Mathematical epidemiology and the control of classical swine fever virus

Abstract

The thesis describes epidemiological research carried out to improve the control of classical swine fever virus (CSFV) epidemics. First, control measures and strategies are investigated in order to enable quick and adequate action upon CSFV detection. CSFV transmission experiments are described, as well as a mathematical model of CSFV transmission between herds, both used to assess the effectiveness of the E2 subunit marker vaccines with respect to CSFV transmission. Quantification of the effectiveness was done with the ‘basic reproduction ratio’ Ri (Rh), defined as the number of animals (herds) that is infected by one typical infectious animal (herd) in a completely susceptible population. The transmission experiments pointed out that Ri decreases to below the threshold value 1 as of three weeks after vaccination until at least six months after vaccination. Already one week after vaccination of a pig herd, virus entry will lead to only minor outbreaks. It appeared that the presence of maternal antibodies at the time of vaccination may reduce the antibody levels at later age, but these lower levels still keep Ri below 1. Because vertical transmission is not completely blocked by the vaccine, persistently infected piglets might be born on vaccinated herds. These herds cannot be clinically detected and might thus be a risk for further virus spread. Detection will be possible if some animals, e.g. the breeding sows, on those herds remain unvaccinated. Whether that would lead to an effective control strategy was tested in a mathematical model of CSFV transmission. The model was based on the Dutch CSFV epidemic of 1997/1998, with a moderately virulent virus strain in a pig dense area with relatively many multiplier herds. The model results point out that an effective control strategy (Rh < 1) requires a complete prohibition of transport of unvaccinated animals. Moreover, in addition to the control measures that are obliged by EU legislation (like the tracing of infectious contacts and hygiene measures), the virus transmission between herds should be halved, e.g. by vaccinating 50% of all pig herds. If all animals but the breeding sows would be vaccinated, the demands for a successful control strategy are well met. In addition to the investigation of control measures and strategies, a tool was tested, which had been developed to analyse the effectiveness of the control strategy in an ongoing epidemic. The tool was designed to use data of an ongoing epidemic for estimation of Rh and the number of infected, yet undetected herds. The only data needed for the calculations are the numbers of detected herds in each week of the epidemic. Unfortunately, the only result the model could generate was whether Rh was smaller or larger than 1. Further mathematical research is needed to understand the shortcomings. Until then, some extra data might be used, e.g. an estimate of the number of infected herds at the day of the first detection, or an estimate of the average time between infection of a herd and detection.