Abstract
Highly Pathogenic Avian Influenza viruses (HPAIVs) display a tissue pantropism, which implies a possible spread in feathers. HPAIV detection from feathers had been evaluated for H5N1 or H7N1 HPAIVs. It was suggested that viral RNA loads could be equivalent or higher in samples of immature feather compared to tracheal (TS) or cloacal swabs (CS). We investigated the suitability of feathers for the detection of clade 2.3.4.4b H5N8 HPAIV in ducks and geese field samples. In the six H5N8 positive flocks that were included in this study, TS, CS and immature wing feathers were taken from at least 10 birds. Molecular loads were then estimated using real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) targetting H5 and M genes. In all flocks, viral loads were at least equivalent between feather and swab samples and in most cases up to 103 higher in feathers. Bayesian modelling confirmed that, in infected poultry, RT-qPCR was much more likely to be positive when applied on a feather sample only (estimated sensitivity between 0.89 and 0.96 depending on the positivity threshold) than on a combination of a tracheal and a cloacal swab (estimated sensitivity between 0.45 and 0.68 depending on the positivity threshold). Viral tropism and lesions in feathers were evaluated by histopathology and immunohistochemistry. Epithelial necrosis of immature feathers and follicles was observed concurrently with positive viral antigen detection and leukocytic infiltration of pulp. Accurate detection of clade 2.3.4.4b HPAIVs in feather samples were finally confirmed with experimental H5N8 infection on 10-week-old mule ducks, as viral loads at 3, 5 and 7 days post-infection were higher in feathers than in tracheal or cloacal swabs. However, feather samples were associated with lower viral loads than tracheal swabs at day 1, suggesting better detectability of the virus in feathers in the later course of infection. These results, based on both field cases and experimental infections, suggest that feather samples should be included in the toolbox of samples for detection of clade 2.3.4.4b HPAI viruses, at least in ducks and geese.