BackgroundSince its first report in 2007, avian influenza (AI) has been endemic in Bangladesh. While live poultry marketing is widespread throughout the country and known to influence AI dissemination and persistence, trading patterns have not been described. The aim of this study is to assess poultry trading practices and features of the poultry trading networks which could promote AI spread, and their potential implications for disease control and surveillance.Data on poultry trading practices was collected from 849 poultry traders during a cross-sectional survey in 138 live bird markets (LBMs) across 17 different districts of Bangladesh. The quantity and origins of traded poultry were assessed for each poultry type in surveyed LBMs. The network of contacts between farms and LBMs resulting from commercial movements of live poultry was constructed to assess its connectivity and to identify the key premises influencing it.ResultsPoultry trading practices varied according to the size of the LBMs and to the type of poultry traded. Industrial broiler chickens, the most commonly traded poultry, were generally sold in LBMs close to their production areas, whereas ducks and backyard chickens were moved over longer distances, and their transport involved several intermediates. The poultry trading network composed of 445 nodes (73.2% were LBMs) was highly connected and disassortative. However, the removal of only 5.6% of the nodes (25 LBMs with the highest betweenness scores), reduced the network’s connectedness, and the maximum size of output and input domains by more than 50%.ConclusionsPoultry types need to be discriminated in order to understand the way in which poultry trading networks are shaped, and the level of risk of disease spread that these networks may promote. Knowledge of the network structure could be used to target control and surveillance interventions to a small number of LBMs.Electronic supplementary materialThe online version of this article (10.1186/s12917-018-1331-5) contains supplementary material, which is available to authorized users.

Avian influenza viruses (AIVs) continue to pose a global threat. Waterfowl are the main reservoir and are responsible for the spillover of AIVs to other hosts. This study was conducted as part of routine surveillance activities in Bangladesh and it reports on the serological and molecular detection of H5N1 AIV subtype. A total of 2169 cloacal and 2191 oropharyngeal swabs as well as 1725 sera samples were collected from live birds including duck and chicken in different locations in Bangladesh between the years of 2013 and 2014. Samples were tested using virus isolation, serological tests and molecular methods of RT-PCR. Influenza A viruses were detected using reverse transcription PCR targeting the virus matrix (M) gene in 41/4360 (0.94%) samples including both cloacal and oropharyngeal swab samples, 31 of which were subtyped as H5N1 using subtype-specific primers. Twenty-one live H5N1 virus isolates were recovered from those 31 samples. Screening of 1,868 blood samples collected from the same birds using H5-specific ELISA identified 545/1603 (34%) positive samples. Disconcertingly, an analysis of 221 serum samples collected from vaccinated layer chicken in four districts revealed that only 18 samples (8.1%) were seropositive for anti H5 antibodies, compared to unvaccinated birds (n=105), where 8 samples (7.6%) were seropositive. Our result indicates that the vaccination program as currently implemented should be reviewed and updated. In addition, surveillance programs are crucial for monitoring the efficacy of the current poultry vaccinations programs, and to monitor the circulating AIV strains and emergence of AIV subtypes in Bangladesh.