U. S. Food and Drug Administration
Center for Food Safety and Applied Nutrition
March 29, 2001

1. Introduction

On December 18, 1997, the U.S. Food and Drug Administration (FDA) implemented a dramatic change in the manner with which domestically produced and imported fishery products are regulated. The new regulations mandated the application of the Hazard Analysis and Critical Control Point (HACCP) principles to the processing of seafood. Many compelling motivations are driving the use of HACCP, but four of the most prominent driving forces are that HACCP (1) is focused on food safety, (2) is science-based, (3) relies on preventive controls rather than retrospective end-product testing, and (4) focuses control on those food safety hazards that are reasonably likely to occur. In essence, HACCP requires food processors to understand the safety hazards associated with the food, the process, as well as distribution and marketing conditions, and to use appropriate controls so that any identified hazard(s) is prevented, eliminated, or reduced to acceptable levels. Although HACCP holds great promise for minimizing the risk of foodborne disease, application of HACCP principles to a few foods and food processes is challenging because no useful strategies are available to control some identified food safety hazards. Cold-smoked fish and the processes generally used for this product are examples of foods and processes that pose such challenges. Although the cold-smoking process may be applied to seafood commodities other than finfish, this report is limited to cold-smoked finfish. When applicable, the report refers to hot-smoked fish, but an in-depth analysis of hot-smoked fisha completely different product from cold-smoked fishis outside the scope of this report.

2. Cold-smoked fish

The definitions associated with cold-smoked fish are vague. For example, both the Codex Alimentarius Commission (1979) and the Association of Food and Drug Officials (AFDO 1991) use the following definition: Cold process smoked fish means a smoked fish that has been produced by subjecting it to smoke at a temperature where the product undergoes only incomplete heat coagulation of protein.

Given the non-specific nature of the Codex and AFDO definition, one would expect to find a wide variety of processing parameters being used by industry. In fact, processors do use different parameters, which in turn impart specific sensory attributes to their products. In certain markets there are distinct preferences for cold-smoked fish with specific sensory characteristics (that is, salty/dry or low salt/moist).

Gram and Huss (2000) categorized cold-smoked fish as "lightly preserved." In Europe, this group includes fish products preserved by low levels of salt (< 6% NaCl in the water phase) and, for some products, the addition of preservatives (sorbate, benzoate, NO₂, or smoke). The pH of the products is high (> 5.0) and they are often packaged under vacuum and must be stored and distributed at refrigeration ( 5 ^oC, 41 ^oF) or frozen temperatures. These products are typically consumed as ready-to-eat with no heat treatment. The authors further reported that drying times, after salting, ranged from 1 - 6 at 20 - 28 ^oC (68 - 82 ^oF), and smoking parameters had a maximum of 30 ^oC (86 ^oF) for 3 - 6 h, and shelf life (based on sensory evaluation) ranging from 3 to 6 wk at 5 ^oC (41 ^oF).

According to current U.S. HACCP regulations, a suggested critical limit for air-packaged product is at least 2.5% NaCl (water phase in the loin muscle), for vacuum-packaged or modified atmosphere-packaged product at least 3.5% NaCl (water phase in the loin muscle), or a combination of 3.0% water phase salt (WPS) and at least 100, but not more than 200 ppm, of sodium nitrite (allowed in the United States for sable, salmon, shad, and chub). The application of smoke, which can be "normal" (wood generated), liquid, or a combination of both, is typically done in the following time and smoke chamber temperature combinations: a) not to exceed 90 ^oF (32 ^oC) for more than 20 h, b) not to exceed 50 ^oF (10 ^oC) for more than 24 h, or c) not to exceed 120 ^oF (48 ^oC) for more than 6 h (for cold-smoked sablefish). All products must be maintained at 37 - 38 ^oF (3.0 - 3.3 0C) at all times, if no other controls are present (that is, an adequate NaCl concentration and the application of smoke). If adequate concentrations of NaCl and smoke are present, the product must be maintained at 40 ^oF (4.4 ^oC). These formulation and process scenarios are based on the need to inhibit the germination of Clostridium botulinum spore germination, growth, and toxin formation. The interplay of the inhibitory effects of salt, temperature, smoke, and nitrite is complex. Control of the brining or dry salting process is significant to ensure that there is sufficient salt in the finished product; however, preventing C. botulinum type E (and non-proteolytic types B and F) toxin production is made even more complex by the fact that adequate salt levels are often not achieved during brining. Therefore, proper drying is also important to achieve the finished product WPS level (the concentration of salt in the water portion of the fish flesh) needed to inhibit the growth and toxin formation of C. botulinum.

