Processing Parameters Needed to Control Pathogens in Cold Smoked Fish Chapter III. Potential Hazards in Cold-Smoked Fish: Clostridium botulinum type E

(Table of Contents


Clostridium botulinum type E is an indigenous organism in the aquatic environment and is the type mainly associated with botulism from seafood products. This section outlines what is currently known about the organism and its behavior in cold-smoked fish. Major findings of importance to the safety of cold-smoked fish and the control of C. botulinum type E are included. While the section provides some of the data relevant for a science-based risk assessment, it does not give complete information on all four components: hazard identification, hazard characterization, exposure assessment, and risk characterization. Recent book chapters about the organism in Lund and Peck (2000) and Dodds and Austin (1997) cover these issues in greater depth. This section ends with lists of conclusions and research needs.

1. Introduction 

Clostridium botulinum comprises a range of gram-positive, anaerobic, spore-forming bacteria that produce botulinum neurotoxin. The bacteria are vastly different on a phylogenetic scale and only similar in toxins and disease pattern. The toxins (as well as the microorganism) have been divided into types A, B, C, D, E, F, and G depending on their antigenic properties. Only the confusion that would arise upon renaming them has prevented a redistribution of the different types into different bacterial species (Lund and Peck 2000; Dodds and Austin 1997).

Botulism may be caused by four different scenarios: a) foodborne botulism which results from ingestion of toxin preformed in foods, b) infant botulism in which spores are ingested, germinate, and lead to toxin formation in the intestine, c) wound botulism and d) adult botulism caused by intestinal germination and growth. In general, botulism occurs today only rarely; indeed, the latter scenario is extremely rare. The botulinum neurotoxins, however, are the most potent toxins known. Before anti-toxins for treatment became available, mortality was as high as 75%. Due to the severity of the disease, research has been directed to understanding the ecology, toxin production, and pathogenicity of the different types and strains. Much has been learned about the ecology of these organisms and about prevention of growth and toxin-production in foods. Two of the C. botulinum groups are responsible for foodborne botulism. Group I C. botulinum is proteolytic and mesophlic, and strains produce toxins of types A, B or F. Group II C. botulinum is non-proteolytic and psychrotrophic, and strains produce toxins of types B, E or F. Group II C. botulinum,, and in particular strains that produce neurotoxin of type E, is a major concern for cold-smoked fish. In this document, these strains will be referred to as C. botulinum type E. One reason for the emphasis on strains that produce type E toxin is that these strains appear to be truly aquatic. It is important to recognize that controls for C. botulinum type E will also be valid for strains of Group II C. botulinum that produce toxins of types B or F.

Many research groups focused their activities on C. botulinum type E following a range of outbreaks of botulism from hot-smoked freshwater fish in the 1960s and 1970s. The more general findings described below are cited from chapters by Lund and Peck (2000), Dodds and Austin (1997), Eklund (1992), and Huss (1981).

Clostridium botulinum type E is a gram-positive, anaerobic, spore-forming foodborne pathogen. As mentioned, it is part of the psychrotrophic, non-proteolytic group of C. botulinum, which also comprises C. botulinum types B and F. Disease is caused by a neurotoxin produced in the food product. Spores are widely distributed in the aquatic environment and occur naturally, albeit at low levels, in many raw aquatic foods (Lund and Peck 2000). The organism is strictly anaerobic and sensitive to oxygen. Clostridium botulinum type E can grow and produce toxin at temperatures of 3.0 - 3.3 °C (37-38 °F) or in up to 5% NaCl (water phase salt, WPS). Toxin production at these limits, however, requires that all other growth conditions are optimal. Thus, for example, toxin production is inhibited at lower salt concentrations than 5% if the temperature is lower than optimum. The tolerance to lowering of water activity depends on the solute; thus, the aw minimum is 0.97 when NaCl is used and 0.94 when glycerol is used. The organism is acid sensitive and does not grow below pH 5.0. In conclusion, as a consequence of its capabilities to grow at low temperature and its resistance to salt, it may, if anoxic conditions are created, germinate and produce toxins in many food products with extended chilled shelf lives and cause foodborne botulism.

The toxins produced by all of the different C. botulinum types is heat labile. The heat resistance of C. botulinum type E toxin depends on the pH; the toxin is destroyed by moderate heating at neutral pH and is more resistant at lower pH values (pH 4.0 - 5.0). Thus, the toxin was destroyed after 5 min at 60 °C (140 °F) in a cooked meat medium (pH 7.5) (cited from Huss 1981) but at 62 °C (144 °F) and 65 °C (149 °F) in meat broth (pH 6.2) (Abrahamsson and others 1965). Woodburn and others (1979) investigated the heat inactivation of several botulinum toxins and found that in canned corn at pH 6.2, a 3D reduction occurred in 2 min at 74 °C (165 °F). In phosphate buffer, a similar reduction occurred in 1 min at pH 6.8 but took 6 min if 1% gelatin was added to the buffer (Woodburn and others 1979). In general, heat treatments used for food preparation are sufficient to destroy the toxin; however, it would not be considered acceptable to rely on cooking to inactivate any C. botulinum toxin that may be present in food. Type E toxin is resistant to NaCl; it is stable for weeks in 26% NaCl (Huss 1981).

