The 1996 Pathogen Reduction: Hazard Analysis and Critical Control Point Final Rule was one of the earliest claims that processing plants can control the prevalence of pathogenic bacteria such as Salmonella on raw meat and poultry. Previously, most authorities believed that Salmonella controls could be successful only at the farm level or in post-processing kill steps such as cooking or irradiation. The Pathogen Reduction: Hazard Analysis and Critical Control Point Final Rule said that antibacterial treatments and microbiological testing would allow plants to improve their control over Salmonella prevalence, mostly through using visual inspection and E. coli testing to avoid or mitigate fecal contamination, the main source of Salmonella in the processing plant according to the Food Safety and Inspection Service.

Preliminary testing of HACCP in processing plants was inconclusive, so the industry-wide introduction of HACCP was not an application of recognized scientific principles, but was essentially an experiment to test the FSIS Salmonella control hypothesis in several hundred processing plants. The three-way relationship among fecal contamination, E. coli counts, and Salmonella prevalence would allow control steps to reduce the prevalence of Salmonella, with other pathogens affected in a similar way, according to the Final Rule.

The relationship between E. coli counts and Salmonella prevalence in chilled carcass samples was discussed previously in this series of articles, but what about the ability of visible fecal contamination to indicate increased likelihood of the presence of Salmonella, or the use of high E. coli counts to indicate the occurrence of fecal contamination earlier in processing? What is known from scientific studies completed since the introduction of HACCP in processing plants?

Visible fecal contamination as an indicator of Salmonella  

What evidence exists concerning the correlation between visible fecal contamination and the presence of Salmonella on poultry? Consider the following:


  • In several studies in operating processing plants, chicken carcasses contaminated with visible ingesta or visible feces were not more likely to be Salmonella positive after chilling.
  • In a different study, non-compliance reports for fecal contamination were not positively correlated with Salmonella prevalence in processed chicken.
  • Several studies indicated that external Salmonella caused more cross-contamination than Salmonella in feces within the intestinal tract, including studies done by FSIS before adoption of HACCP.


At the Texas Poultry HACCP Roundtable on May 4, 2006, one of the participants asked, “Will the FSIS personnel in the Dallas district begin to recognize the growing conventional wisdom that there is no direct correlation between fecal material and the presence of Salmonella? How will that affect local inspection and frontline supervision?”


The FSIS representatives answered, “District office personnel recognize this and are making an effort to get the point across. You still have to meet the zero-tolerance standard (381.65e), but this is not linked to Salmonella contamination. We hope that everyone has interventions directed at Salmonella other than zero tolerance, since 381.65e is not an intervention strategy for Salmonella.”

Post-chill E. coli as an indicator of fecal contamination  

What evidence exists that post-chill E. coli counts are an indicator of previous fecal contamination? Equating higher E. coli counts with previous fecal contamination was never more than an assumption by FSIS. There has never been any claim that baseline carcass rinse samples were identified as being taken when the processing plant was in or out of control with respect to fecal contamination.


  • In several studies in operating processing plants, chicken carcasses with visible ingesta or visible feces were not more likely to have higher E. coli counts after chilling.
  • In several lab studies, half carcasses that were intentionally contaminated with feces did not have higher E. coli counts after chilling when compared to paired carcass halves that were not fecally contaminated. When the site of fecal contamination was tested against a skin sample from a similar control site on the opposite side of chilled carcasses, E. coli counts did not identify the previously contaminated sample, even though it had been completely covered with feces.
  • In multiple studies, samples taken near the cloaca did not have higher E. coli counts than samples taken from more distant locations on the carcass.
  • Multiple papers have reported poor or non-existent relationships between counts of various indicator bacteria with observed levels of hygiene during processing.


The net result of these observations is that FSIS assumptions about fecal contamination as the in-plant source of most Salmonella have not been confirmed by operating experience or scientific research. The usefulness of E. coli as an indicator of fecal contamination is also in doubt. The original use of E. coli as an indicator was in water samples where it was not expected to be found, but most chicken carcasses are routinely positive for E. coli, and higher counts have not turned out to be good predictors of fecal contamination that occurred in the processing plant and could have been avoided or cleaned up. The possibility that high counts of E. coli could indicate increased likelihood of Salmonella is even more unlikely.

FSIS assumptions not confirmed  

The basic microbiological assumptions for the argument that Salmonella can be controlled in the processing plant have not been confirmed.