Interestingly, current FDA guidance (FDA 1998) for cold-smoked fish indicates that it is important that the product not be subjected to so much heat that the number of spoilage organisms is significantly reduced. Spoilage organisms are necessary to compete with the growth and toxin formation of C. botulinum type E and non-proteolytic types B and F. Thus, competitive inhibition may be important in cold-smoked products because the heat applied during the process is insufficient to inactivate or damage the C. botulinum spores and may not even affect the vegetative cell form. Packaging may have a significant influence on the efficacy of competitive inhibition. It is likely that the lactic acid bacteria, which will dominate in a non-oxygen storage environment, may in some circumstances inhibit C. botulinum. In an oxygen storage environment, however, a mix of surviving aerobic and facultative anaerobic spoilage organisms could create anaerobiosis, which would facilitate growth and toxin formation by C. botulinum. Relying on the competitive flora to restrict growth of C. botulinum or to indicate spoilage, however, is not an effective or reproducible control point and cannot be trusted to control safety.

The following flowchart is a generalized description of the cold smoking process, and as such, it does not account for other process variations encountered in actual industry practice. Guidelines from the AFDO (AFDO 1991) have been indicated where appropriate.

The microbial flora associated with freshly harvested fish is principally a function of the environment in which the fish are caught, not the fish species (Shewan 1977). Although this generalization appears simple, there is great diversity in aquatic environments (that is, fresh, salt, estuarine, cold, tropical, temperate, coastal, open ocean, polluted, and pristine) and therefore, the indigenous microbial populations of fish can vary significantly.

The microflora on temperate-water fish is predominately psychrotrophic or psychrophilic, gram-negative bacteria belonging to the genera Pseudomonas, Moraxella, Acinetobacter, Alcaligenes, Shewanella, and Flavobacterium. Aeromonas spp. are characteristic of freshwater fish, whereas Vibrio spp. are typical of marine waters. Additionally, several types of C. botulinum are found in the aquatic environment; however, the non-proteolytic C. botulinum, especially strains producing type E toxin (called C. botulinum type E), is truly indigenous, particularly in the temperate and subarctic zones (Huss 1980). While spores of C. botulinum are found in most sediments and on fish from around the world, the limited quantitative data available suggest that the numbers are low, usually less than 100 cfu/g. Higher numbers have been reported from the Great Lakes and the Sound of Scandinavia (Dodds 1993).

Since the bacterial flora on freshly caught fish is a reflection of the environment in which the fish is caught, it is not surprising that aquacultured fish are more likely to be contaminated with certain non-indigenous species. This is due to the closer proximity of fish farms to human and animal populations and the waste generated by each. Research has demonstrated that Listeria monocytogenes is a frequent isolate from surface waters (up to 62% positive samples) and polluted seawater (up to 33% positive samples), while it is not isolated from unpolluted ocean waters and spring water (Huss and others 1995).

The microflora on a fish product is a function of the indigenous flora and the microflora of the processing environment. Typically, the term "processing environment" is limited to that of an actual processing plant. Any handling of fish, and the associated sanitary practices from the point of harvesting, however, have the potential to contribute to the microflora on the final product. Consequently, the presence or absence of foodborne pathogens on a fish product is a function of the harvest environment, sanitary conditions, and practices associated with equipment and personnel in the processing environment, as well as any lethality associated with the actual process.