Clostridium botulinum type E can be isolated from water, aquatic sediments, and aquatic organisms. In general, numbers of the microorganism in fish are low, ranging from 1-2 spores to a few hundred per kg. Some studies have found higher levels, such as 2,000-3,000 spores / kg (Lund and Peck 2000). Because the disease caused by C. botulinum type E is an intoxication, low numbers of spores are not considered a hazard per se, but products allowing germination, growth, and toxin production must be evaluated carefully to determine the combinations of parameters that can either eliminate the organism, for example by heat treatment, or guarantee that growth and toxin production do not occur within the shelf life of the product.

The risk of toxin formation of C. botulinum in cold-smoked fish must be taken seriously, because the processing of cold-smoked fish does not involve a heating step that eliminates the spores (or any vegetative cells). Cold-smoked fish is often vacuum-packed. Low levels of salt and low temperatures are used only as a means of preservation. Numerous studies, models and predictions (see 4.3 and 4.4 below), however, have documented that appropriate combinations of NaCl (water activity) and low temperature are sufficient to guarantee that no growth of the organism occurs within the shelf life of the product under vacuum. Considering the amount of the scientific evidence, the combination of salt and refrigeration can be considered critical parameters to control C. botulinum growth and toxin production when stored under vacuum.

Lightly preserved fish products, particularly fermented fish, have been linked to cases of botulism (Lund and Peck 2000). In most cases, the cause of the disease has been linked to home processing or faulty processing by small producers. Hot-smoked fish has led to several outbreaks. As outlined by Eklund (1992), however, the products were typically underprocessed, often with NaCl of less than 1%; vacuum-packed; and grossly temperature-abused during distribution. In contrast, the panel is unaware of outbreaks of botulism involving cold-smoked fish.

2. Prevalence in water, raw fish, and smoked fish 

Clostridium botulinum type E is a truly indigenous organism of the aquatic environment (Table III-1 and Dodds and Austin 1997). Spores are detected at different levels in sediments. In some areas of the world, such as Great Britain and Ireland, C. botulinum type B is more commonly isolated. It has been shown in studies on the Pacific Coast of the United States that C. botulinum type E is the most frequently isolated type off the coast of Alaska, Washington and Oregon. Off the coast of South Carolina, however, non-proteolytic C. botulinum types B and F were isolated more frequently than C. botulinum type E (Eklund 1992). Due to the widespread occurrence in aquatic environments, C. botulinum type E can also be isolated from fresh fish (Table III-2). Levels are typically low, although high levels have been detected in one sample of Danish fish (Huss and others 1974). In this case, fish were sampled from trout reared in mud ponds during the time when wet feed was allowed. Both of these factors are believed to have increased numbers of C. botulinum

Table III-1. Prevalence of Clostridium botulinum type E in freshwater and seawater sediments (modified from Dodds 1993, Huss 1981, Huss 1980, Sugiyama and others 1970, Eklund and Poysky 1967)

Sampling location Sample size, g % positive MPN2 / kg
  Green Bay (USA) 1 77 100,000-36,000,000
  Rivers leading into Green Bay (USA) ? 5-50 10,000
  Not specified (Denmark) 10 86 20-800
  Lake Washington (USA) 5 91 18,400-35,000
  Coastal area (Scandinavia) 6 100 More than 780
  Caspian Sea (Iran) 2 171 93
  Pacific coast (USA) 5 53 ?
  Coast (Canada) 10 0.6 40-280
  Bellingham Bay (USA) 5 50-93 540-32,000
  Not specified (Denmark) 10 92 800-350, 000

1 a few strains of type B isolated

2 MPN, most probable number

Table III-2. Prevalence of Clostridium botulinum Type E in fish (modified from Dodds and Austin 1997, Huss 1981, Hyytiä and others 1998, Huss and others 1974, Nickerson and others 1967, Cann and others 1966)

Fish type Sample size, g % positive MPN2 / kg
Whitefish 10 12 (also C) 14
Flounder, vacuum-packed, frozen 1.5 10 70
Rockfish, dressed 10 100 2400
Cod, whiting, flounder ? 0.40 -
Salmon, smoked 20 2 (type B) < 1
Carp, salted 2 63 490
Cod, haddock a.o. 10 intestines 4.5 40-100
Trout, farmed whole fish 5-100 340-5300
Trout, herring 80-234 0 -
Herring whole fish 45-65 5 - 60
Rainbow trout, cold-smoked 100-200 31 40-290
Rainbow trout, fresh 100-200 15-201 30-1900

1 Determined by PCR-detection of the E toxin gene

2 MPN, maximum probable number

Only a limited number of studies have determined real-life prevalence of C. botulinum in cold-smoked fish. Heinitz and Johnson (1998) sampled from 201 vacuum-packaged smoked fish and shellfish products and did not detect spores of C. botulinum in any package. Nielsen and Pedersen (1967) found 1.7% positive samples in cold-smoked salmon with a level of < 1 spore / kg. Hyytiä and others (1998) found that 3% of 64 samples of cold-smoked rainbow trout were positive, with numbers between 40 and 290 spores/kg. In the latter case (Hyytiä and others1998), detection was based on PCR of the E toxin gene, whereas the other studies relied on classical mouse test detection of the toxin, which may contribute to the different results obtained. In hot-smoked fish studies from the mid 1960s, between 1% and 20% of samples were positive with levels of 1-20 MPN / kg. Hot-smoked fish seem to present a higher risk than cold-smoked fish with respect to C. botulinum (Southcott and Razzell 1973). As outlined by Eklund (1992), this may be due to severe underprocessing. Also, the heating step in the hot-smoking process may act as a spore-activating step, inducing germination and posing a greater risk than without a heating step.

