2 days in Madin-Darby canine kidney (MDCK) cells with tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin. Aliquots of culture supernatant were frozen at −80°C until use. An aliquot of each virus was thawed, and serial 10-fold dilutions of the virus stock with serum-free minimal essential medium (MEM) were tested. Further 2-fold dilutions were done in duplicate from this endpoint to determine the limit of detection of each rapid antigen test more accurately. Similar dilutions were concurrently assayed in two multiplex reverse transcription-PCRs (RT-PCRs) for the detection of respiratory viruses, including Resplex II plus v2.0 (Qiagen GmbH, Germany) and xTAG RVP FAST (Luminex Molecular Diagnostic, Toronto, Canada). The seven commercially available POC rapid influenza antigen detection kits included the Espline influenza A&B (Fujirebio Inc., Tokyo), BinaxNow influenza A+ B (Emergo Europe, The Netherlands), Wondfo (Wondfo Biotech Co., Ltd., Guangzhou, People's Republic of China), QuickVue influenza A+B and Sofia Influenza A+B (Quidel Corporation, CA), Directigen EZ Flu A+B, and BD Veritor system (Becton, Dickinson, Sparks, MD). All POC antigen and multiplex molecular tests were performed according to the manufacturers' instructions (4, 5). The viral M gene genome copy number of each virus dilution was determined using real-time quantitative RT-PCR as described previously with modification (6, 7). The limit of detection of the M gene RT-PCR is 10 copies per reaction. The same virus dilutions were inoculated onto MDCK cells to determine the 50% tissue culture infective dose (TCID₅₀).

Our results showed that the limit of detection of these seven POC antigen tests was between 5.4 to 8.9 log copies/ml (1.0 to 4.3 log TCID₅₀/ml) and 4.8 to 7.3 log copies/ml (1.0 to 4.3 log TCID₅₀/ml) for the H3N2v and H7N9 viruses, respectively (Table 1). Unexpectedly, the limits of detection of the two multiplex RT-PCRs were 4.9 to 7.9 log copies/ml (1.0 to 3.3 log TCID₅₀/ml) and 5.8 to 7.2 log copies/ml (2.0 to 4.3 log TCID₅₀/ml) for the H3N2v and H7N9 viruses, respectively, which did not appear to be better than those of the antigen tests. As for the real-time quantitative RT-PCR targeting the M gene, which was used as the “gold standard,” the limits of detection were 2.9 log copies/ml (0.01 to 0.02 log TCID₅₀/ml) and 2.2 to 3.3 log copies/ml (0.2 to 0.3 log TCID₅₀/ml) for the H3N2v and H7N9 viruses, respectively.

The analytic sensitivity results of our evaluation of the rapid antigen tests for the swine origin H3N2v and avian origin H7N9 influenza viruses are comparable with those for the 2009 pandemic influenza A H1N1 virus (Table 1) (1, 4). We therefore expect similar in-use clinical sensitivities of 40 to 60% for H3N2v and H7N9 (5). Surprisingly, the multiplex RT-PCR test kits had a poor analytical sensitivity similar to that of the rapid antigen tests. Such low in-use sensitivity was previously reported with the 2009 pandemic influenza A H1N1 virus (8). Moreover, in the case of avian origin H7N9 and H5N1 viruses, which may have a predilection for the lower respiratory tract mucosal cells with more 2,3-linked sialic acid receptors (9), the upper respiratory tract specimens may not contain sufficient viral load to provide a positive result even when the viral load peaks with rapid clinical deterioration (2). Clinicians using these rapid POC antigen and multiplex RT-PCR tests must be aware that neither of these tests can confidently rule out H3N2v or H7N9 infection. Moreover, even if positive, none of these tests can differentiate seasonal flu from H3N2v and H7N9. A monoplex influenza virus RT-PCR would be preferred in patients with high clinical suspicion of these infections.