1. INTRODUCTION

1.1 Background

In recent years the numbers of BSE cases in the Republic of Ireland have been increasing, although they continue to remain at a relatively low level. This has led to a review of the BSE controls that are in place, and questions about the potential risk from beef on the bone.

One control introduced to safeguard human health early in the epidemic is the removal of all tissues in which any infectivity may be present in an infected animal. These tissues, referred to as Specified Risk Material (SRM), include the brain and spinal cord as well as other tissues. In 1997 experiments being carried out by MAFF in the UK found infectivity in the dorsal root ganglia (DRG) of experimentally infected cattle. DRG are part of the peripheral nervous system and are located within the vertebral column but were not included in the definition of SRM.　A problem arose because the DRG would not normally be removed with the spinal cord when that is removed at the time of slaughter.　Some cuts of beef, notably T-bone steaks and some cuts of rib of beef are sold to the consumer with the vertebral column; there is therefore the possibility that people could be exposed to potentially infectious material.

In 1998, the UK Government took the decision to ban sales of beef on the bone following the discovery of infectivity in dorsal root ganglia. Sales of beef on the bone were allowed to resume in 2000 following a review by the Chief Medical Officer. In the UK this only applies to cattle aged under 30 months as older animals are excluded from food for human consumption. Subsequently, the EC took a decision to ban sales of beef on the bone across Europe. This came into effect in April 2001.

At the request of the CJD Advisory Committee of the Department of Health and Children, the Food Safety Authority of Ireland (FSAI) established a DRG Task Force and decided to carry out a risk assessment. DNV were asked to undertake this risk assessment as they had carried out the study for the UK Ministry of Agriculture Fisheries and Food in 1997 that helped inform the advice given by the Spongiform Encephalopathy Advisory Committee to government.

1.2 Study Objectives

The objective of the study is to quantify, so far as is possible, the risks to the population of Ireland from the consumption of beef products due to the possible contamination with BSE infectivity in the dorsal root ganglia (DRG) of cattle, and to assess the risk reduction from banning the sale of beef on the bone.

1.3 Approach

The method used for this study has closely followed the approach used for the UK MAFF in 1997 (Assessment of Risk from Possible BSE Infectivity in Dorsal Root Ganglia, DNV Report C7831, December, 1997). The approach has been updated to take account of any new research or data, and applied to the situation in Ireland.

In order to assess the risk from infectivity in DRG it is necessary:

1. To estimate the expected number of cattle slaughtered for human consumption with significant levels of BSE infectivity .
2. To assess the level of infectivity in DRG in an infected animal.
3. To assess what happens to the DRG when meat is cut from the bone, or when bone-in cuts (e.g. rib of beef or T-bone steaks) are consumed.

This information is then combined in an event tree that shows what happens to the DRG in a slaughtered animal and allows the fraction of DRG consumed to be calculated. The event tree model is developed as an excel spreadsheet. There is significant uncertainty about much of the data on which an assessment of this sort is based. The range of uncertainty in the data can be modelled using a probabilistic risk assessment approach such as Monte Carlo simulation.　To do this a　range and distribution is defined for each of the input factors, and a Monte Carlo simulation tool is used to predict the range of possible values for the result. A software tool called Crystal Ball (Crystal Ball Version 4.0; Decisioneering Inc, Denver, Colorado), that works with Excel spreadsheets has been used to carry out the simulation.

Monte Carlo simulation is a well established technique that enables the modeller to take account of the chance variation that is inherent in most real life problems. The simulation is based on the use of random number generators to select a value from each of the input parameters for which a distribution has been defined. Over a number of iterations the simulation enables the full range of possible values for each variable to be tested, but weights each scenario by its probability of occurrence. In each iteration a new value is selected for each variable and an output value (or values) calculated. This is repeated a large number of times so that a distribution of the output value is built up.