EFSA Issues Opinion on Salmonella Link in Poultry Production Chain

If the current flock prevalence target is changed to a prevalence target for all serovars there may be a greater correlation between the occurrence of flock infection and the prevalence of Salmonella in carcasses.

30 Mar 2010 --- Following a request from the European Commission, the Panel on Biological Hazards (BIOHAZ Panel) was asked to provide technical assistance on the link between Salmonella criteria at different stages of the poultry production chain. In particular, the BIOHAZ Panel was firstly asked to provide guidance on the detailed rules of a food safety criterion (FSC) for fresh poultry meat during shelf life (n and c values) with the purpose to be in accordance with the existing process hygiene criterion (PHC) on carcasses under conditions of processing and storage. The BIOHAZ Panel was further asked to provide guidance on a possible revision of both the FSC in fresh poultry meat and the PHC when the target for reduction of Salmonella in broilers has been achieved (1% or less flocks remaining positive for Salmonella enterica serovar Enteritidis or Salmonella enterica serovar Typhimurium).

With reference to the first question figures are presented in this Scientific Opinion to illustrate an FSC for poultry carcasses and meat portions which will result in a regulatory outcome that is in accordance with or has a similar performance as that of the current PHC for poultry carcasses. The assumptions and limitations that need to be considered for this comparison are discussed.

The calculations should theoretically apply to meat from all poultry species and sectors, but specific studies have mainly been carried out in broiler meat production including an EFSA commissioned longitudinal study on the fate of Salmonella on broiler meat before and after cutting and/or deboning. The results of this study are presented in a separate document and its analysis is included in the current document. However, there are major differences in the slaughter and further processing of other types of meat birds that would be expected to result in significantly different levels of contamination. This variability is not reflected in this Scientific Opinion because of the lack of data.

Further on and in order to answer the second question, a reasoned discussion is presented aimed at providing an indication on the way both the Salmonella PHC and FSC in poultry meat should be reviewed once the current target for Salmonella in life poultry is met (i.e. 1% or less flocks positive to S. Enteritidis or S. Typhimurium). It has to be understood that this Scientific Opinion does not take into account either public health aspects of Salmonella levels or targets in broiler meat or in broiler flocks, nor the relevance of the different Salmonella serovars.

For the first question the BIOHAZ Panel concluded that there are major differences in the purpose between a PHC and an FSC, which make comparison difficult and which make it impossible to suggest food safety criteria which are precisely in accordance with the existing process hygiene criteria. Different points of sampling in the food chain may provide different estimates of the prevalence, dependant on the degree of cross-contamination and inactivation of organisms and differences in the definition of batches (origin of carcasses or products). In addition the sampling methods (pooled neck skin samples, carcass washes, 25 g single meat samples) may differ. It is estimated that with the existing PHC for Salmonella in poultry carcasses of broilers and turkey (n=50 and c =7) and with 7 samples positive (compliance), the best estimate of the pooled prevalence is 14% (with 2.5th and 97.5th percentiles of 5.8% and 26.7%). If 8 samples are positive (non-compliance) the best estimate of the pooled prevalence is 16% (with 2.5th and 97.5th percentiles of 7.2% and 29.1%). On the other hand, the best estimate of the carcass prevalence is 4.9% (with 2.5th and 97.5th percentiles of 2.0% and 9.9%). If 8 samples are positive (non-compliance) the best estimate of the carcass prevalence is 5.6% (with 2.5th and 97.5th percentiles of 2.4% and 10.8%).

In order to detect batches which are contaminated from the mean pooled prevalence of 14% and up 26.7%, which is still in accordance with the PHC n=50, c=7, it is concluded that the FSC should lie between n=12-24, c=0 or n=19-36, c=1. If the comparison is based on the carcass prevalence even n=36, c=0 or n=55, c=1 samples are needed to have equal prevalence detected with 97.5% certainty. It has to be noted that independently of the within batch prevalence that an FSC is designed to detect at retail, a food business operator needs to operate at a much lower level in order to ensure that batches are likely to pass. For example, for a sampling scheme with n=12, c=0, a producer needs to operate at a within batch prevalence of less than 0.43% to have more than 95% probability that his batches will pass.

For the second question the BIOHAZ Panel concluded that the prevalence of Salmonella in carcasses at slaughter will depend both on the prevalence of positive flocks, the within flock prevalence and the slaughter hygiene. Thus, some correlation between flock prevalence and carcass prevalence at slaughter is expected. If a target for only Salmonella Enteritidis and Salmonella Typhimurium is set for flock prevalence, the impact on the prevalence of Salmonella on carcasses will depend on the relative contribution of those two serovars to all Salmonella serovars at primary production. Thus, if the current flock prevalence target is changed to a prevalence target for all serovars there may be a greater correlation between the occurrence of flock infection and the prevalence of Salmonella in carcasses.

It is also concluded that if a low flock prevalence is achieved (i.e. 1% or less positive flocks for Salmonella Enteritidis and/or Salmonella Typhimurium or all serovars), reducing the c value or increasing n of the PHC would ensure that monitoring of hygiene performance is maintained. Alternatively, Salmonella PHC could be replaced by a microbiological PHC based on the concentration of indicator microorganisms.  In a situation with this low flock prevalence, the n and c values of the FSC may need to be revised in accordance with progressive changes in the acceptable Salmonella prevalence.

The BIOHAZ Panel makes a series of recommendations on the data that should be collected in order to support further quantitative analysis on the possible correlation between Salmonella targets in broiler flocks and expected contamination levels in the resulting poultry meat, including serovar distribution. Also, the Panel recommends the access to Salmonella PHC data from slaughterhouses in MSs (including serotyping results) since these would provide improved food chain information and would enable the assessment of the extent to which poultry meat conform with different FSC. Finally, the harmonisation of the methodology for testing of poultry meat in the EU is recommended.