The most universal means of preserving fish quality is chilling (ice or mechanical refrigeration systems). As the temperature on the surfaces of fish is reduced below optimal, bacterial growth begins to slow. Given the microbial diversity typical of fish, it is not surprising that chill temperatures impact some microbial species more dramatically than others. The growth of some species is totally inhibited while the growth of other species proceeds, albeit more slowly. Consequently, the rather diverse microflora will shift until just a few species predominate, due to the selective pressure of the chilled environment. The two groups that will ultimately become dominant during aerobic ice storage are Pseudomonas spp. and Shewanella putrefaciens (Gram and others 1987; Levin 1968).

The salting and drying/smoking processes often reduce the numbers of microorganisms and cause a change in the spoilage microflora. While autolytic changes are believed to be the cause of some of the textural changes observed in vacuum-packed, cold-smoked fish during chill storage (Truelstrup Hansen and others 1996), microorganisms are responsible for the unpleasant off-odors and flavors that develop. Aerobic storage of cold-smoked fish at refrigerated temperatures results in the development of spoilage microflora consisting mostly of Pseudomonas spp. and yeast. Growth directly on the product may become so pronounced that microbial colonies can actually be observed without magnification. During vacuum- or CO₂-packing, lactic acid bacteria rapidly become the dominant microflora. Typically, their numbers increase from 10² to 10⁷ - 10⁸ cells/g within 2 wk (Truelstrup Hansen and others 1996; Jorgensen and others 2000; Civera and others 1995). Gram-negative bacteria, such as psychrotrophic Enterobacteriaceae or marine vibrios, are often part of the microbial community as well. The spoilage of this product is complex. For example, the nature of the microorganisms that cause spoilage and the manner in which the interactions among the organisms influence the spoilage scenario are not yet fully understood (Gram and Huss 2000). This is important because these spoilage bacteria may be expected to help control pathogens.

The processes used for cold smoking of fish are not exceptionally rigorous; thus, there is concern that some foodborne pathogens, if present, could survive. Organisms of primary concern are C. botulinum (psychrotrophic non-proteolytic type B, E, and F), and L. monocytogenes. In addition to contributing to pathogen survival, the extensive handling of products following the cold smoking process provides ample opportunities for other foodborne pathogens (that is, Salmonella spp., Shigella spp., Escherichia coli, Staphylococcus aureus, Bacillus cereus, Vibrio parahaemolyticus, and Vibrio cholerae) to contaminate and survive in the products if insufficient attention is given to Good Manufacturing Practices (GMPs), Sanitation Standard Operating Procedures (SSOPs), and hygienic practices of plant employees. While noting that negligence in sanitary processing and employee hygienic practices could nullify any HACCP control strategies, there is little history of classic pathogens such as Salmonella sp., Shigella sp., and S. aureus occurring in cold-smoked fish products. The reasons for this vary but may include the fact that these pathogens are poor competitors (for example, S. aureus), or that they may be injured or reduced by the cold-smoking process or the high salt concentration of the product. Therefore, in this document, these organisms are not included in the discussion of potential health hazards that are reasonably likely to occur.

