Sustainable surveillance strategies for the detection of swine influenza in Vietnam

The occurrence of the 2009 H1N1 influenza pandemic (H1N1pdm09) highlighted the need for sustainable surveillance strategies to monitor swine influenza virus (SIV) evolution and to improve knowledge on swine influenza (SI) epidemiology. Vietnam has the second highest pig population in Asia, and its local context is conducive to inter-species transmission of influenza. I performed an analysis of the pig value chain in Northern Vietnam, from which I selected a collective slaughterhouse, three local slaughterhouses, a weaner market and 15 sentinel familial farms for inclusion in SI pilot surveillance protocols, along with 39 health actors and traders for syndromic surveillance. I found that SIV isolation was only successful at the collective slaughterhouse with 78 viruses obtained. Four groups were identified based on hemagglutinin and neuraminidase sequences. The H1N1pdm09 viruses showed no reassortment, while H1N2 and H3N2 were all reassortants, often with some internal genes from H1N1pdm09. There was virological and serological evidence of circulation of the different hemagglutinin lineages in both the company and familial farming sectors. Monthly seroprevalence was high in slaughterhouses (27%-89%) but low in the market (0-17%). Based on pig origin data, seroprevalence at the pig-level in local slaughterhouses and farm-level in the collective slaughterhouse was significantly higher in companies (100% and 97% respectively) compared to familial farms (60-78% and 74%). In six sentinel farms, seroconversion occurred at an average age of 4 months (2.4-6.4), following maternal antibody decay. The health-related syndromic indicators did not show any clear trends related to influenza circulation. I then considered the cost-effectiveness of the alternative surveillance protocols. Overall the collective slaughterhouse surveillance was the most cost-effective protocol due to successful virus isolation and informative seroprevalence data. Sentinel farm surveillance was also useful as it provided valuable data on SI transmission dynamics in farms. The cost-effectiveness of the other protocols was considered to be lower. I conducted a risk assessment to estimate the risk of SIV circulation in the different farm categories linked to pig introduction, and found that the familial farms most at risk of SI circulation were large farrow-to-finish farms and secondarily large fattening farms. The main outcome of this work was the evaluation of SI surveillance strategies for long term monitoring of SIV evolution and pandemic preparedness, and the identification of surveillance at collective slaughterhouses as having particular value. The implementation of a similar protocol in collective slaughterhouses in Southern Vietnam would allow further study of SIV evolutionary dynamics at the regional level. Risk-based surveillance could be developed in swine farms using the epidemiological results of this study, targeting at risk farms and at risk pig age period (3-4 months old). Further studies are needed to determine the cost-effectiveness of a syndromic surveillance strategy which would allow spatio-temporal sampling in farms based on high incidence of influenza-like illness. One immediate next step in my current study would be an assessment of the pandemic potential of the SIV isolates. More broadly, a One Health approach including surveillance of influenza viruses in animals and humans and the animal-human interface should be developed.published_or_final_versionPublic HealthDoctoralDoctor of Philosoph