A summary of some of the practical difficulties that arise in using NSP ELISA to help substantiate FMD freedom is provided in Supplementary Table 4. Three workshops in 2007 examined the design and interpretation of post FMD-vaccination serosurveillance by NSP tests [52]. Their aim was to test the feasibility and consequences of applying the above-described rules after applying emergency

vaccination in three plausible scenarios involving different outbreak sizes, affected species and livestock densities. The summary recommendations of the workshops are provided in Supplementary Table 5 and the following key issues are further discussed below: (1) the requirement to sample all vaccinated selleck products animals; (2) the follow-up investigation required to establish the significance of seroreactors identified;

(3) the criteria for removal of seropositive animals and herds; (4) what can be done with such animals (slaughter for consumption or destruction); (5) the impact of finding seroreactors during the process of surveillance with the Dactolisib cell line objective of regaining the status “FMD free where vaccination is not practised”. Even with tests of suboptimal sensitivity (70–90%), a low prevalence of infection can be detected with high confidence in large groups of animals without sampling and testing every animal. However, in large herds, the animals are often segregated in smaller groups that may be considered as separate epidemiological

units and in this case, the number of animals per epidemiological unit would be the denominator for calculation of sample sizes. For NSP serosurveillance, using a test with Sp = 0.995 and Se = 0.7, then detection of seroconversion at 95% confidence, at a prevalence of 2%, in an epidemiological unit of 1000 animals, would require 513 animals to be sampled and the cut-point would be five (i.e. finding five or fewer reactors could still be consistent with absence of true seroconversion, i.e. probability of 2% or more seropositive animals is less than 5%). If it were accepted that only strongly seroconverting animals are likely to (have) spread infection, then the Se figure could be increased to 0.9, in which case 366 samples would need to be tested and the cut-point would become four (FreeCalc; [53]). Reduction secondly of the numbers sampled in large herds is often relevant for pigs which also do not have risks associated with the development of FMDV carriers. Clinical disease is also rather obvious in pigs so that NSP surveys add less value. Therefore, surveillance in pigs should be targeted towards the identification of disease and virus circulation. Studies on vaccinated pig herds in Hong Kong suggested an all-or-nothing effect, with widespread clinical disease and NSP seroconversion (49–82% seroprevalence) or neither clinical disease nor seroconversion [54].