Epidemiological approaches to safety investigations

Abstract

This paper considers the different approaches to post-authorisation safety monitoring of veterinary medicinal products that is essential to ensure confidence in their safety. Most safety testing is undertaken prior to granting of a marketing authorisation and is generally on a small scale. Field trials are usually much larger, but still involve relatively low numbers of animals compared to the number to which authorised products are administered. Safety testing is generally aimed at detecting common events; the numbers of animals used in the tests are too small for detection of all but the most common reactions. The efficiency of the tests depends on the frequency and severity of the adverse reaction and the ability to associate the adverse event with the product. The latter is affected by the period of time between administration and the event, as well as by its underlying frequency. Adverse reaction surveillance is critical in monitoring the safety of a marketed product. Most is entirely passive and so reporting rates are likely to underestimate true incidence. It is relatively efficient for rare, serious adverse effects and for those with a low underlying frequency in the population, but it is less useful when there is long period between administration and the event, or where the event has a relatively high underlying frequency. Greater emphasis should be placed on active surveillance after production registration. Detailed epidemiological investigations, including cohort, case control and cross-sectional designs, offer the only approaches that provide more information on the association between a product and events that have a high underlying frequency in the population or where there is a long period between administration and the adverse event. The relative merits of different approaches are discussed, with particular reference to our recently published study of the temporal association between canine vaccination and non-specific signs of ill health and plans to undertake studies of associations with feline injection site sarcoma. Emphasis is placed on the need for clearly stated hypotheses and the consideration of equivalence, rather than significance testing when considering safety studies.