The Food and Agriculture Organization (FAO 1999) reported that L. monocytogenes and other Listeria species have been isolated from fishery products on a regular basis since the late 1980s. Listeria monocytogenes can survive the cold-smoking process and is capable of growing at the temperature-NaCl combinations of the final product. Studies of inoculated vaccum-packed cold-smoked fish have shown that the organism may grow from 10³ cfu / g to 10⁷-10⁸ cfu / g in 2 - 4 wk. However, the growth in naturally contaminated products is significantly slower and levels above 10⁴ cfu / g are rarely detected, even at end of shelf life. Interestingly, given the relatively high incidence in ready-to-eat and heat-treated fishery products, there have not been any large outbreaks of listeriosis due to the consumption of contaminated fishery products. A couple of sporadic cases, however, have been linked to lightly preserved fish products such as smoked mussels and cold-smoked trout (Brett and others 1998; Miettinen and others 1999). The FAO report did suggest, however, that an increase in the incidence of listeriosis over the next decades acquired from all food products is likely due to the increasing numbers of susceptible people. Highly susceptible groups include pregnant women, infants, the elderly, and immunocompromised people. Although listeriosis occurs infrequently, at an annual rate of 2 to 10 per million, the fatality rate usually ranges from 20 - 30% in the highly susceptible groups (Farber and Peterkin 2000). In an effort to understand the extent of the L. monocytogenes food safety problem, the United States Department of Health and Human Services' FDA's Center for Food Safety and Applied Nutrition in collaboration with the U.S. Department of Agriculture's Food Safety and Inspection Service and the Centers for Disease Control and Prevention conducted a risk assessment on listeriosis associated with various foods, including smoked fish. The risk assessment presents an estimate of the level of exposure of consumers to L. monocytogenes and its relationship to public health. Because the period of request for public comments is still open at the time of this writing, after which the draft assessment may be subject to revision, we did not include any material from this risk assessment document. Once the risk assessment is finalized, we anticipate that the results from it will become invaluable information in support of future scientific evaluation of public health hazards in ready-to-eat foods, such as cold-smoked fish. The draft risk assessment is available at http://www.foodsafety.gov/~dms/lmrisk.html.

With respect to C. botulinum, the concerns with psychrotrophic non-proteolytic C. botulinum are not associated with the mere presence of the organism or its spores. The organism is found both in freshwater and saltwater species of fish; hence, its prevalence is widespread but its incidence is low. Packaging environment and temperature significantly influence risk factors associated with C. botulinum. Since C. botulinum is an anaerobe, packaging that eliminates or reduces the oxygen concentration enhances the opportunity for germination, growth, and toxin production of the organism. This is especially true when competing organisms, such as the aerobic spoilage microflora, are suppressed due to the anaerobic environment. Because refrigeration temperatures alone will inhibit the growth of proteolytic C. botulinum and of non-proteolytic C. botulinum, appropriate temperature control could be an important control measure. Control could be established by maintaining temperatures below 3.0 - 3.3 ^oC (37 - 38 ^oF) throughout distribution, retail storage, and by the consumer to inhibit growth of all non-proteolytic and proteolytic strains. Maintaining temperatures consistently below 3.0 - 3.3 ^oC, however, is not a realistic expectation, based on current distribution, warehousing, retailing, and consumer handling practices. Consequently, a combination of both low temperature control and salt are vital. Because there are no reports in the scientific literature linking cold-smoked fish to an outbreak of botulism, it is speculated that the combination of NaCl and low temperature has been sufficient for control of the hazard.

The food safety concerns associated with cold-smoked fish are not limited to microbiological hazards. Consideration must also be given to the likelihood of another biological hazard--parasites--surviving the processes employed in cold smoking. As stated earlier, the processes employed in the cold smoking of fish are not rigorous; thus, parasite survival is a distinct possibility. Therefore, species carrying parasites that are known to be pathogenic to humans must be frozen at some stage during processing. Current FDA guidance indicates that freezing at the following temperatures will kill parasites of concern: 4 ^oF (-20 ^oC), measured internally or externally, for 7 d, or - 31 ^oF (- 35 ^oC), measured internally, for 15 h (FDA 1998). European legislation requires that raw fish used in the production of matjes-herring and cold-smoked fish (wild salmon, herring, mackerel, sprat, cod, and halibut) be frozen at least 24 h at - 20 ^oC (EEC 1991). In the United States and Europe, farmed salmon reared on pelletized feed are not subject to the freezing requirement because the feed is considered void of parasites due to the feed processing method.