3. Growth in refrigerated smoked fish 

Innumerable studies have documented that fish is an excellent substrate for C. botulinum type E. Several trials have specifically evaluated how the combination of low temperature and atmosphere (that is, N2 or CO2) influence toxin production as compared to spoilage patterns. The majority of these studies have been done with fresh (non-salted, non-smoked) fish. Recent reviews contain more information on these studies (Gram and Huss 2000, Lund and Peck 2000).

Similarly, a vast range of studies has shown that hot-smoked and cold-smoked fish are good substrates for C. botulinum and that the organisms may grow and produce toxin, depending on salt and temperature levels. As outlined in Chapter II, growth in naturally contaminated products may be different from growth in inoculated samples. Few studies have evaluated growth and toxin production in naturally contaminated samples of cold-smoked fish. Due to the large outbreaks in the 1960s, several trials have evaluated growth and toxin production in hot-smoked fish. Hot-smoking is a different process from cold-smoking (60 - 70 °C [140 - 158 °F] vs. 20 - 30 °C [68 - 86] °F, internal temperatures). The resulting product is different in structure (that is, the proteins are heat coagulated) and microflora. Cann and Taylor (1979) prepared hot-smoked whole trout from fish naturally contaminated with C. botulinum. Levels of C. botulinum were not reported, but the trout were from a farm where the prevalence consistently was 80% or above (Cann and others 1975). The fish were salted to different levels, vacuum-packed, and stored at 10 °C (50 °F). At 2.5% NaCl, no toxin was detected after 30 d of storage at 10 °C (50 °F). At 2.0% NaCl, only products produced from whole, ungutted fish became toxic. In addition, ungutted fish have caused botulism outbreaks in salted fish such as Kapchunka (Anonymous 1985, 1987).

In recent Finnish studies (Hyytiä and others 1999), vacuum-packed, cold-smoked rainbow trout inoculated with less than 10 C. botulinum spores / g became toxic after 3 wk at 8 °C (46 °F) or after 4 wk at 4 °C (39 °F) with a NaCl concentration of 3.2% (WPS). This is a rapid toxin formation compared to other studies, in particular considering the low spore level and the salt/temperature combinations (Table III-3). Surprisingly, the authors did not detect growth of the bacterium. Graham and others (1996a), who studied the growth and toxin production by C. botulinum using a high inoculum at different combinations of pH, NaCl, and heat activation, concluded that "on no occasion was toxin detected under conditions which did not result in visible growth." In a later study, Graham and others (1997) did detect toxin formation in a few samples where growth was not detectable as turbidity in the tube. The authors (Hyytiä and others 1999) offer some suggestions for the difference between their results and those of other studies (sampling problems, micro-niches, inconsistent NaCl distribution). Currently, no clear conclusions can be reached and further studies are required. 

Table III-3. Growth and toxin production by psychrotrophic Clostridium botulinum depending on salt, temperature, and pH

Inoculum to toxicity Medium to growth

Temperature °C

pH Time Reference
104 /ml Two-phase meat medium 5 °C 2.0 6.2   4 wk Graham and others 1997
  3.0 5.5-6.0   > 13 wk  
6.5 10 wk 10 wk
3.5 5.5-7.0 > 13 wk > 13 wk
10 °C 3.0 6.0   2 wk
6.5-7.0   1 wk
3.5 6.0   12 wk
6.4-7.0   2 wk
105/ml Meat medium 5 °C 2.5 6.5 24-27 d 24-27 d Graham and others 1996a
4.3 6.5 > 104 d > 104 d
8 °C 2.5 6.5 8-9 d 8-9 d
4.3 6.5 14-17 d 14-17 d
105/ml TPG medium 8 °C 2.5 7.0   14 d Segner and others 1966
3.0 7.0   14-16 d
3.5 7.0   21-23 d
105/ml Peptone-sucrose-medium 5 °C 3.0 7.2   > 180 d Emodi and Lechowich 1969
10 °C 4.5 7.2   31 d
102/g and 104/g Hot-smoked salmon 25 °C > 3.8 ns > 2 wk   Pelroy and others 1982
Natural level Hot-smoked fish 10 °C 2.0 ns 30 d1   Cann and Taylor 1979
      3.0   > 30 d    
102/g 3.0 > 30 d  
103/g Hot-smoked fish 15 °C ? ? 5 d   Huss and others 1980
3 x 103/g Hot-smoked fish 27 °C 3.5-5.2 ns > 35 d   Cuppett and others 1987
102/g Cold-smoked trout 4 °C 1.7 6.1-6.3 >28 d   Dufresne 2000
8 °C 1.7 6.2-6.3 >28 d    
12 °C 1.7 6.4-6.5 14 d    
ns2 Laboratory media 5 °C 2.0 ns 50 d   McClure and others 1994
3.0 60 d  
10 °C 3.0 18 d  
1 only ungutted fish; ns = not stated

4. Effect of processing steps and preservation parameters

4.1. Freezing 

Frozen storage does not affect spores but is an efficient way to completely prevent germination and growth of C. botulinum. If the product is vacuum-packaged, frozen storage will have little adverse effects on sensory quality (that is, lipid oxidation).

