Comparison of three rapid detection systems for type A influenza virus on tracheal swabs of experimentally and naturally infected birds.

The present paper reports of the comparison between three rapid virus detection systems and virus isolation (VI) from pooled tracheal swabs collected from naturally and experimentally infected birds with a low pathogenicity avian influenza virus of the H7N3 subtype. The relative sensitivity, specificity and agreement (K value) were calculated for a commercial antigen capture enzyme immunoassay (AC-EIA) and for two nucleic acid detection tests, a one-step reverse transcriptase-polymerase chain reaction (RT-PCR) and a real-time RT-PCR (RRT-PCR), both targeting the M gene. The results indicate that in experimentally infected turkeys VI was positive from the pooled tracheal swabs collected from day 3 to day 10. One-step RT-PCR was able to detect influenza RNA from samples collected from day 3 to day 12, while RRT-PCR amplified influenza RNA in swabs collected from day 3 to day 15. The AC-EIA test yielded positive results between day 5 and day 10 post-infection. On field samples, the K value between the AC-EIA and VI tests was 0.82. Compared with VI, the relative sensitivity of this test was 88.9% (CI95 = 85.2-92.6) and the relative specificity was 95.7% (CI95 = 93.7-97.7). The K value between the RT-PCR and VI tests was 0.88. Compared with virus isolation, the relative sensitivity of the one-step RT-PCR was 95.6% (CI95 = 93.1-98.0) and the relative specificity was 96.3% (CI95 = 94.4-98.1). The K value between the RRT-PCR and VI tests was 0.92. Compared with virus isolation, the relative sensitivity and specificity of RRT-PCR was 93.3% (CI95 = 90.4-96.3) and 98.4% (CI95 = 97.2-99.6), respectively. Generally speaking, comparison between virus isolation, the AC-EIA test and the two nucleic acid detection methods indicated excellent agreement. Data obtained from both experimental and field study suggest a higher sensitivity of the PCR-based methods compared with the AC-EIA. The economical and practical implications of using one of the rapid tests as an alternative to VI during an avian influenza epidemic are discussed. Copyright 2004 Houghton Trust Ltd