Forty-six chickens and 48 ducks were sampled from four Vietnamese poultry premises in 2009 infected with H5N1 highly pathogenic avian influenza (HPAI) clade 2.3.2 and 2.3.4 viruses, which also differed by cleavage site (CS) sequences in their haemagglutinin (HA) genes. All clinical specimens (n0282), namely tracheal and cloacal swabs plus feathers, were tested by five Eurasian reverse-transcriptase AI RealTime polymerase chain reaction (RRT-PCR) methods. Bayesian modelling showed similar high sensitivity for the validated H5 HA2 RRT-PCR and a new modified M-gene RRT-PCR that utilizes lyophilized reagents. Both were more sensitive than the validated ''wet'' M-gene RRT-PCR. Another RRT-PCR, which targeted the H5-gene CS region, was effective for clade 2.3.4 detection, but severely compromised for clade 2.3.2 viruses. Reduced sensitivity of the H5 CS and ''wet'' M-gene RRT-PCRs correlated with mismatches between the target and the primer and/or probe sequences. However, the H5 HA2 RRT-PCR sensitively detected both clade 2.3.2 and 2.3.4 viruses, and agreed with N1 RRT-PCR results. Feather testing from diseased chicken and duck flocks by AI RRT-PCRs resulted in the most sensitive identification of H5N1 HPAI-infected birds. Evolution of new H5N1 HPAI clades remains a concern for currently affected Asian countries, but also for more distant regions where it is important to be prepared for new incursions of H5N1 HPAI viruses. Genetic evidence for adamantane resistance and sensitivity was also observed in isolates from both clades.

Please cite this paper as: Slomka et al. (2012) Evaluation of lateral flow devices for identification of infected poultry by testing swab and feather specimens during H5N1 highly pathogenic avian influenza outbreaks in Vietnam. Influenza and Other Respiratory Viruses 6(5), 318–327. Background  Evaluation of two commercial lateral flow devices (LFDs) for avian influenza (AI) detection in H5N1 highly pathogenic AI infected poultry in Vietnam. Objectives  Determine sensitivity and specificity of the LFDs relative to a validated highly sensitive H5 RRT PCR. Methods  Swabs (cloacal and tracheal) and feathers were collected from 46 chickens and 48 ducks (282 clinical specimens) and tested by both LFDs and H5 RRT PCR. A subset of 59 chicken and 34 duck specimens was also tested by virus isolation (VI), the ‘gold standard’. Results  Twenty‐six chickens and 15 ducks were shown to be infected by at least one RRT PCR positive clinical specimen per bird. Bird‐level sensitivity for the Anigen LFD was 84·6% for chickens and 53·3% for ducks, and for the Quickvue LFD 65·4% for chickens and 33·3% for ducks. Comparison of the three clinical specimens revealed that chicken feathers were the most sensitive with 84% and 56% sensitivities for Anigen and Quickvue respectively. All 21 RRT PCR positive swabs from ducks were negative by both LFDs. However, duck feather testing gave sensitivities of 53·3% and 33·3% for Anigen and Quickvue respectively. Specificity was 100% for both LFDs in all investigations. Conclusions  Although LFDs were less sensitive than AI RRT PCR and VI, high titre viral shedding in H5N1 highly pathogenic avian influenza (HPAI) infected and diseased chickens is sufficient for a proportion of birds to be identified as AI infected by LFDs. Feathers were the optimal specimen for LFD testing in such diseased HPAI scenarios, particularly for ducks where swab testing by LFDs failed to identify any infected birds. However, specimens should be forwarded to the laboratory for confirmation by more sensitive diagnostic techniques.