4.2. Cold smoking 

Although the smoking process uses temperatures at which C. botulinum grows well, cold smoking is a highly aerated process; therefore, no growth of the organism occurs during this processing step.

4.3. Combinations of salt and low temperature 

In combination, salt and low temperature are the two major factors controlling growth of psychrotrophic C. botulinum in cold-smoked fish. This section attempts to draw data and conclusions from a wide range of studies that have evaluated the inhibitory potential of these two factors. To draw conclusions, the following points on the methodology of the studies should be considered:

  • Due to the technical difficulty in achieving low levels of spores, high numbers (often exceeding the natural levels by several orders of magnitude) have been used in most studies. This will overestimate the risk because probability of outgrowth becomes greatly reduced at low spore levels. The use of high inocula can, however, be justified as an introduction of a safety margin.

  • Studies are based on inoculation trials. It is possible that growth of the natural population is slower or more inhibited than that of the inoculated strains. This has been documented for L. monocytogenes (see Chapter II). In contrast, culture collection strains may be less adapted to the food environment than naturally occurring strains.

Psychrotrophic C. botulinum will, under otherwise optimal conditions, grow in up to 5% WPS (aw = 0.97); however, at reduced temperatures, less salt is tolerated. Also, pH influences salt tolerance. NaCl-concentrations of less than 5% are required to inhibit growth, as pH is reduced from neutral (pH = 7.0) (Table III-3). The pH of cold-smoked salmon varies between 6 and 6.3. In general, pH of fish varies between 6 and 7. It has been suggested that the high redox potential (Eh) in some fish (caused by the presence of trimethylamine oxide) can render the bacterium more susceptible to NaCl than under conditions of low Eh (Huss 1980).

Several studies have evaluated the time for growth and/or toxin production of psychrotrophic C. botulinum as related to temperature, salt and pH. A compilation of data relevant for cold-smoked fish is given in Table III-3. The studies can be divided into studies using laboratory media and, typically, high inocula (Segner and others 1966, McClure and others 1994; Graham and others 1996a, 1996b, 1997), and studies using fish, typically with lower inocula (Cann and Taylor 1979; Cuppett and others 1987; Dufresne and others 2000). The levels of spores used in fish inoculation studies, however, are still several magnitudes of order above natural levels (compare Tables III-2 and III-3).

The inhibition of the formation of botulinum toxin by psychrotrophic strains can be ensured using a minimum level of 5% NaCl (WPS). Combining NaCl control with chill storage can prevent toxin formation for certain periods of time. In laboratory media using high inocula (104-105 spores / ml), toxin growth is prevented for at least 2 - 3 wk at 2% - 2.5% NaCl and for at least 4 - 5 wk at 3% - 3.5% NaCl at 5 °C (41 °F). At 10 °C (50 °F), 2% - 2.5% NaCl prevents toxin growth for at least 1 wk and 3% - 3.5% NaCl for at least 1 - 2 wk. Some studies (McClure and others 1994, Emodi and Lechowich 1969) report longer "safe periods" than those just stated.

It is striking that studies using fish as substrate have shown that similar salt/temperature combinations result in significantly longer time to growth and toxin production. Thus, at 1.7% NaCl, toxin was not detected for 4 wk in cold-smoked trout stored at 4 °C (39 °F) or 8 °C (46 °F) (Dufresne and others 2000) and at 2.5%-3.5% NaCl, toxin was not detected in hot-smoked trout stored at 10 °C (Cann and Taylor 1979). The model studies using laboratory media would point to very restricted safe storage times, for example, maximum 4 wk at 5 °C with 3.5% NaCl and only 1 wk at 10 °C with 3.5% NaCl. Based on the fish studies, the history of cold-smoked fish not being involved in cases of botulism, a level of 3.5% NaCl appears sufficient to prevent toxin formation for up to 4 wk at chill temperatures. Thus, the recommendation by the Advisory Committee on the Microbiological Safety of Foods (ACMSF 1992) was that 3.5% NaCl (WPS throughout the product) results in a safe (from psychtrotrophic C. botulinum) product for at least 4 wk at 4 °C - 5 °C. This panel therefore concludes that 3.5% NaCl (WPS) combined with a maximum storage time of 4 wk at 40 °F (4.4 °C) will result in C. botulinum-safe products of vacuum-packed, cold-smoked fish. Similar, although less specific in terms of storage time, conclusions have been reached by Eklund (1992) and Huss and others (1995).

