Comparison of three non-nested RT-PCR for the detection of influenza A viruses

Background: The viral isolation technique (VIT) is largely used as a gold s tandard for the detection of influenza A and B viruses in respiratory sampl es. Some recent studies have pointed out that the polymerase chain reaction (PCR) assays allow sensitive and rapid detection of influenza viruses, als o providing excellent correlation with traditional methods. Objectives and design study: The aim of this study was to evaluate the efficiency of three non-nested PCR, two PCR-hybridization assays using primers defined in M an d NS genes, and one PCR which uses primers defined in NP, NS and HA genes a nd combines the detection of H3N2 and H1N1 hemagglutinin genes using define d primers in NP, NS and HA genes (PCR3), in comparison with an IF assay (IF A) and viral isolation technique (VIT). The study was carried out on 244 na sal samples collected mainly by practitioners of the GROG surveillance netw ork during winter 1998-1999 for the detection of influenza A virus. Results : Overall influenza viruses were detected more frequently by PCR techniques in 157 (64.3%), 147 (60.2%), 110 (45%) cases for PCR1, PCR2,;PCR3, respect ively, than by VIT or IFA, in 100 (40.9%) and 74 (30.3%) cases, respectivel y. Taking the positive culture samples as a reference, 100 (41.8%) samples were found to be positive for influenza A, and the sensitivity of IFA, PCR 1, PCR 2 and PCR3 techniques were 70, 100, 99, and 90%, respectively as com pared with viral isolation cultures. On the other hand, as 86.5% of positiv e samples were positive with at least two different techniques, the sensiti vity, specificity, VPP and VPN of each technique were recalculated taking i nto account a further criterion defining a positive sample: positivity with two techniques. We observe that techniques PCR 2 and particularly PCR 1 ha ve very good sensitivity, respectively 98.6 and 100%, far better than the t raditional techniques, IFA and culture, whilst maintaining acceptable speci ficity: 94.1 and 86.1%, respectively. In both cases they enable 141 (57.7%) A-positive influenza samples to be detected instead of the 100 (40.9%) obt ained when culture is the reference test. IFA, culture and PCR 3 are highly specific (VPP = 100%), but in comparison with PCR 1 and 2 their sensitivit y, respectively 51.7, 69.9, 77.6%, and negative predictive value are unsati sfactory. PCR 1 and 2 are superior to the other techniques to a statistical ly highly significant degree in terms of sensitivity, but the difference be tween the two is not significant. (C) 2000 Elsevier Science B.V. All rights reserved.