Microbiological cross-contamination has been a contributing factor to several well-documented outbreaks of foodborne illness.[1,2] In most Hazard Analysis and Critical Control Points (HACCP) or other hazard analysis-based food safety systems, cross-contamination is controlled and managed predominately by prerequisite programs (PRPs). PRPs can be defined as the measures that provide the basic environmental and operating conditions in a food operation that are necessary for the production of safe and wholesome foods,[3] such as cleaning and disinfection, and personnel hygiene. The implementation of an appropriate PRP is also seen as the foundation on which a good HACCP plan is built; there are many examples of best practices to follow for PRPs at an international level,[4] via retailers’ requirements[5] or from recognized food research bodies[6–8] or trade associations.[9,10]

There is little information, however, on how to align the use of specific PRPs to control actual routes of cross-contamination in food processing plants. The concept of a ranking system for PRPs has been addressed by ISO 22000,[11] which differentiates operational PRPs (OPRPs) from PRPs. An OPRP is defined as a PRP identified by the hazard analysis as essential in order to control the likelihood of introducing food safety hazards to and/or the contamination or proliferation of food safety hazards in the product(s) or processing environment. In other words, ISO 22000 suggests that a hazard analysis may identify some routes of cross-contamination that are so important to the safety of the food product that their control is essential and are thus elevated to a higher classification of PRP, that is, an OPRP. It is also interesting to note that ISO 22000 recognizes that it is important to not only control cross-contamination of food safety hazards into the product, but also to control cross-contamination within the processing environment. No guidance has been provided, however, as to the hazard analysis steps to be undertaken to determine OPRPs from PRPs.

This article explores two critical concepts of microbial cross-contamination: sources and vectors. It also presents a method for identifying and risk-ranking sources and vectors of contamination, which builds upon previous work described by Smith.[12] Finally, the potential use of OPRPs for control and management of cross-contamination to the product is discussed.

Cross-Contamination Concepts: Sources & Vectors
Pathogenic microorganisms can enter food processing areas from several main routes: the external environment, raw materials, people, equipment and in-plant microbiology laboratories. Once inside, pathogens can be temporary or sporadic visitors (present until they lose viability or are removed via cleaning and disinfection procedures) or they may persist for long periods. When pathogens persist in the environment, they generally survive in harborage sites, which can be defined as physical areas in which pathogens can lodge and be protected from cleaning and disinfection actions, for example, poor hygienic design features of processing equipment or damaged areas of the plant’s building structure. When a harborage site also provides an environment suitable for growth, it can be considered a growth niche. Both harborage sites and growth niches are potential sources of contamination within the processing environment.

For a pathogen to move from a source within the processing environment to other locations (and perhaps even into product), a vector is required. A vector can be defined as anything (air and other gases, water and other liquids, physical objects or people) that carries or transfers a pathogen from one place to another. Vectors may be further described as those that carry a pathogen from a source to another location within the processing environment, that is, an environmental vector, or those that carry a pathogen from a source to the product or product ingredients, that is, a product vector (Figure 1). It should be noted that cross-contamination usually occurs as an event in which a number of vectors may be involved. For example, collecting a product sample from an enclosed process line for quality control analysis by inserting a sampling bag into the product stream by hand may have potential product vectors of the operator’s hand (or glove), the operator’s sleeve, the sampling bag and the air. In another example, a line mechanic may contaminate his or her hands through interaction with a source, subsequently transfer the contamination to a tool and then contaminate a product contact surface with the tool while performing simple maintenance on the line. In this example, the mechanic’s hands may be considered an environmental vector while the mechanic’s tool may be considered a product vector. In other circumstances, a cross-contamination event may have only a single vector, for example, contaminated water droplets from a compressed air line entering a product stream.

The type of vector affects the potential for actual transfer of a pathogen into a product. For example, if a pathogen is being carried in a liquid vector, the liquid may be absorbed into another surface or food completely, which would increase the potential for transfer to be nearly absolute. Conversely, if a pathogen is being carried on a solid vector, such as a mechanic’s tool, the potential for transfer to a secondary surface, including a food product, depends on the physical properties and interaction between the pathogen and the surface as well as the interaction of the vector with the surface. Smith[12] demonstrated that the transfer of microorganisms from one surface to another on contact can be approximated to 50 percent for practical purposes. For stationary air, transfer of microorganisms from the air via sedimentation, which has defined rates for particles of a given size and buoyancy according to Stokes’ law,[13] and the number of microorganisms transferred depends on the microbiological loading of the air and the exposure time. When product is transported via air, or when air is blown over a product for cooling or drying, microorganisms can enter the product via impingement in addition to sedimentation, and the number of microorganisms transferred may be related to the volume of air to which the product is exposed.