4.4. Atmosphere 

Clostridium botulinum is an anaerobic organism and is sensitive to oxygen. Sensitivity to redox potential (Eh) is not as pronounced. Therefore, growth and toxin production may occur at high Eh if compounds other than O2 are used to establish a positive Eh (Lund and Peck 2000). Due to the intolerance to O2, most attention has been paid to vacuum- and CO2-packaged products. Many studies have documented that O2 removal enhances toxin formation (Eklund 1992), but several studies have found that toxicity may also occur with oxygen present (Table III-4). Thus, Huss and others (1980) found that air-packaging delayed toxin formation by C. botulinum type E in hot-smoked herring stored at 15 °C (59 °F), compared to vacuum-packaging if the fish were handled under aseptic conditions and C. botulinum type E was able to grow and form toxin under 100% O2 atmosphere. Kautter (1964) also reported that toxin could be produced without packaging. In fish contaminated with aerobic spoilage bacteria, toxicity occurred after 4 - 5 d when vacuum-packed, compared to 5-6 d when air-packed (Table III-4). Thatcher and others (1962) reported that hot-smoked fish packed in plastic wrappers had caused cases of botulism. They, therefore, investigated the influence of atmosphere on toxin formation in fish surface inoculated with 103 spores / g. After 8 d at 30 °C (90 °F), both samples incubated under anaerobic and aerobic conditions were toxic. In a study of spoilage and botulinum toxin formation in cold-smoked trout, Dufresne and others (2000) found that at 8 °C (46 °F), fish packed in high O2-transmission films became toxic before fish packed in low O2-transmitting films (Table III-5 and 6). As implied in these studies, although there is no doubt that vacuum-packing and CO2-packing may enhance toxin formation, aerobic packaging or the inclusion of O2 in modified atmosphere packaging (MAP) cannot be relied upon as a safeguard. ACMSF (1992), an advisory body reporting to the Department of Health under the UK-Ministry of Agriculture, Fisheries, and Food (UK-MAFF), concluded on the safety hazards of C. botulinum in vacuum-packed foods: "It is now recognized that the growth of C. botulinum in foods does not depend on the total exclusion of oxygen, nor does the inclusion of oxygen as a packaging gas ensure that growth of C. botulinum is prevented. Anaerobic conditions may occur in microenvironments in foods that are not vacuum- or modified-atmosphere packaged. For example, in the flesh of fish, conditions which are favorable to toxin production can exist in air-packaged fish as well as in vacuum- or modified atmosphere-packaged fish."

In cold-smoked fish, aerobic conditions lead to faster spoilage than under vacuum- or MA-packaging (Table III-5). Under aerobic conditions pseudomonads, yeast, and some lactic acid bacteria develop, whereas anoxic packaging conditions result in development of a lactic acid bacteria flora with a minor component of gram-negative bacteria. Typically, shelf life is reduced by a factor of 1.5 to 2 by aerobic storage as compared to vacuum-packed storage (Table III-6). 

Table III-4. Toxin production in hot-smoked herring surface inoculated with Clostridium botulinum type E spores (103 /g). Some fish were also surface inoculated with 103/g of spoilage bacteria (Shewanella putrefaciens). Fish were packed in air or vacuum-packed and stored at 15 °C (59 °F) (Huss and others 1980).

spoilage bacterium packaging Eh, mV

 mm Hg

No of toxic samples (two sampled on each day) after storage at 15°C
11 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
- Air +250 - 0 120-80 nt2 nt nt 0 0 1 1 0 1 0 1 1 1 1 0 2
- vacuum +200 - 0 30-20 nt 0 0 0 2 2 2 nt nt nt nt nt nt Nt nt nt
+ Air +200 - +150 130-100 nt nt 0 0 1 2 2 2 nt nt nt nt nt Nt nt nt
+ vacuum +200 - 0 0 nt nt 0 1 2 2 2 2 nt nt nt nt nt Nt nt nt

1 Days of storage

2 nt = not tested

Table III-5. Shelf life (determined by sensory assessment) of cold-smoked fish depending on temperature and packaging atmosphere (From Jeppesen 1988, Dufresne and others 2000)


oxygen transmission cc/m2/d/atm


Shelf life, days Reference
4 °C 5 °C 8 °C 10 °C
vacuum; riloten film 20/60 41 4-5   21-28   14 From Jeppesen 1988
aerobic; polyethylen 70 "oxygen permeable" 4-5 28 14   7-14 From Jeppesen 1988
  122 1.7 28   14   Dufresne 2000
  2,9502 1.7 22   13    
  4,9202 1.7 18   16    
  10,0402 1.7     6    

1 at 75% RH, 25 °C

2 at 0% RH, 24 °C

Table III-6. Sensory shelf life (odor) and time to botulinum toxin formation (inoculated with 102 spores of Clostridium botulinum / g) in cold-smoked trout with 1.7% NaCl (WPS) stored at 4 (39 °F) or 8 °C (46 °F) in packaging films with different oxygen transmission rate (modified from Dufresne and others 2000)




at 24 °C, 0% RH

Shelf life, days

No of toxic samples (2 sampled)

after 7, 14, 21 and 28 d

7 14 21 28
4 °C 12 ~ 28 0 0 0 0
  2,950 ~ 28 0 0 0 0
  4,920 22 0 0 0 0
  10,040 18 0 0 0 0
8 °C 12 14 0 0 0 0
  2,950 13 0 0 0 0
  4,920 16 0 0 0 1
  10,040 6 0 0 0 1
12 °C 12 11 0 1 2 1
  2,950 12 0 1 2 2
  4,920 11 0 1 2 2
  10,040 6 0 1 2 2

