Early detection of avian influenza virus circulation in wild birds is critical for anticipating infection pressure on poultry and regional virus dissemination. Current European legislation primarily targets dead, sick or hunted wild birds, whereas environmental sampling is not formally integrated into surveillance programmes. In this study, we compare the sensitivity of wild bird and environmental sampling strategies for avian influenza surveillance and evaluate the added value of environmental sampling.
During the 2024–2025 high pathogenicity avian influenza season in north-eastern Italy, we sampled wild birds and their environment in high-risk wetlands. Captured and hunted wild birds were tested using oropharyngeal, cloacal, feather swabs and brain samples; environmental samples included fresh faecal droppings and surface water collected using passive samplers. Bayesian latent class models were used to estimate the sensitivity of individual sample types (bird level) and surveillance strategies (field-visit level), and virus prevalence.
At the bird level, the estimated sensitivity for H5 high pathogenicity avian influenza detection was highest for brain samples (0.82, 95% credible interval?=?0.58–0.99), followed by oropharyngeal swabs (0.63, 95% credible interval?=?0.25–0.98), whereas cloacal (0.05, 95% credible interval?=?0.00–0.22) and feather swabs (0.09, 95% credible interval?=?0.00–0.32) had substantially lower sensitivity.
At the field-visit level, environmental-based strategies had the highest estimated sensitivities for M-gene detection: 0.77 (95% credible interval?=?0.63–0.89) for surface water and 0.71 (95% credible interval?=?0.57–0.84) for faecal samples. Estimated prevalence among hunted and captured wild birds was 5.2% (95% credible interval?=?2.4%–8.6%) for M gene, 3.1% (95% credible interval?=?1.3%–6.3%) for H5 subtype and 2.7% (95% credible interval?=?1.1%–6.0%) for H5 high pathogenicity avian influenza.
Synthesis and applications. Environmental sampling strategies showed higher sensitivity than bird-based samples and, although relying mostly on M-gene detection, can complement wild bird sampling by increasing opportunities to detect avian influenza virus circulation at high-risk wetlands without depending on access to infected birds. Integrating rapid environmental screening with targeted follow-up sampling in birds could strengthen early-warning systems, improve spatial targeting of surveillance and support more timely measures to reduce the risk of virus introduction into poultry populations.