Identifying Potential Sources and Vectors of Contamination
Potential pathogen sources and cross-contamination vectors in a processing plant can be determined by a physical examination of the processing environment and may include microbiological sampling. Sources and vectors may be associated with a specific process step or may affect the processing line in general. For example, contaminated air in the production environment might affect many processing steps within a production line, whereas vectors associated with a specific line procedure may be associated solely with a specific process step.

In an exercise similar to determining the product process flow within the HACCP plan, a cross-functional team (comprising personnel knowledgeable about plant operations, sanitation, hygienic design, microbiology and engineering) can be assembled to identify potential sources by walking the line and examining the processing equipment and environment. Potential sources can be determined by numerous means, including dismantling equipment to identify potential harborage sites and niches, as well as an inspection of the environment and building structures. A review of data collected as part of an environmental monitoring program may help identify potential sources in the production environment. However, a history of negative results for a particular pathogen on an environmental site does not indicate that the site is not a potential source for other pathogens, or that the site could not become a source in the future, especially when the construction of the site is not consistent with accepted hygienic design principles.

The observation of all potential sources should be recorded, for example, as indicated in Table 1. In these examples, meat residues were seen inside a meat slicer on/off switch and fluid was seen oozing from underneath a meat slicer equipment foot support plate.

If observational and/or microbiological data identify likely pathogen sources, all potential environmental cross-contamination vectors from this source should be determined to identify the potential to create secondary or temporary sources. Using the equipment foot plate example in Table 1, liquid oozing from under the plate was transferred throughout the process area on an operative’s shoes and equipment wheels and was redeposited at random sites on the floor to act as potential temporary or short-term sources.

The same cross-functional team described above should perform a comprehensive review of the process and environment for potential product cross-contamination vectors. The process and processing environment should be observed during all shifts, when all types of products are produced and when infrequent procedures are performed. In some cases, all personnel may not perform the same task in the same manner. Therefore, interviews of line operators, maintenance staff and quality and sanitation personnel may also help determine potential cross-contamination vectors. Vectors may occur at either fixed and defined time intervals, or randomly. Observations for potential cross-contamination vectors should be made independently of known or likely pathogen sources, as contamination could arise from temporary sites and be transferred to other sites and/or the product stream. It is unlikely that microbiological sampling of vectors would be helpful, as the likelihood of observing a pathogen on a potential vector would be very small.

Observational data for vectors should also be recorded, for example, as indicated in Table 2 for two theoretical spray dryer interventions in a milk-spray drying operation.

Addressing Cross-Contamination Control within a Food Safety Program
The assessment, management and ultimately control of cross-contamination sources and vectors will likely include both the HACCP plan and its foundational PRPs. For example, the cross-functional assessment of the processing environment could be drafted within the plant’s environmental monitoring program. The identification and assessment of product vectors could be incorporated into the HACCP plan. This approach is consistent with established HACCP models, in which hazards that may be introduced at a process step should be considered in the process hazard analysis.

Regardless of how the identification and assessment of cross-contamination sources and vectors are incorporated into a food safety plan, efforts to control them should include reducing the number of possible sources and vectors within a processing environment and developing specific measures to reduce the risk associated with those that remain or are intrinsic to the food production process. For example, the usage of water in certain processing environments may be significantly reduced or eliminated in an effort to control the establishment, growth and movement of pathogens (sources and vectors).

When observing and identifying potential contamination sources and vectors, any current direct controls of observed sources and vectors should be recorded as illustrated in Tables 1 and 2. For vectors, subsequent controls at the process step may have an effect on the hazard that could be transferred by the cross-contamination event, and these should also be recorded.

Practical Application of Risk-Ranking Tools for Sources and Vectors
Although many potential sources and cross-contamination vectors may be identified during the assessment of a processing environment, the degree of control necessary for each source and vector may be determined using risk analysis tools, such as risk ranking. A familiar approach to risk analysis is to consider the likelihood and severity of a hazard on a three-point scale [e.g., LMH (low, medium and high) risk]. A risk analysis for a contamination source may be similar and can be described as the likelihood of a pathogen being present at the potential source and the ability of a pathogen to transfer from this source via an environmental and/or product vector. A risk analysis for cross-contamination vectors may be more complex as it involves three factors: the likelihood of a pathogen being transferred by the vector, the frequency of the event and the severity of the illness if the target consumer ingested the pathogen.