The United States requires that vacuum-packed, cold-smoked fish contain 3.5% NaCl (water phase) or 3.0% if combined with 200 ppm nitrite. Only 2.5% NaCl is required of aerobically packed fish, which spoil more rapidly. No clear definition of an aerobic pack exists. The more rapid spoilage not the presence of oxygen is relied upon as a safeguard against C. botulinum. Recent data by Dufresne and others (2000) showed that in aerobic-packaged, cold-smoked trout (with 1.7% WPS) stored at 8 °C (46 °F), toxin formation occurred more rapidly when packaged under high O2-transmission than under low O2-transmission. The data emphasize that although spoilage did occur more rapidly under the highest O2 transmitting film (10,000 cc / m2 / d / atm @24 °C, 0% RH), toxin formation also occurred more rapidly, and oxygen was no safeguard against botulinum toxin formation.

4.5. Nitrite 

Nitrite has for some time been used in cured meats and some fish. It is an efficient anti-botulinogenic compound. The effect of nitrite is influenced by pH, NaCl, and temperature. The exact mechanism of its effect is not known. Nitrate, probably via conversion to nitrite, inhibits or delays botulinum toxin formation. The addition of nitrite may reduce the amounts of NaCl required to inhibit C. botulinum toxin formation (Pelroy and others 1982). NaCl concentration may be reduced in vacuum-packed fish by adding 200 ppm nitrite. Due to concerns raised about the potential carcinogenic effects of nitrosamines, however, there is some reluctance to increase use of this compound.

4.6. Lactate 

Lactate is inhibitory to psychrotrophic C. botulinum. Meng and Genigeorgis (1993) found that the lag phase of 104 spores / sample of turkey roll was prolonged from 8 h to 28 h at 8 °C (46 °F) when 2% lactate was added. The effect of lactate was more pronounced with the concurrent addition of NaCl; adding 2% NaCl to the 2% lactate increased the lag phase to 58 d.

4.7. Sorbate 

Sorbate is a GRAS compound and may therefore be used in cold-smoked fish. In some studies, sorbate has been reported to inhibit spore germination, whereas others have found no effect of 1% sorbate (at pH 6.7) on germination (cited from Kim and Foegeding 1993). Vegetative cell growth is inhibited by sorbate, particularly at low pH. For instance, sorbic acid is most effective at pH < 6.0 - 6.5 (Lund and others 1987, cited by Kim and Foegeding 1993).

4.8. Role of background microflora 

As indicated in Table III-4, the aerobic microflora of some products may serve to enhance the risk of toxin formation, probably due to the depletion of O2 by the respiration of the non-botulinum microflora (Abrahamsson and others 1965, Huss and others 1980, ACMSF 1992). In contrast, the growth of some other microorganisms may inhibit growth and toxin production by C. botulinum. Thus, Lyver and others (1998) reported that certain Bacillus isolates were inhibiting toxin production. Also, lactic acid bacteria, either by acid or by bacteriocin production, may inhibit growth and toxin formation of C. botulinum (Kim and Foegeding 1993). Relying on competition from naturally occurring background flora to restrict C. botulinum, however, is not an effective or reproducible way of preventing growth.

5. Conclusions 

The following conclusions are based on a thorough analysis and evaluation of the current science on control methods of C. botulinum in cold-smoked fish:

  • Psychrotrophic C. botulinum occurs naturally in the aquatic environment, so its presence in low numbers on fresh fish must be anticipated. Spores may also be isolated infrequently from cold-smoked fish, although numbers, if present, are low. Given this low number, the probability of germination and toxin production is low but present.

  • Experiments with naturally contaminated hot-smoked fish produced from fish with high levels of C. botulinum show that toxin may be formed under conditions of temperature abuse.

  • Toxin production by psychrotrophic C. botulinum is controlled with a combination of a moderate level of NaCl (3.5% NaCl WPS) and storage at chill temperature (<4.4 °C, <40 °F) for at least 4 wk. Based on the scientific data and because commercially produced cold-smoked fish has never been reported as a source of botulism, it is reasonable to conclude that the salt and cold keep the hazard under adequate control.

  • Based on a range of model studies in broth and inoculation studies with hot- or cold-smoked fish, it can be concluded that a combination of 3.5% NaCl (WPS) and chill storage (4.4 °C, 40 °F), allowing for short periods of elevated temperatures up to 10 °C (50 °F), will prevent toxin formation in reduced oxygen packaging cold-smoked fish for several weeks beyond its sensory shelf life.

  • As a general safeguard, salting to 3.5% (WPS) for chilled, stored cold-smoked fish is essential for reduced oxygen packaged (ROP) cold-smoked fish. The requirement for chilling with a sufficient salt concentration is an option to considered in national or international regulations (for example, E.U. directives).

