New developments in quantitative risk assessment of campylobacteriosis

Quantitative microbiological risk assessment (QMRA) is now broadly accepted as an important decision support tool in food safety risk management. It has been used to support decision making at the global level (Codex Alimentarius, FAO and WHO), at the European level (European Food Safety Authority), in many contries world wide and by the industry. The human health risks associated with Campylobacter on broiler meat have been the subject of many different risk assessment studies, whereas only a few studies have considered other exposure pathways. Models are now available for all stages of the farm-to-fork chain. At the farm, all available models suggest a very rapid spread of Campylobacter infection in commercial broiler flocks. After infection of the first birds, a within-flock prevalence > 95% may be reached within one week. This implies that high-level infection in flocks at slaughter may frequently go undetected after on-farm monitoring. Processing models indicate that defeathering and evisceration are key processes leading to carcass contamination, but there is still considerable uncertainty of the quantitative effects of these processes. In the kitchen, cross-contamination from contaminated meat to ready-to-eat foods is the main pathway of consumer exposure. Undercooking appears to be of minor importance. However, this conclusion may need to be reconsidered in the light of increasing consumption of minced meat preparations. Five QMRA models have been compared in detail, and detailed technical information is now available in CRAF- the Campylobacter Risk Assessment Framework (www.rivm.nl/craf). Newly published QMRA studies on broiler meat arrive at similar conclusions. A general conclusion of all models is that highly contaminated carcasses contribute most to risk, and that reducing Campylobacter counts by >100-fold will result in significant risk reduction. There are many possible options to reduce counts, and QMRA models combined with economic models are increasingly used to estimate effectiveness and efficiency of a broad range of interventions. QMRA models can also be used to support the formulation of quantitative criteria for broilers and other meats. New modelling methods (e.g. Bayesian approaches) are being proposed, but are currently less detailed than conventional Monte Carlo simulation models. There is considerable uncertainty in dose-response relationships and in particular, current models do not consider the effects of acquired immunity. Evidence is increasing that asymptomatic infections occur at high frequencies in developing countries as well as in the industrialised world. Partly, these asymptomatic infections may be explained by the protective effects of acquired immunity. Alternative explanations include differences in virulence characteristics among the naturally occurring Campylobacter strains, or low disease risks associated with infection by low doses. Further research and modeling efforts are needed to understand the relative importance of these different factors.