Comparison of real-time reverse transcription-PCR and virus isolation for estimating prevalence of avian influenza virus in hunter-harvested wild birds at waterfowl wintering grounds along the Texas mid-Gulf Coast (2005-2006 through 2008-2009).
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
Historically, virus isolation has been the method of choice for conducting surveillance for avian influenza virus (AIV) in avian species. More recently, the primary screening method has become real-time reverse transcription-polymerase chain reaction (RRT-PCR). We wanted to determine how these two testing methods (virus isolation and RRT-PCR) affected AIV prevalence estimation, particularly in an understudied, low-prevalence region-the waterfowl wintering grounds along the Texas mid-Gulf Coast. Cloacal swabs were collected from hunter-harvested waterfowl and other wetland-associated game birds during four consecutive hunting seasons (2005-2006 through 2008-2009). Overall prevalence by RRT-PCR (5.9%, 6.5%, 11.2%, and 5.5%) was approximately an order of magnitude higher than prevalence by virus isolation (0.5%, 1.3%, 3.9%, and 0.7%) for the four hunting seasons, respectively. Apparent AIV prevalence by virus isolation conducted only on RRT-PCR-positive samples resulted in estimates nearly identical in magnitude to those derived from parallel testing (0.5% vs. 0.6%, 1.3% vs. 1.7%, and 3.9% vs. 4.0% for 2005-2006, 2006-2007, and 2007-2008, respectively). Unlike most reports of seasonal variation in AIV prevalence, we documented differences in prevalence estimates among months by RRT-PCR only during 2008-2009 and by virus isolation only during 2006-2007 and 2007-2008. Our data indicate that screening samples by RRT-PCR followed by virus isolation only on RRT-PCR-positive samples provides a reasonable means to generate prevalence estimates close to the true prevalence as determined by virus isolation. We also confirmed the low prevalence of AIV in waterfowl wintering grounds along the Texas mid-Gulf Coast and demonstrated little variation in prevalence among months during the four hunting seasons sampled.