Surveillance for early detection of low pathogenicity avian influenza in poultry

Infection with low pathogenicity avian influenza (LPAI) virus is widespread and has led to outbreaks in domestic birds in many countries. Although infection does not pose a serious concern for animal heath, LPAI virus subtypes H5 and H7 can mutate into the highly pathogenic form (HPAI), which can cause massive epidemics in poultry and threaten human health. However, the knowledge of the underlying mechanisms of mutation from LPAI to HPAI is insufficient for predicting these mutations, especially because the molecular changes necessary for the virulence shift seem to occur at random. Hence, to reduce the likelihood of HPAI emergence, LPAI circulation in general should be reduced. To this end, early detection of LPAI infections in domestic poultry is fundamental. The main objective of this thesis was to develop a framework of surveillance strategies for the early detection of LPAI infections in poultry. The studies described in this thesis were based on surveillance data collected during the LPAI epidemics in Italy in the period 2000-2009, which involved both the industrial and rural sectors. First of all, the Italian monitoring system, which is the longest running system in the EU, was evaluated in terms of its capacity to detect outbreaks of LPAI infection in poultry, with particular regard to fattening turkeys, which has been the most affected species. It was shown that the majority of outbreaks were detected by active surveillance, which was especially successful in vaccinated flocks, given the vaccine-induced decrease in the severity of clinical signs and the consequent reduction in the sensitivity of passive surveillance. Subsequently, the involvement of the rural poultry sector in LPAI outbreaks was investigated. Grower and dealer holdings probably play a major role in perpetuating the infection, due to the peculiar husbandry and trade practices adopted, which facilitate the circulation and persistence of LPAI viruses in the rural sector. After that, the transmission dynamics of LPAI infections in industrial turkey flocks was modelled based on field data, and it was shown that these dynamics varied greatly among flocks. Subsequently, the sensitivity and specificity of diagnostic assays for detecting LPAI antibodies were estimated, without considering any assay as a gold standard. The estimates confirmed that the hemagglutination inhibition (HI) test, which is generally accepted as the gold standard, has a near perfect accuracy and may be considered as a reference test; however, to perform LPAI surveillance, for which large numbers of samples have to be processed, valid alternatives exist. Finally, different surveillance strategies were evaluated in terms of their performance for early warning; it was shown that the prompt detection of LPAI outbreaks in turkeys can be achieved through active surveillance by increasing the sampling frequency to once every 15 days, though very frequent sampling may not be sustainable in the long term. In the last chapter, the main findings of the above mentioned studies were integrated and critically discussed coming up with a framework of surveillance strategies for the early detection of LPAI infections in industrial poultry.