  • For air-packaged products, levels of NaCl can, theoretically, be reduced; however, scientific data that support this argument do not exist and are needed before any reduction is recommended. Even when not packed under vacuum- or modified atmosphere, pockets of anaerobic conditions may be created where slices of fish overlap or where aerobic spoilage bacteria consume the oxygen present.

6. Research needs 

The following is a list of research areas that the panel suggests need further attention:

  • Evaluate growth and toxin production in naturally contaminated cold-smoked fish products to validate models and predictions for growth and toxin production.

  • Determine the influence of redox potential, various concentrations of trimethylamine oxide (TMAO), and NaCl on toxin production by psychrotrophic C. botulinum in gadoid and non-gadoid species.

  • Determine the potential facilitation by TMAO on formation of nitrosamines, if nitrite is added, during cold smoking.

  • Identify processing conditions and gas transmission rates of films under various time/temperature conditions for products to be considered "air packaged." Determine the oxygen transmission rates (OTR) needed for a product with 2.5% salt concentration to provide equivalent safety compared with cold-smoked ROP products.

  • Conduct challenge studies on air-packaged, cold-smoked fish in films with OTR between 7,000 and 10,000 cc / m2 / 24 h and compare to unpackaged cold-smoked fish.

  • Establish minimum WPS concentrations required to inhibit growth and toxin formation by C. botulinum in air-packaged and unpackaged cold-smoked fish.

  • Determine the shelf life of the product, relative to product quality as well as safety, under different packaging methods and storage temperatures.

  • Determine appropriate sell-by dates and evaluate the use of time-temperature indicators to ensure a safe product. 


Abrahamsson K, De Silva NN, Molin N. 1965. Toxin production by Clostridium botulinum, type E, in vacuum-packed, irradiated fresh fish in relation to the changes to the associated microflora. Can J Microbiol 11:523-9.

[ACMSF] Advisory Committee on the Microbiological Safety of Foods. 1992. Report on vacuum packaging and associated processes. London (UK): Her Majesty's Stationery Office.

Anonymous. 1985. Botulism associated with commercially distributed Kapchunka-New York City. MMWR 34(35):546-7.

Anonymous. 1987. International outbreak of type E botulism associated with ungutted, salted whitefish. MMWR 36(49):812-3.

Cann DC, Taylor LY. 1979. The control of the botulism hazard in hot-smoked trout and mackerel. J Food Technol 14:123-9.

Cann DC, Taylor LY, Hobbs G. 1975. The incidence of Clostridium botulinum in farmed trout raised in Great Britain. J Appl Bacteriol 39:331-6.

Cann DC, Wilson BB, Shewan JM, Hobbs G. 1966. Incidence of Clostridium botulinum type E in fish products in the United Kingdom. In: Nature. p 205-6.

Cuppett SL, Gray JI, Pestka JJ, Booren AM, Price JF, Kutil CL. 1987. Effect of salt level and nitrite on toxin production by Clostridium botulinum type E spores in smoked great lakes whitefish. J Food Prot 50(3):212-7.

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.

Dodds KL, Austin JW. 1997. Clostridium botulinum. In: Doyle MP, Beuchat LR, Montville TJ, editors. Food Microbiology. Fundamentals and Frontiers: American Society for Microbiology. p 288-304.

Dufresne I, Smith JP, Liu JN, Tarte I, Blanchfield B, Austin JW. 2000. Effect of films of different oxygen transmission rate on toxin production by Clostridium botulinum type E in vacuum packaged cold and hot smoked trout fillets. J Food Saf 20:251-68.

Eklund MW. 1992. Control in fishery products. In: Hauschild AHW, Dodds KL, editors. Clostridium botulinum: Ecology and control in foods. New York: M Dekker. p 209-32.

Eklund MW, Poysky F. 1967. Incidence of Cl. botulinum type E from the pacific coast of the United States. In: Ingram M, Roberts TA, editors. Botulism 1966. [unknown]: Chapmann and Hall.

Emodi AS, Lechowich RV. 1969. Low temperature growth of type E Clostridium botulinum spores. 1. Effects of sodium chloride, sodium nitrite and pH. J Food Sci 34:78-81.

From Jeppesen. 1988. Results in report: "Predictive microbiology - measurement and control of food quality". Lyngby: Danish Institute for Fisheries Research, Dept. of Seafood Research.

Graham AF, Mason DR, Maxwell FJ, Peck MW. 1997. Effect of pH and NaCl on growth from spores of non-proteolytic Clostridium botulinum at chill temperature. Letts Applied Microbiol 24:95-100.

Graham AF, Mason DR, Peck MW. 1996a. Inhibitory effect of combinations of heat treatment, pH, and sodium chloride on growth from spores of nonproteolytic Clostridium botulinum at refrigeration temperature. Appl Environ Microbiol 62(7):2664-8.

Graham AF, Mason DR, Peck MW. 1996b. Predictive model of the effect of temperature, pH, and sodium chloride on growth from spores of non-proteolytic Clostridium botulinum. Int J Food Microbiol 31:69-85.

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.

