Risk factors of new Aujeszky's disease virus infection in swine herds in Brittany (France)

Abstract - Brittany is the main pig production area in France. About 14 million fatteners are produced every year in this province. There are about 9 300 farms holding 690 000 sows. Vaccination against Aujeszky's disease is compulsory, and all vaccines which are used are marker vaccines. Since 1994, a yearly general serological screening has been performed. All herds are bled at least once a year. Fifteen sows are bled in each breeding stocks and 15 fatteners in each fattening stock. All serological analyses are GE ELISA. The seroprevalence of herds has been decreased dramatically during this period. Twenty four percent of the herds had at least one positive serum in 1995 whereas only 4.5% were positive in 1998. Nevertheless, some new infections still occur. A case control study was carried out in 80 commercial herds to identify and to rank the risk factors which are associated with new infections. Different possible risk factors were assumed and classified regarding into 3 categories according to consecutive steps of contamination of a herd. These 3 categories were: Cases and control herds were chosen out of the population of all herds with at least one serological control per year for 3 consecutive years. These 3 years were from the first of July 1994 to the 30th of June 1997. Both cases and controls were seronegative during the first 2 years. Control herds remained free of infection on the third year. On the contrary, newly infected cases were defined as those herds which became seropositive on the third year. In order to eliminate potential false positives, herds with less than 3 positive sows or 2 positive fatteners were excluded from the analysis. Forty four case herds met these criterion. Forty of them were studied. Seventy percent of the cases were Farrow-to-Finish herds (FAFI), 28% were Feeder-to-Finish herds (FEFI) and the last farm was a Farrow-to-Feeder (FAFE). Controls were individually matched to cases considering their herd type (FAFE, FAFI, FEFI), size (the difference in herd size had to be lower than 30%), location (the control had to be located within 25 kilometres from the case) and producer organisation (in order to facilitate data collection). Variables describing the possible risk factors were obtained (1) with a questionnaire describing the herd and herd management (a single investigator performed the survey in all herds) and (2) by using the regional database which describes the location and the serological status of all herds in Brittany, in order to assess the risk of being contaminated by the neighbourhood. The herd questionnaire contained 95 variables which described the purchase of pigs (9 variables), passive vectors (21), quarantine (8), bio-security (16), herd structure (8) and vaccination scheme (33). Seven variables were calculated to describe the risk associated to neighbourhood. These variables described pig density and the number of infected herds within 1, 3 and 6 kilometre radiuses. The statistical analysis was performed with 2 different stages: Bio-security is an example of a summary variable defined after the first step of analysis. 9 variables are summarised in it. Although the real world is not black and white, large differences between the groups of herds with good or bad bio-securities were observed. Seventy-eight percent of the farms with a good bio-security had a locker and clothes for visitors whereas 79% of the bad group had none. In the same way the good bio-security group was better for locking the door (78% did whereas 90% of the bad group did not), more of them avoided the entry of slaughterhouse lorry drivers within the premises (88% did whereas 62% of the bad group did not). Considering the logistic regression, three risk factors were associated with the new infections: (1) the presence of at least one herd within 1 kilometre from the case ( , ); (2) a high number of infected herds between 1 and 6 kilometres ( , ). A "high number" was more than 3 infected herds between 1 and 3 kilometres and more than 11 infected herds between 3 and 6 kilometres; (3) the frequency of piglet deliveries ( , ). No relation was shown between new infection by Aujeszky's disease virus and other variables including, quarantine, herd-structure, vaccination and bio-security. The results suggest that the two main risk factors for new infection by Aujeszky's disease virus are airborne transmission and purchase of infected piglets. The fact that the other potential risk factors, in particular the quality of vaccination, were not found to be associated with risks of new infection, has to be further discussed. A good vaccination does not prevent herds from being infected by Aujeszky's disease virus. Nevertheless, vaccination remains necessary in high density areas to prevent neighbouring herds from being contaminated by infected herds and to achieve eradication of Aujeszky's disease.