Within the context of HACCP, the food safety hazard associated with high levels of histamine is classified as a chemical hazard. Biogenic amines, like histamine, are by-products of bacterial growth on the surfaces of susceptible species of harvested fish. Some surface bacteria excrete enzymes capable of decarboxylating amino acids, particularly histidine to produce histamine. Once formed, biogenic amines are quite stable and are not destroyed or eliminated by any of the steps associated with the processing of cold-smoked fish. The most significant factor associated with formation of biogenic amines is temperature. If freshly caught fish are cooled rapidly and maintained at cold temperatures, levels remain low. If, however, the fish are temperature-abused prior to cooling, levels of biogenic amines can rise, even under subsequent refrigerated conditions, due to the activity of the preformed decarboxylase enzymes released by the bacteria during the period of temperature abuse. Histamine may be formed in lightly preserved fish products. Levels between 3 and 240 ppm have been detected in cold-smoked salmon (Jørgensen and others 2000). The high levels are above the FDA histamine guidance level of 50 ppm (FDA 1998) and of the 100 ppm by the European regulation for Scombroidae and Clupeidae (EEC 1991). Also, the high levels exceed the European maximum limit of 200 ppm. Lightly preserved fish products, such as cold-smoked fish, however, have not been linked epidemiologically to foodborne disease caused by scombroid toxicity (Gram and Huss 2000).

The HACCP concept is based on a simple yet fundamental premise of "identify and control." Specifically, HACCP requires that all food safety hazards associated with the food and with the processes used in manufacturing the food, that are reasonably likely to cause illness or injury in the absence of control, must be identified. Once all the food safety hazards are identified, a control procedure must be established to prevent, eliminate, or reduce the hazards to acceptable levels. If a control option cannot be identified in the existing process, HACCP ideology dictates that the process must be modified to create a control opportunity. With respect to cold-smoked fish, hazard analysis suggests that biological hazards may exist (that is, L. monocytogenes, C. botulinum), but a definitive control point is either problematic (that is, temperature control for C. botulinum) or non-existent (that is, kill step for L. monocytogenes). Suggestions that the process be modified, such as adding additional salt or a terminal heating step, are generally not welcome by producers because such changes would significantly alter the sensory attributes of the final product and result in the loss of customers.

The concerns associated with C. botulinum in smoked fish are not new. Since the early 1960s, FDA has grappled with the issue of this organism in fishery products; specifically, those products packaged in vacuum or modified atmospheres. Prior to the implementation of the HACCP regulation, FDA discouraged the use of these forms of packaging on both fresh and processed products. In fact, the FDA's Food Code (FDA 1997) and AFDO's Retail Guidelines for Refrigerated Foods in Reduced Oxygen Packages (AFDO 1990) specifically prohibit these forms of packaging at the retail level, unless the products are frozen before, during, and after packaging. Nonetheless, at the processing level vacuum packaging is acceptable if barriers (that is, 3.5% NaCl, smoke, and chill temperature control) are in place. Since the advent of the HACCP rule, however, there has been a dramatic shift in responsibility. Under HACCP, FDA is requiring smoked fish processors to document that the C. botulinum hazard is being controlled. The Fish & Fisheries Products Hazards and Controls Guide (FDA 1998) indicates that control can be accomplished through a combination of temperature, salt, packaging, and, where appropriate, preservatives (nitrite). According to Gram and Huss (2000), less salt is needed to inhibit growth of the psychrotrophic (non-proteolytic) C. botulinum types B, E, and F at chilled fish temperatures than at higher temperatures. Moreover, reduced pH in combination with salt enhances the inhibition of the organism.

While C. botulinum has been of concern for many years, L. monocytogenes is a relatively new concern. Listeria monocytogenes has been isolated from fishery products on a regular basis (FAO 1999) although no clear contamination route is known (Eklund and others 1995). Additionally, it survives both the salting and the cold smoking processes and is capable of growth at refrigeration temperatures (Hudson and Mott 1993). The prevalence of L. monocytogenes in cold-smoked fish is highly variable. Jørgensen and Huss (1998) found the organism to range from < 1.4% to 100% in cold-smoked salmon from three production sites. In the context of HACCP control, Huss and others (1995) indicated that since the contamination source is not known and the preservation steps do not prevent growth, no definitive critical control point can be identified. The problem is further complicated by FDA's current policy of "zero-tolerance" (non-detectable on samples by the current methods [AOAC 1995] in 25 g sample) for L. monocytogenes on ready-to-eat products. This policy is a continuation of the agency's initial response to the public health concerns that ensued when L. monocytogenes was first implicated as the causative agent in a foodborne disease outbreak in the mid 1980s. At that time, little was known about the organism, its pathogenicity, or infectious dose. Although much has been learned, the dose/response relationship of the organism for humans is not yet known. Based on the reported numbers of Listeria in contaminated foods responsible for epidemic and sporadic foodborne cases, however, there is little evidence that a very low number of L. monocytogenes in foods causes disease (FAO 1999). In fact, the FAO (1999) reported that data indicates that a per capita human exposure to doses of L. monocytogenes exceeding 1,000 cfu (total ingested dose) is likely to occur several times each year. Despite this exposure, however, the total incidence of invasive listeriosis is estimated to be somewhere between 2 to 10 cases per million population per annum in countries where data are available. Ross and others (2000), using a quantitative risk assessment approach, concluded that unless the "zero tolerance" was rigorously enforced, this standard was not better for food safety compared to a level of 100 per gram at time of consumption.