To provide a quantitative approach to evaluate significance, rankings of low, medium or high may be replaced with values of 1, 2 or 3, respectively. These numerical rankings can then be multiplied together to result in an overall risk score associated with a source or vector. Risk scores should always be assessed in the absence of control. In the example provided, this would result in a score range of 1–9 for sources and 1–27 for vectors.

Risk ranking of sources and vectors should be recorded as illustrated in Tables 1 and 2. Undertaking a risk analysis before and after the application of any controls can help identify whether controls are necessary and/or whether current or intended controls are sufficient to reduce either the risk of the source or the contamination event. At a minimum, this allows consideration of the adoption of controls for the uncontrolled sources and vectors identified, which may have an immediate impact on improved food safety. If current controls are not sufficient to adequately control the hazard risk, additional controls must be undertaken. To illustrate this and using the foot plate source as described in Table 1, the frequent use of chlorine disinfectant may not be a sufficient control, and it may be necessary either to lift the equipment, decontaminate the area under the foot plate and then reseal the foot plate to the floor or purchase new foot plates or equipment supports of a more hygienic design.

Subsequent controls should also be considered when assessing the risk of a cross-contamination event. In the theoretical example in Table 2, operatives must insert a guillotine or spray cap into the powder line to prevent clean-in-place (CIP) fluids entering sensitive areas during the dryer CIP program, for example, the bag house where powder is removed from the airflow exiting the dryer. Any microbial contamination entering the dryer, particularly during the removal of the guillotines, would then be subjected to the dryer start-up procedure, which could include the circulation of heated air for several hours (e.g., 205 °C/400 °F for 2 hours).

In the second dryer intervention example in Table 2, the removal, cleaning and insertion of the milk spray nozzles could occur a number of times between CIP events of the dryer, such that any microorganisms entering the dryer during these potential cross-contamination events would not be subjected to a process control step. In this example, it is possible to do a risk assessment on the cross-contamination event (particularly if the cross-contamination event results in a high-risk score) or individual vectors related to the event to determine which vectors are most important to control. In this case, entry of air has been identified as a vector and the risk assessment for the air indicates that other vectors associated with the cross-contamination event may be more important.

Higher risk scores for sources may be used to help prioritize resource allocation, identify where additional control measures are needed and/or to justify capital expenditure. Likewise, higher risk scores for vectors may help prioritize actions taken to reduce the frequency of the vector, identify where additional control measures can lower the risk associated with the vector and/or to eliminate the vector altogether.

The risk analyses as described in Tables 1 and 2, for sources and cross-contamination event vectors, respectively, can further be developed by considering the risk scores for the sources and vectors without controls. For the maximum risk scores associated with the meat slicer foot plates (Table 1) or the removal, cleaning and reinstallation of the spray nozzles (Table 2), these scores indicate that if these sources or cross-contamination events were uncontrolled, or more practically, if the required controls failed, there could be a significant risk of pathogens being present in the processing environment (meat slicer foot plate) or product (spray nozzles).

OPRPs as a Control for Product Vectors
Since the control of significant product vectors is especially critical to product safety, these controls could be described as OPRPs. An OPRP requires the establishment of operating limits (or control limits), monitoring activities, corrective actions for when a control limit is not met, verification activities and record-keeping procedures.

Table 3 describes the controls associated with the theoretical milk spraying nozzle removal and reinsertion procedure described in Table 2. The hazard is that Salmonella could be taken into the dryer on the nozzle and supporting wand, and via the air surrounding the top of the dryer. The nozzle and wand could be cross-contaminated from the operative’s hands and clothing and from the tools used. Control measures could include changing into clean clothing at the point of nozzle removal and reinsertion, using dedicated tools and cleaning equipment, decontaminating wands and nozzles and all surfaces touched prior to reinsertion and tamperproof tagging of the wands so that they cannot be unintentionally removed. By microbiologically filtering environmental air surrounding the dryer, contamination from the air at routine dryer interventions would be controlled.