Heinitz ML, Johnson JM. 1998. The incidence of Listeria spp., Salmonella spp., and Clostridium botulinum in smoked fish and shellfish. J Food Prot 61(3):318-23.

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

Huss HH. 1981. Clostridium botulinum type E and botulism [DSci thesis]. Lyngby (DK): Technical University, Technological Laboratory of the Ministry of Fisheries. 58 p.

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

Huss HH, Pedersen A, Cann DC. 1974. The incidence of Cl. botulinum in Danish trout farms. II: Measures to reduce the contamination of the fish. J Food Technol 9:451-8.

Huss HH, Petersen ER. 1980. The stability of Clostridium botulinum type E toxin in salty and/or acid environment. J Food Technol 15:619-27.

Huss HH, Schaeffer I, Pedersen A, Jepsen A. 1980. Toxin production by Clostridium botulinum type E in smoked fish in Ireation to the measured oxidation reduction (Eh) potential, packaging method and the associated microflora. In: Connell JJ, editor. Advances in Fish Science and Technology: Fishing News Books Ltd. England. p 476-9.

Hyytia E, Hielm S, Korkeala H. 1998. Prevalence of Clostridium botulinum type E in Finnish fish and fishery products. Epidemiol Infect 120:245-50.

Hyytia E, Hielm S, Mokkila M, Kinnunen A, Korkeala H. 1999. Predicted and observed growth and toxigenesis by Clostridium botulinum type E in vacuum-packaged fishery product challenge tests. Int J Food Microbiol 47:161-9.

Kautter DA. 1964. Clostridium botulinum type E in smoked fish. J Food Sci 29:843-9.

Kim J, Foegeding PM. 1993. Principles of Control. In: Hauschild AHW, Dodds KL, editors. Clostridium botulinum. Ecology and control in foods. New York: M Dekker. p 121-76.

Lund BM, George SM, Franklin JG. 1987. Inhibition of type A and type B (proteolytic) Clostridium botulinum by sorbic acid. Appl Environ Microbiol 53:935-.

Lund BM, Peck MW. 2000. Clostridium botulinum. In: Lund BM, Baird-Parker TC, Gould GW, editors. The microbiological safety and quality of foods. Gaithersburg (MD): Aspen. p 1057-109.

Lyver A, Smith JP, Austin J, Blanchfield B. 1998. Competitive inhibition of Clostridium botulinum type E by Bacillus species in a value-added seafood product packaged under a modified atmosphere. Food Res Int 31(4):311-9.

McClure PJ, Cole MB, Smelt JPPM. 1994. Effects of water activity and pH on growth of Clostridium botulinum. J Appl Bacteriol Symp Suppl 76:105S-114S.

Meng J, Genigeorgis CA. 1993. Modeling lag phase of nonproteolytic Clostridium botulinum toxigenesis in cooked turkey and chicken breast as affected by temperature, sodium lactate, sodium chloride and spore inoculum. Int J Food Microbiol 19:109-22.

Nickerson JTR, Goldblith SA, DiGioia G, Bishop WW. 1967. The presence of Cl. botulinum, type E in fish and mud taken from the gulf of Maine. In: Ingram M, Roberts TA, editors. Botulism 1966. [unknown]: Chapman and Hall.

Nielsen SF, Pedersen HO. 1967. Studies of the occurrence and germination of Cl. botulinum in smoked salmon. In: Ingram M, Robers TA, editors. Botulism 1966. [unknown]: Chapman and Hall. p 66-72.

Pelroy GA, Eklund MW, Paranjpye RN, Suzuki EM, Peterson ME. 1982. Inhibition of Clostridium botulinum types A and E toxin formation by sodium nitrite and sodium chloride in hot-process (smoked) salmon. J Food Prot 45(9):833-41.

Segner WP, Schmidt CF, Boltz JK. 1966. Effect of sodium chloride and pH on the outgrowth of spores of type E Clostridium botulinum at optimal and suboptimal temperatures. Appl Microbiol 14(1):49-54.

Southcott BA, Razzell WE. 1973. Clostridium botulinum control in cold-smoked salmon: a review. J Fish Res Bd Can 30(5):631-41.

Sugiyama H, Bott TL, Foster EM. 1970. Clostridium botulinum type E in an inland bay (Green Bay of Lake Michigan). In: Herzberg M, editor. Toxic Microorganisms. Washington D.C.: U.S. Dept. of the Interior. p 287-91.

Thatcher FS, Robinson J, Erdman I. 1962. The "vacuum pack" method of packaging foods in relation to the formation of the botulinum and staphylococcal toxins. J Appl Bacteriol 25:120-4.

Woodburn MJ, Somers E, Rodriguez J, Schantz EJ. 1979. Heat inactivation rates of botulinum toxins A, B, E and F in some foods and buffers. J Food Sci 44:1658-61.

Page Last Updated: 08/04/2015
Note: If you need help accessing information in different file formats, see Instructions for Downloading Viewers and Players.
Language Assistance Available: Español | 繁體中文 | Tiếng Việt | 한국어 | Tagalog | Русский | العربية | Kreyòl Ayisyen | Français | Polski | Português | Italiano | Deutsch | 日本語 | فارسی | English