The United States is not alone in its "zero-tolerance" criteria for ready-to-eat products. Austria, Australia, New Zealand, and Italy currently require the absence of L. monocytogenes in a 25 g sample. Germany, Netherlands, and France have a tolerance of less than 100 cfu/g at the point of consumption, while Canada and Denmark have a tolerance of less than 100 cfu/g for some foods and zero tolerance for other foods (that is, those with extended shelf-life that can support the growth of the organism).

Clearly, the "zero-tolerance" policy (non-detectable, by the current methods [AOAC 1995] in 25 g sample) for L. monocytogenes in ready-to-eat products is a significant issue for the cold-smoked fish industry. Scientific evidence indicates that finding L. monocytogenes in cold-smoked fish is merely a question of "when" not "if." Some studies have demonstrated significant plant-to-plant variation in detecting positive samples; however, plants that may have been negative on one sampling date have produced positive findings on another date (Jørgensen and Huss 1998; Fonnesbech Vogel, Ojeniyi and others 2001).

As previously noted, cold-smoked fish presents a dilemma. As a food, there is little question that cold-smoked fish has been a part of our dietary heritage for centuries. Only recently, however, have serious questions of safety been raised. Is the product safe? If yes, why? If no, why not? Can the process be modified to establish a greater margin of safety? Will the modified product be acceptable? For a relatively minor seafood commodity, cold-smoked fish introduces some rather significant challenges.

To address the safety issues with cold-smoked fish, this report is structured in sections. The first sections discuss the potential hazards: L. monocytogenes, C. botulinum toxin, biogenic amines, and parasites. The last section addresses the hazards, control points, and processing parameters for each step of the process, from harvesting to consumption.

[AFDO] Association of Food and Drug Officials. 1990. Retail guidelines for refrigerated food in reduced oxygen packages. J Assoc Food Drug Of 54(5):80-4.

[AFDO] Association of Food and Drug Officials. 1991 June. Cured, salted, and smoked fish establishments good manufacturing practices [model code]. [York (PA)]: Association of Food and Drug Officials. 7 p.

[AOAC] Association of Official Analytical Chemists. 1995. FDA Bacteriological Analytical Manual (BAM), 8th ed. Gaithersburg (MD): AOAC Int.

Brett MSY, Short P, McLauchlin J. 1998. A small outbreak of listeriosis associated with smoked mussels. Int J Food Microbiol 43:223-9.

[CAC] Codex Alimentarius Commission. 1979. Recommended International Code of Practice for Smoked Fish. Rome: Codex Alimentarius Commission. CAC/RCP 25-1979.

Civera T, Parisi E, Amerio GP, Giaccone V. 1995. Shelf-life of vacuum-packed smoked salmon: microbiological and chemical changes during storage. Arch Lebensmittelhyg 46:13-7.

Dodds KL. 1993. Clostridium botulinum in the environment. In: Hauschild AHW, Dodds KL, editors. Clostridium botulinum: ecology and control in foods. New York: M Dekker. p 21-52.