ATP testing prior to entry could apply an operating limit to an assessment of the cleanliness of the wands, nozzles and tools, and verification of cleanliness could be periodically undertaken by microbiological sampling. During the nozzle removal procedure, observations could be made to ensure the procedure was being done correctly and that there were no extrinsic factors that could act as additional cross-contamination vectors. Records would be kept of all interventions into the dryer, whether removal and reinstallation procedures had been correctly followed, ATP and microbiological counts and tamperproof tag numbers. Corrective actions would review the training of the staff against removal and reinstallation procedures and the effectiveness and validation of the tools and cleaning equipment decontamination programs.

In the same manner as CCP records, the records of an OPRP should be incorporated into a food safety plan to ensure all essential conditions were met during the manufacture of a product. If a deviation in an OPRP were to occur, then the affected product should be placed on hold while a cross-functional team is assembled to review the associated risk and make a decision on product disposition.

The management strategy described above for OPRPs is essentially the same as for CCPs as defined under HACCP. So then what is the difference between a CCP and an OPRP? CCPs are generally described for specific steps in the manufacturing process to eliminate or reduce a significant hazard to an acceptable level, for example, cooking of a meat patty, cooling of a sauce or running a liquid product through a screen of defined particle size. On the other hand, OPRPs are generally described for procedures or programs that address some aspect of the processing environment or the interaction of the processing environment with the process, for example, the manual removal, cleaning and reinsertion of milk spray nozzles into the spray dryer during a production run.

A Developing Study
The concept of identifying sources and vectors of cross-contamination, assessing their risk and managing their risk through OPRPs in a fashion similar as CCPs is a developing study.

Elevating the control of sources and vectors to the level of OPRPs and managing them similar to CCPs focuses attention on the control of what is thought to be the highest risk of cross-contamination from the processing environment to the product. Controlling sources and vectors by developing and documenting OPRPs as discussed in this article may provide a means to demonstrate increased confidence in product safety should a pathogen be found in the manufacturing environment.

Taken beyond microbiological hazards, the same source-and-vector approach may be used to evaluate and control nonmicrobiological hazards, such as allergens or foreign material. As these are developing concepts, comments are welcomed as to how they can be improved.

John Holah, Ph.D., is an applied microbiologist working both in food factories and the laboratory and is responsible for food hygiene at Campden BRI.

Edyta Margas is a hygiene and novel technologies specialist at Campden BRI.

Robert Hagberg is the director of QA technical services at Land O’Lakes, Inc.

Benjamin Warren, Ph.D., is the director of product safety & regulatory affairs at Land O’Lakes, Inc.

Judy Fraser-Heaps is the senior manager of QA, microbiology & food safety at Land O’Lakes, Inc.

Sara Mortimore is the vice president of quality & regulatory affairs at Land O’Lakes, Inc.

References
1. www.cdc.gov/print.do?url=http%3A//www.cdc.gov/salmonella/agona.
2. Jackson, K.A., M. Biggerstaff, M. Tobin-D’Angelo, D. Sweat, R. Klos, J. Nosari, O. Garrison, E. Boothe, L. Saathoff-Huber, L. Hainstock and R.P. Fagan. 2011. Multistate outbreak of Listeria monocytogenes associated with Mexican-style cheese made from pasteurized milk among pregnant, Hispanic women. J Food Prot 74: 949–953.
3. Gaze, R. 2009. HACCP: A practical guide (fourth edition). Campden BRI, Chipping Campden, Gloucestershire, UK.
4. www.codexalimentarius.net.
5. www.mygfsi.com.
6. Middleton, K.E. and J.T. Holah. 2008. Cleaning and disinfection of food factories: A practical guide. Campden BRI, Chipping Campden, Gloucestershire, UK.
7. George, R.M. 2004. Foreign bodies in foods: Guidelines for their prevention, control and detection (second edition). Campden BRI, Chipping Campden, Gloucestershire, UK.
8. Smith, D. 2009. Hand hygiene: Guidelines for best practice. Campden BRI, Chipping Campden, Gloucestershire, UK.
9. www.bcas.org.uk.
10. www.cieh.org.
11. Anon. 2005. ISO 22000:2005 Food Safety management systems — Requirements for any organization in the food chain.
12. Smith, D. 2007. Ranking of cross-contamination vectors of ready-to-eat foods: A practical approach. Campden BRI, Chipping Campden, Gloucestershire, UK.
13. Lamb, H. 1994. Hydrodynamics (6th edition). Cambridge University Press.