| Abstract: | Data combining sequential bacteriology and serology from a longitudinal study of a dairy herd were used to demonstrate the limitations of serology as a diagnostic method in cross-sectional sampling of bovine populations. Whole-herd point serological prevalences showed considerable variation over a two-year sampling period (38.8 to 76.2 per cent), and this was mainly due to varying age-specific prevalence. Owing to the rapid decline in titres and the varying persistence of infection, point serological prevalences failed to approximate to cumulative infection rates (either past or present) at different times of the year. A higher estimate of the number of susceptible animals in the herd than is the case results in inaccurate information on true incidence rates and can confuse assessments of the susceptibility of different age groups, especially if only small numbers are sampled. A sampling exercise demonstrated that a 10-cow sample usually provided little useful information other than establishing the presence or absence of hardjo in the herd. Increasing the sample size markedly improved epidemiological information, investigations of clinical disease, assessments of vaccination needs and public health tracebacks. Preferably 10 sera from each of the yearling, first calver, second calver and older age groups should be tested. Serology was an inadequate indicator of infection in individual animals. Group geometric mean titres taken from a mean serological response curve were shown to have limited application in the interpretation of field data, unless infection had occurred in the previous two months. |