EEC. 1991. Council directive 91/493/EEC of 22nd July 1991 laying down the health conditions for the production and the placing on the market of fishery products. Off J Eur Comm(No. L268):15-32.

Eklund MW, Poysky FT, Paranjpye RN, Lashbrook LC, Peterson ME, Pelroy GA. 1995. Incidence and sources of Listeria monocytogenes in cold-smoked fishery products and processing plants. J Food Prot 58(5):502-8.

[FAO] Food and Agriculture Organization. 1999 May. Report of the FAO expert consultation on the trade impact of Listeria in fish products. Rome: FAO. FAO Fisheries Report nr 604. 34 p.

Farber JM, Peterkin PI. 2000. Listeria monocytogenes. In: Lund BM, Baird-Parker TC, Gould GW, editors. The microbiological safety and quality of foods. Gaithersburg (MD): Aspen. p 1178-1232.

[FDA] Food and Drug Administration. 1998. Fish & Fisheries Products Hazards & Controls Guide. 2nd ed. Washington, D.C.: FDA, Office of Seafood. 276 p.

[FDA] Food and Drug Administration. 1999. Food Code. Washington, DC: U.S. Department of Health and Human Services, Public Health Service, Food and Drug Administration.

Fonnesbech Vogel B, Ojeniyi B, Ahrens P, Due Skov L, Huss HH, Gram L. 2001. Elucidation of Listeria monocytogenes contamination routes in cold-smoked salmon processing plants detected by DNA-based typing methods. Appl Environ Microbiol. Forthcoming.

Gram L, Huss HH. 2000. Fresh and processed fish and shellfish. In: Lund BM, Baird-Parker TC, Gould GW, editors. The microbiological safety and quality of food. Gaithersburg (MD): Aspen. p 472-506.

Gram L, Trolle G, Huss H. 1987. Detection of specific spoilage bacteria from fish stored at low (0° C) and high (20° C) temperature. Int J Food Microbiol 4:65-72.

Hudson JA, Mott SJ. 1993. Growth of Listeria monocytogenes, Aeromonas hydrophila and Yersinia enterocolitica on cold-smoked salmon under refrigeration and mild temperature abuse. Food Microbiol 10:61-8.

Huss HH. 1980. Distribution of Clostridium botulinum. Appl Environ Microbiol 39:764-9.

Huss HH, Ben Embarek PK, From Jeppesen V. 1995. Control of biological hazards in cold smoked salmon production. Food Control 6(6):335-40.

Jorgensen LV, Dalgaard P, Huss HH. 2000. Multiple compound quality index for cold-smoked salmon ( Salmo salar ) developed by multivariate regression of biogenic amines and pH. J Agric Food Chem 48:2448-53.

Jorgensen LV, Huss HH. 1998. Prevalence and growth of Listeria monocytogenes in naturally contaminated seafood. Int J Food Microbiol 42:127-31.

Levin RE. 1968. Detection and incidence of specific species of spoilage bacteria on fish. 1. Methodology. Appl Microbiol 16:1734-7.

Miettinen MK, Siitonen A, Heiskanen P, Haajanen H, Bjorkroth KJ, Korkeala HJ. 1999. Molecular epidemiology of an outbreak of febrile gastroenteritis caused by Listeria monocytogenes in cold-smoked rainbow trout. J Clin Microbiol 37(7):2358-60.

Ross T, Todd E, Smith M. 2000. Exposure assessment of Listeria monocytogenes in ready-to-eat foods: preliminary report for joint FAO/WHO expert consultation on risk assessment of microbiological hazards in foods. Rome: Food and Agriculture Organization of the United Nations. 242 p.

Shewan JM. 1977. The bacteriology of fresh and spoiling fish and the biochemical changes induced by bacterial action. Proceedings of the conference on handling, processing, and marketing of tropical fish. London: Tropical Products Institute. p 51-60.

Truelstrup Hansen L, Gill T, Drewes Rontved S, Huss HH. 1996. Importance of autolysis and microbiological activity on quality of cold smoked salmon. Food Res Int 29:181-8.

Hypertext updated by ear 2001-MAY-18